This patch is adding the _NET_WORKAREA property
which according to EWMH spec MUST be implemented.
Some application like rdesktop or nautilus are using
it, that's maybe why you encountered some errors like these:
** (nautilus:6457): WARNING **: Can not get _NET_WORKAREA
** (nautilus:6457): WARNING **: Can not determine workarea, guessing at layout
For now, the property is only updated when a workarea is created or destroyed.
This patch is updating the coding style based on checkpatch output
and as suggested by Rodolfo it also inverts the condition test
in getAnimationGeometry fct to return earlier when possible.
This patch is fixing the shortcut label used in window menu.
As now when a 'control' shortcut was used, the label displayed
was 'bug' cause GetShortcutString fct which is calling wXModifierFromKey
fct was waiting for a 'Control' static string not the default one
set to 'Ctrl' string.
This patch is fixing the RenderBadPicture X errors on deiconifying
window using double-click.
Background:
I also experienced this issue these days, after checking it was not
introduce with latest apercu feature I decided to dig into it.
As far as I saw in the archive it was first reported by Rodolfo in June 2012.
Each time a window is deiconify using double-click, the error below
is reported in the session error log:
wmaker(catchXError(startup.c:177)): warning: internal X error:
RenderBadPicture (invalid Picture parameter)
Request code: 152
Request minor code: 7
Resource ID: 0x6000a4
Error serial: 9269
So I decided to track when the issue is happening,
from src/action.c wDeiconifyWindow then in src/icon.c wIconDestroy
then in src/wcore.c wCoreDestroy.
At that stage, all the structures I checked are cleaned up properly
during the icon destroy process.
So I also checked when the issue is generated, I ended in
wrlib/xutil.c RCreateXImage (in the share memory extension code)
XSync(context->dpy, False);
oldErrorHandler = XSetErrorHandler(errorHandler);
XShmAttach(context->dpy, &rximg->info);
XSync(context->dpy, False);
XSetErrorHandler(oldErrorHandler);
At this point I was quite lost, cause it means the error could
have been in the server error queue already.
Facts, seems the issue is appearing only on double-clicking the miniwindow,
but not when "Deminiaturize" is clicked from the right-click contextual menu.
I also checked the code from winmenu.c when the action MC_MINIATURIZE
is selected from execMenuCommand fct, but it was just calling
wDeiconifyWindow function. Hum.
So, after one full day of thought I tested to just change the focus before
double-clicking the miniwindow, and bingo it works.
That means when the icon is destroyed the X server still has a reference
on it that's generating the issue.
This patch is used to fix the Debian bug #140181
It updates the NumLockMask function to report the mask and
call XFreeModifiermap when needed.
It also modifies capture_shortcut function to be able to use Super keys
and report numeric keypad when pressed (aka now detects correctly
if the numlock is enabled or not).
This patch is changing the default application icon to
a one that reminds the NeXTcube but also a blackbox of
something we don't know about.
It's from LuBu OpenMagic 1.0 for Sparc project at
www.alge.no/index.php/OpenMagic_1.0
According to the license,
"LuBu OpenMagic 1.0 is put into the public domain"
Since the resolution of the Retina display tends to make everything small,
the default apercu preview size (twice the icon size) couldn't be used to
distinguish the window contents without tiring too much my eyes.
Therefore, let's make the apercu size a configurable option. You can set
it through the ApercuSize variable with
$ wdwrite WindowMaker ApercuSize 4
in multiples of the icon size (in this case the apercu size will be four
times the icon size).
The default size remains 2 (twice the icon size).
This patch is adding miniwindow apercu when the mouse
is over the miniwindows.
To enable it you have to run WPref, in Miscellaneous Ergonomic
Preferences, check miniwindow apercus.
Then, you will be able to see a screenshot of the app when the mouse
is over the miniwindow.
This patch is setting the app logo max size to 128 as what can
be found in the rest of the code, so the logo is not restricted
anymore to the icon size defined by WPrefs.
The goal is to solve the resize issue that appears when the logo
is used in Info and Exit panel.
This patch is fixing compilation issue:
moveres.c:1801:65: warning: unused parameter 'dx' [-Wunused-parameter]
by removing variable 'dx' from getResizeDirection function.
This patch is removing the experimental status of
drag-nd-drop support on the dock and adding an option
to disable the feature from the configure step.
This patch is ensuring that the data dropped is a file format URI
and converting it as stated by XDND specs.
It solves:
*crash in my env when arbitrary stuff is dropped
*correct handling of non ascii chars
As reported by cppcheck:
[WINGs/array.c:129] -> [WINGs/array.c:131]: (warning) Possible null pointer dereference: array - otherwise it is redundant to check it against null.
[WINGs/array.c:151] -> [WINGs/array.c:153]: (warning) Possible null pointer dereference: array - otherwise it is redundant to check it against null.
[WINGs/array.c:170] -> [WINGs/array.c:172]: (warning) Possible null pointer dereference: array - otherwise it is redundant to check it against null.
This patch is checking that the var name 'array' exists.
As reported by cppcheck:
[src/main.c:141]: (error) Common realloc mistake: 'wVisualID' nulled but not freed upon failure
The patch is using wrealloc instead of the standard realloc().
As reported by cppcheck:
[src/superfluous.c:238] -> [src/superfluous.c:199]: (warning) Possible null pointer dereference: aicon - otherwise it is redundant to check it against null.
[src/superfluous.c:239] -> [src/superfluous.c:199]: (warning) Possible null pointer dereference: aicon - otherwise it is redundant to check it against null.
[src/superfluous.c:240] -> [src/superfluous.c:199]: (warning) Possible null pointer dereference: aicon - otherwise it is redundant to check it against null.
The patch is adding a test to check if variable aicon exists.
As reported by cppcheck:
[src/dock.c:568] -> [src/dock.c:571]: (warning) Possible null pointer dereference: aicon - otherwise it is redundant to check it against null.
The aicon test is useless, the patch is removing it.
cppcheck is reporting:
[wrlib/tests/testgrad.c:148]: (error) Memory leak: color_name
but seems some other variables were never freed.
This patch is cleaning them property.
WRaster:
- new function 'RShutdown'
- removed flag 'optimize_for_speed' from RContext
- new functions '...(operate_xxx)', '...(flip_image)'
WUtil:
- new function 'wutil_shutdown'
- new macro 'wlengthof'
WINGs:
- new function 'WMReleaseApplication'
- new function 'WMCreateScaledBlendedPixmapFromFile'
(And maybe a few more I missed)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The commit #d1f9b801 removed it because the code behind it is actually
useless, unfortunately removing the flag itself from the structure is
breaking the compatibility of the API.
To avoid problems, this patch re-introduces the flag in the structure, but
not the code behind it so it is useless, and adds an attribute so that gcc
will report its use as deprecated, to help application migration.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original behaviour was a bit frustrating because it used the first
small mouse move to deduce the user's wished direction, but that is a bit
imprecise.
This patch divides the window in 9 rectangles:
- 4 for corners, used for diagonal resizing
- 4 for middles of sides, for resizing Up, Down, Left and Right
- 1 useless in the middle of the the window, which falls back on diagonal
resizing
This leads to a more predictive behaviour.
Recommend usage for the configure flag 'enable-debug' when testing code and
invite to make use the the checkpatch.pl script.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It makes the code easier to read to explicitly define a type for the
functions that are used for callbacks, so this patch does this for the
wmmenugen tool.
It was an opportunity to highlight some variable definitions that looked
like function prototypes, and were as such misplaced in the code, being a
source of confusion.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When a command is given for a menu to be run through a terminal, there is
a piece of code that removes everything from a '!' to the end of the line.
The original code was too dense and not really optimal, so this patch
proposes a more explicit code, mainly for maintainability.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Having all these information spread in different places makes it error
prone when wanting to add/remove/change something in the list are there are
many unrelated places to keep in sync.
By merging everything in a single struct array it is a lot easier to
maintain this list.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Having all these information spread in different places makes it error
prone when wanting to add/remove/change something in the list are there are
many unrelated places to keep in sync.
By merging everything in a single struct array it is a lot easier to
maintain this list.
Took the opportunity to properly document a little hack which is used to
avoid complex handling for a special case.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When a list has a fixed size, it is better to use the size of the list
instead of an end-of-list mark because it avoid unnecessary memory usage
and ease compiler's optimisation task.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The code uses an extra variable that is specific the the Xkb code but it
is not really needed, removing it makes the code simpler, thus easier to
maintain.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In the function that creates the background image for the switch panel
there are a number of checks on the size of the image, but this has
already been checked at the beginning of the function with an early return
in this case.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In the function wHandleAppIconMove there is an event handling loop
which was supposed to be finished with the variable 'done'.
As the code is using 'return' inside the loop, which is both more
convenient and easier to read, this variable was never set, making
the loop an infinite looking loop.
This patch removes the variable so it will be clear how it behaves.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The code includes a check whose goal is to avoid reloading more than once
the configuration, in case Inotify would report more than one change during
the event processing.
But the flag was not set so mechanism was ineffective, which is probably
ok anyway but as it is here let's make it work.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function did spend its time re-copying data and searching the
end of the read buffer, which is not really efficient.
The new code reads directly from file to the end of previous data,
avoiding unneccesary duplication, and keeps track of the end of
data to avoid searching it for next read.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It is generally not a good idea to have an API with a high number of
functions because it adds complexity for user and for maintainability,
so both function have been "merged" into a single RFlipImage with
a parameter to specify what flip is expected.
As a bonus, the function can perform both flips at once if wanted.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As it does not cost anything in the current code to not add this channel,
it is then probably a good idea to keep the output image with the same
format as the input image, this avoid wasting +33% of memory for something
that may be at best unused and at worst will induce extra cost when
manipulating the image.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Instead of calling all variables with variants of 'image', changed
the name to reflect what image it is about for clarity.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function RRotateImage contains code to handle efficiently the 4 simple
angles 0, 90, 180 and 270 degrees, which makes it a long function.
This patch separate the code for the different cases into dedicated
functions so the main function's code ends up being simple (aka: easier to
understand/review/maintain).
As a side effect, the function for the 180 degree function is not static
because it can be reused to flip an image both horizontally and vertically.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Added a comment to describe each terminal, so it will be easier to maintain
the list in the future; added const attributes to help compiler generate
better code; changed list parsing to use the size of the array instead of
a null pointer mark for better compiled result.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This macro supposes that the called lib function clears the 'errno'
variable on success which is not the case. The macro was (luckily) not
used in these file yet, by removing it we make sure it won't happen.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Use a better name for local variable and allocate the (almost) correct
number of characters for the path instead of a fixed offset that could be
a problem if the constant part of the path were to be updated.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, when reading the size of data using 'readmsg' that
size cannot be fully trusted (possibly in case of bugs in present case),
so this patch adds a check to ensure it is valid before continuing.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the buffer used to store the command for the menu
has a fixed size, so a check is welcome to avoid buffer overflow.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, it is possible that the variable 'image' remains
NULL in the function parseTexture, so this case must be checked
appropriately where it is used.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, some locally created PropList objects are not freed
when the function returns. The truth is their refCount is 1 more than what
they should be, but that's more than what Coverity can detect today.
This patch adds the appropriate release calls when they are not needed
anymore, which will actually not free them but get their refCount right so
they will be freed as soon as the PropList in which they are used are
released.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the function wstrconcat is allocating memory to
return its result, which is not freed in old coding of the function.
This patch uses a local storage buffer to have a simpler code to generate
the command to bu run with 'system' without leak.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, in some case of texture with image that needs to
be scaled the temporary scaled image would be leaked if and error occured
during its conversion.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the function is creating a temporary array with
the list of keys of the Style dictionnary, but it was not freed in the end.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the string passed to 'findPositionInMenu' is not
freed, and as the function does not modify it it is not necessary to
allocate a new storage for the argument.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there are a few cases where the list of the keys
of a dictionary is being requested, but at the end the array that was
created to hold these keys was not freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the indentation for a block of code could lead to
misinterpretation on when it is executed. To make code safer, re-indented
the code properly and added some blank lines for clarity.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, it is not useful to call RReleaseImage if we are in
the branch of code where we know that the image for which it is being
called is NULL.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the code was making an explicit use of a null
pointer, which is certainly not what was initially expected. The code was
simplified to fix the case and to make it easier to understand and
maintain.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When defining enums as types instead of simple enums allows to use these
types at the places where the corresponding enum values are expected, then
allowing the compiler to check that, potentially reporting incorrect use
of values to the user.
This patch adds the type for the gradient style for RRender*Gradient.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When defining enums as types instead of simple enums allows to use these
types at the places where the corresponding enum values are expected, then
allowing the compiler to check that, potentially reporting incorrect use
of values to the user.
This patch adds the types for the drawing operations used by the functions
who drawing shapes into an RImage.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When defining enums as types instead of simple enums allows to use these
types at the places where the corresponding enum values are expected, then
allowing the compiler to check that, potentially reporting incorrect use
of values to the user.
This patch adds types for the configuration fields in the RContextAttributes
structure.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Using local function prototype redefinition is dangerous, now a single
prototype is seen in all the files that are concerned.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
To be consistent with RCreateXImage and because it should not be up to the
caller to handle this, when XGetImage returns a NULL pointer then the
function RGetXImage will also return NULL instead of an RXImage structure
with a NULL pointer.
This consistent behaviour helps fixing a memory leak in WMaker reported by
Coverity (#50125).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, this value is never used, the assignation has no
effect, so Coverity reports that it is unnecessary complexity that is not
compatible with code maintainability.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there's a free on the content of this buffer at
function's end, but if the image's width or height is lower than 1 then we
could reach this free before allocating anything, thus having an
uninitialised pointer.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, environment variables cannot be considered
trustworthy ("tainted"), so we need to make sure the values provided are
within reasonable bounds for safe operations.
Seized the opportunity to try to provide clearer name for the constants
defining the default value.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the behaviour of fopen/fread/fclose in case of
error is not really what the macro RETRY assumes. So the macro is removed
and appropriate action is implemented.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the shift operation performed for color-to-pixel
transform may not behave as well as expected because of the types used
(Coverity pointing suspicious sign extension because of int type being
involved).
The new code tries to leave no open interpretation to the compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the shift operation performed for color-to-pixel
transform may not behave as well as expected because of the types used
(Coverity pointing suspicious sign extension because of int type being
involved).
The new code tries to leave no open interpretation to the compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch adds the ability to set the colors for IconTitleColor, IconTitleBack,
ClipTitleColor, and CClipTitleColor, the only four colors set by setstyle which
previously were not configurable using WPrefs.
The icon in the preview window appears as a miniwindow if IconTitleColor or
IconTitleBack are being configured and as the clip if ClipTitleColor or
CClipTitleColor are being configured.
Based on libexif feature, this patch is adding orientation
detection and provides 2 more shortcuts for live
right/left rotation. Internal version was bumped to 0.7.
Having the warning displayed will help devs to understand what is going
on and where to look at for a fix.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the allocated value returned by XGetModifierMapping
is assumed to be non-NULL everywhere except when releasing it.
Removed this last check (useless) and added a little check at the beginning
to avoid an (improbable) crash.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the functions WMCreatePLString return a newly
allocated WMPropList, which must be released when not needed anymore.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Coverity, the return string of WMGetTextFieldText is
malloced so it needs to be freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the string returned by 'WMGetTextFieldText' is
allocated dynamically, so it must be freed when not needed anymore.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the function 'WMGetTextFieldText' already allocates
memory for its return string, so it is not necessary to wstrdup it.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if there is a problem when reading the binary data
of an RImage from a file, the area allocated to store it would not be
freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the PLArray that is created to store the return
value of the function 'processData' can be left allocated if some case
handling decide to return NULL instead.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is a possible NULL pointer dereference in the
code that extracts the selected modifier (from a popup button list).
The code assumes that there is always a space between the core name of the
modifier (which is used for saving) and the more user friendly name. This
is true if the list could be successfully generated in fillModifierPopUp,
but it may not be the case if that function has to fall-back to the
default list (XGetModifierMaping failure).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, it is possible to dereference the NULL pointer in
the function 'browseForFile'.
This should have been reported by the compiler ("possible use of
uninitialised variable") but the assignation of a default value to the
variable 'lPtr' hides this, but as the default value is not a good one then
the code could crash.
This patch removes the default value, so we may have some info from the
compiler and handles the case smoothly.
It also fixes a potential memory leak pointed also by Coverity in the
case where the string returned by WMGetFilePanelFileName would be a 0
length non-null string.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the return value of 'sscanf' is not checked, which
may not be a problem because the fall-back value for 'i' is provided before.
However, if sscanf fails there is no guarantee that all implementation would
leave 'i' untouched in case of parsing error, and for convenience it is more
logical to write the code is such a way that the value assignment looks like
the fall-back value it is.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
cppcheck is reporting the msg below:
[../WINGs/winputmethod.c:215] -> [../WINGs/winputmethod.c:209]:
(style) Found duplicate branches for 'if' and 'else'.
The patch is fixing it by setting a default return call.
With a set of images i was able to detect that
the flip functions was shifting the image by 1px.
The 90 and 270 degrees rotation were not working as
the functions were also mirroring the img.
The patch is also fixing the C style based on checkpatch.
when the rotation angle value passed to RRotateImage is a modulo of 90,
the function is working well but in other cases the rotateImage()
function is called. That last function is half implemented but
the half already implemented part is also segfaulting (use the testrot.c
to replay the crash).
So the patch is moving the 'not implemented' msg to disable the whole
rotateImage() function and comment it out.
As pointed by Coverity, the function GetCommandForPid did use a macro
'RETRY' for the 'close' of file descriptor, which is not correct because
on success the function 'close' does not reset errno, and in case of
failure the file descriptor is not more valid anyway.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Factually, the program works because the function 'drawerRemoveFromChain'
uses only the address of the drawer and not the content of the structure,
but conceptually this address point to a storage space that has been
de-allocated, so that's calling for trouble for future code changes.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the function 'wdefaultspathfordomain' cannot
return a NULL path, so it is not necessary to check for it. (in the
present case, it led Coverity to think it was possible to have the
structure 'stbuf' uninitialised)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is one case where the string returned by
WMGetTextFieldText could be left allocated.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is one case where the application dictionary
which is created temporarily (when saving the changes in the Window
Inspector) could be left allocated. It is now freed in the common path to
avoid problems.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Coverity, it was possible to return early from the function
'listCallback' without de-allocating a temporary string. The string is now
requested after this early return.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointer by Coverity, the function wstrappend may move the string in
memory (because that's what realloc does if it can't expand the allocated
buffer in-place), so it is important to use the returned address and not
assume the old pointer is still valid.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the caller of the function 'getSavedState' assumes
that if the function returns 0 then the wstate is not initialised. This was
not actually true, so this patch changes 'getSavedState' to behave as
expected.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the memory allocated to contain the argument list
for spawning ourself was never released.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if the function wShowCrashingDialogPanel is not
able to connect to the default screen, then the memory allocated for the
panel would be leaked.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there are some cases where the name of the file
returned by wIconChooserDialog could be left unfreed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function wIconChooserDialog already makes sure that it returns
NULL if its result would have been an empty string, so it is not
necessary to re-check this in setIconCallback, this leads to
more complicated code and can't be optimised by the compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if the function wScreenInit fails to properly
attach to the screen then the storage structure and content were not
released before returning.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if the target width or height for the center ended
being <= 0 then the image created to contain the generated assemblage would
be leaked.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if there's no window to unmap in a workspace, then
the array used to build the list of the windows to be unmapped for
workspace change would not be freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if an invalid OPEN_MENU is used then the memory
allocated when parsing the line containing it would be leaked.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if the word "WITH" was used in the content of the
root menu definition then its string would not be freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the function 'ProgGetWMClass' always allocates the
strings returned in wm_class and wm_instance, so they always must be freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, a string could be left unfreed in the case where
the window's attribute dialog is closed before an icon would be actually
chosen from the Icon Chooser Dialog.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, a string could be left unfreed in the case where
the docked application's settings dialog is closed before an icon would be
actually chosen from the Icon Chooser Dialog.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, in this specific case the string previously stored
in 'file' was lost when replaced with the string from 'db_icon'. The logic
have been simplified to avoid the case.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there was a memory leak when updating the shortcuts
in the menu for windows which occured when a shortcut was associated with
the window and this shortcut did not change.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is a check to make sure a pointer is not NULL
but it is still dereferenced later in the function. As the code is done so
that this pointer cannot be NULL then it is not necessary to check that.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch adds "FrameFocusedBorderColor" option for theming,
It should allow having borders better matching titlebar colors.
By default it's set to black.
Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>
Long time ago I added this option so I could have wallpaper filling my
screen while preserving aspect ratio.
Seems like I didn't add it to WPrefs
Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>
As raised by Carlos R. Mafra and Doug Torrance, when an image is used in a
texture but the image cannot be loaded, there are some case where WPrefs
misbehaves.
This patch display a warning in every case where the image could not be
loaded, and always switch to black texture as a fallback, which was not the
case previously.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
According to NEWS, the DisplayFont option was removed in version 0.62.0 and
replaced by the WINGs system font. However, it was still referenced in a large
number of style files, as well as the utils getstyle, setstyle, convertfonts,
and the manpages for getstyle and setstyle. This patch removes all of these
references.
In addition, it updates the Czech, Slovak, and Russian translations of these
manpages to match the English versions. Note that I relied on Google translate, and it appears I may have re-encoded some characters when I saved the Czech
and Slovak files -- if you speak these languages, please take a look!
A few minor spelling errors were corrected in the English manpages.
The coding style in wmaker.git follows the conventions of the linux kernel.
Therefore, I'm stealing their 'checkpatch' tool in an attempt to maximize
"perfect" patches to wmaker-dev, at least coding-style wise :-)
Before submitting your patch, please ('please' here means you must) check
it with:
./checkpatch your-patch.diff
and use your common sense to address any violations (error and/or warnings) detected.
There will always be false positives and other complaints that sometimes are better
dealt with by leaving them alone. No tool is perfect, but it _will_ detect obvious
errors which you must fix (warnings can be sometimes ignored).
I modified the script by removing kernel-specific stuff, but I don't guarantee
that I removed everything. I also increased the line length warning to 115.
In any case, it seems to work here and hopefully it will lead to a cleaner
and more consistent code base for wmaker.
PS: You can also check existing files, e.g.,
./checkpatch.pl --file src/icon.c
The color well contains two views: view, consisting of the outer "button", and
colorView, consisting of the color itself. Previously, only clicking on view
would bring up the corresponding color panel, resulting in nonintuitive
behavior.
In response to Amadeusz's comments regarding my patch "WPrefs: Add ability to
edit FrameBorderColor/FrameSelectedBorderColor", this new patch attempts to
reduce the confusion between focused and selected windows. In particular,
only the frame border color for non-selected windows appears in the preview.
The frame border color for selected window colors can still be changed, but it
does not appear in the preview. In addition, the hand icon has been removed
from the preview for these two options.
I also looked into the possibility of adding a new border color for focused
windows. This appears to be nontrivial, as the border color belongs to
WFrameWindow, but the state of being focused belongs to WWindow.
When the condition can be simply checked in an if, it is better to do that
than to use a flow breaking goto.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Coverity, the structure containing the callback functions
to handle the drag-and-drop actions could be freed but later re-used.
The correct behaviour is to keep the allocated memory for the structure.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the 'paintItem' function in the WMBrowser widget
is checking for nullity of its text argument, but before that it called the
strlen function which crashes on NULL pointer. This patch moves the strlen
call to the right place and reduce the lifespan of 'textLen' to highlight
incorrect tries to use the variable.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is a NULL pointer check at the end of the
function 'deleteFile' (which happens to not be necessary because wfree was
made to accept NULL pointers), however there are many things done before
that assumes the pointer is not NULL. The check is moved to where it
matters.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the variable 'dir' is passed by reference because
the function may update it; as it is never NULL (the function is local)
checking for that does not make sense. From the actual usage, it seems
logical that the check should have been on the value, not on the pointer.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the code was checking for NULL pointer to avoid
misbehaviour, but it actually dereferenced the pointer beforehand so the
crash would still happen.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the data returned by WMGetTextSelectedStream are
actually a newly allocated buffer, so we need to release it when we're done
with it.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the array used to build the result for the
function 'getStreamObjects' could leak in case of early return.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Coverity, in the (rare) case where the wstrlcat would fail
the array for the result would not be freed before returning.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Coverity, the function 'WMGetTextFieldText' allocates memory
to generate the result string, so it has to be freed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, there is a small check for NULL pointer when
handling triple-click selection in a WMText, however this check is only
partially implemented so the code would crash later anyway.
This patch implement an appropriate skip to avoid crash, and includes a
log message to help debug if the case happens.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, despite apparent checks the function would fail if
trying to set a NULL title. Instead of trying to do over-complicated things
we just ignore (with a warning if NDEBUG is not set) the case. If user
wants an empty title, the correct way is to provide "" as we're not able to
remove a title.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by coverity, the case 'index < 0' cannot be true because it was
already earlier and handled with an early return.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Since 'foo' and 'bar' are not being multiplied...
Done automatically with the sed script:
/^[a-zA-Z][a-zA-Z]*.*(/{
s/ \* \([a-zA-Z]\)/ *\1/g
}
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The information about the File, Line and Function name that were included
in the assertion message are already present from the 'wwarning' macro, so
it is not necessary to include them in the macro's message.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In many places of the code these functions are counting on the "return"
effect of these macros to gracefully handle incorrect arguments. So, when
debug is not enabled, if it is okay to not display a message it is however
not good to completely skip the check and skip the early return.
This patch changes the macro to always perform the check and return to
avoid crashes, displaying a message only when NDEBUG is not set.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by coverity, calling 'wexpandpath' with a path that contains
either '$()', '$(\0' or '$\0' would cause an undefined behaviour because
the 'buffer2' would be uninitialised.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity (#50226), the function getenv can return unreliable
data, so if a sensitive application makes uses of the function 'wgethomedir'
or 'wusergnusteppath' we'd better use the GNU function secure_getenv which
ignore environment variable when used in a known critical cases.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the type used to calculate the size to allocate was
not the right one. It now gets the compiler to deduce it from the variable
for which the memory is allocated.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the return value of the function 'wdefaultspathfordomain'
cannot be NULL, so it is not necessary to check for that.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, the macro RETRY does not behave as expected, as it
assumes that errno is cleared on successful 'fopen' call which is not the
case.
This patch removes the uses of the macro RETRY:
- fopen: with the appropriate check
- fread/fwrite: nothing because they do not set errno
- fclose: nothing because retrying is not recommended
and took the opportunity to add a little bit more information in the error
messages.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function wrealloc never fails, and allocates enough storage to store
the resulting string, so it is useless to check if strcat failed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity (#50074), despite the expected behaviour that
'wmalloc' should never return NULL, it may still happen if an abort handler
set by user (with wsetabort) does not call exit() as expected. In such
case we call exit ourself to be sure not to return NULL.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by Coverity, if the user does not a an entry in the password
file then the function would assume its home path to be "/" but still
continue and later try to check for user->pw_dir which would dereference
the NULL pointer.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original code did have a few weaknesses, including:
- possible buffer overflow, if the translation was too long;
- possible crash, if either instance or class is NULL but not both
Now the appropriate length is calculated (not counting on a margin) and the
string is generated with the available elements.
As a side note, the variable was renamed from 'tmp' to 'title' for clarity.
This was highlighted by cppcheck (thanks to David Maciejak) and by
Coverity (bug #50078).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
moving scr initialization after WApplication existing test
to prevent crashing in case wapp is null
Acked-by: Christophe CURIS <christophe.curis@free.fr>
It is not necessary to call them when the command is not executable.
Took opportunity to change parameter to XSync which is supposed to be a
'Bool' type from Xlib, which means either True or False.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This allow compiler to raise a warning in case a new value would be added
to the enum; it also change order in the check so that WMGetButtonSelected
will be called only once, only for the case of interest.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As part of the process, some #defines were turned into enums. Also, the *_COL
when needed.
This brings the code for colors in line with the code for textures, and
allows us to use them as array indices to improve readability, e.g.,
colorOptions[MTITLE_COL] instead of colorOptions[3].
The manpages for getstyle and setstyle have not been updated and were missing
some options. This patch brings them up to date.
In addition, the *TitleExtendSpace options were appearing twice in getstyle.c.
One of the unnecessary sets was removed.
WPrefs.app/Appearances.c contained code for an unfinished feature
in the appearances panel of WPrefs, a tab for setting the background.
Since this feature has now been implemented as part of the texture
tab (see commit c2aca1a ("WPrefs: Set workspace background")), there
is no reason to keep this code.
Previously, when an image could not be loaded by WPrefs when rendering textures
for the Appearances panel, a segfault would occur. This could happen, e.g., if
a user moved or deleted an image file without editing their ~/GNUstep/Defaults/
WindowMaker file.
This patch first checks if a texture contains an image, and if it does, it
checks to see if that image can be loaded. If it can't, a solid black texture
is loaded instead.
The patch also has the added benefit of combining some of the code used for
rendering both pixmap and textured gradient textures.
Result of 'malloc' is converted to a pointer of type 'unsigned char',
which is incompatible with sizeof operand type 'char'
Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>
It is really unlikely that the free(menu) was the correct thing, it is more
likely that it was a typo instead of the appropriate 'free(menu)':
It would be excessive behaviour to free the complete menu only because an
entry cannot be added, and even if it were the case that would not be the
right way to do it because it would leak all the entries already allocated
for it.
(This fixes Coverity #50154 and #50155)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch enables users to set the workspace background (WorkspaceBack) in the
Appearance Preferences section of WPrefs. It appears as a new item in the
popup menu in the Texture tab, in the list of textures below, and a preview
appears in the background of the preview panel on the left.
The RGB panel of the WINGs color panel lists the red, green, and blue values as
base 10 numbers. However, hexadecimal numbers are very common when dealing with
RGB colors. This patch adds two radio buttons at the bottom of the RGB panel
to allow users to choose their preferred number system.
For version 2 of the patch:
Based on Cristophe's suggestions regarding the new decimal/hexadecimal RGB color
feature introduced in commit 83d8ad6, this patch changes rgbState from an int
to an enum, converts some if statements to switches, and changes from atoi to
strtol for the base 10 string conversion.
In addition, the "hexadecimal" text was too wide for the default size of its
radio button for some fonts, so both radio buttons are now resized to fit the
width of the panel.
This patch allows users to edit the two fonts (SystemFont and BoldSystemFont)
used in WINGs widgets using WPrefs. Note that changes will not take effect for
WINGs applications which are currently running.
In commit b4cb488, wprogressindicator.c was removed. It was restored in commit
d435ea7, but the corresponding declarations in WINGs/WINGs.h were not. This
patch fixes this oversight.
A common feature in several desktop environments is the ability to bring up
a "run application" dialog via a keyboard shortcut (frequently Mod1+F2). This
feature has been available to Window Maker users only through the root menu.
This patch adds the ability for a user to set up a keyboard shortcut to
accomplish this, either via WPrefs or by editing ~/GNUstep/Defaults/WindowMaker,
e.g., with
RunKey = "Mod1+F2";
Code from the execCommand function in rootmenu.c has been copied almost directly
into the handleKeyPress function in misc.c to accomplish this.
Some strings have been changed in the source code, this patch is updating
the corresponding 'msgid' in the 'po' files so that the translated words
will be found, and will not be lost in case of an automatic update on the
file.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The icons were too tightly packed into their buttons and were not even
fully displayed, and the label of the checkbox could be split on 2 lines
unnecessarily.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The flags were stored as 'char' but it is more logical to use int which is
"the natural size for integer on the host" and may help future evolution
if new flags are added;
The array is created with an initial size consistent with the number of
element we feed it with in 'makeWindowFlagsArray', this avoids a bunch
of realloc;
Fixed the type conversion that was not right, as pointed by clang.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In the function 'wdefaultspathfordomain' there was a check to make sure the
generated path would fit in the allocated area, but this allocated area is
sized precisely to fit the path, so it cannot fail.
In the function 'getCurrentFileName' there were checks to make sure the
generated result string would fit in the allocated area, but this allocated
area is sized precisely to fit the path, so it cannot fail.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function would not behave correctly if called twice, and the
library may re-use de-allocated pointers if calling WINGs functions
after calling this function.
The goal is for our library to behave as cleanly as possible, so
that users may have a clearer hint in case of misuse on their side
to debug the problem.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As pointed by the commpiler, the variable 'syslog_prefix' is not used
in the code, so it is removed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Due to some old and recent patches, the msgid used for translating the
error messages were not in line with the source code, now they are.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Recent patches has introduced the ability to exit cleanly from the WINGs
library, but this introduced some side effects because a function is
registered with 'atexit' to save user config on exit, which may not work
anymore because WMReleaseApplication frees some stuff needed for that task.
This patch handles this so that both method works, in case user of the lib
would forget to call the clean exit function.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The library uses internally a cache of tables to convert image to different
depths, there is now an internal function 'r_destroy_conversion_tables' to
free them.
We rely on pkg-config first, because that is the traditional way to
query configs and because that's how the MagickWand API suggests to do.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Instead of calling 'MagickWandGenesis' every time an image is loaded, only
call it the first time, and similarly do not call 'MagickWandTerminus' if
the library has not been used before.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It looks like that the function 'MagickGetImagePixels' is part of
the "deprecate.h" header, at least in ImageMagick 6.8.8. As it is not
used in WRaster's code, it is safer to actually check for 'NewMagickWand'
that is being used and which is not yet deprecated.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Took the opportunity to expand the WM_LIB_CHECK macro because it is not
suited for the case where compilation flags can be queried but was meant
for the case when they have to be guessed
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
- changed header used for MagickWand to the one specified in the online
documentation
- changed a comparison to 'MagickTrue' because its value does not allow a
compiler optimisation that would be expected in present case
- fixed potential crash, if loading with Magick fails it could return a
pointer to a de-allocated area
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In order for users to use pkg-config to compile programs using libwraster,
wrlib.pc should be installed in /usr/lib/pkgconfig, which had not been
happening. This patch fixes this issue.
A number of files have been added to the Window Maker source tree but are not
included in the tarball produced by "make dist" because they have not been added
to EXTRA_DIST in the appropriate Makefile.am. They are:
- Dutch translation files (commit afc90)
- email-clients.txt (commit 385db)
- WPrefs source files removed from WPrefs_SOURCES (commit 21dfd)
This patch adds these missing files.
Only the 'msgid' was updated, to make sure gettext can still find
the old translation, but I am not fluent in Armenian and google
translate is not helpful here, so I can't do better for the
translated string.
Only the 'msgid' was updated, to make sure gettext can still find
the old translation, but I am not fluent in Dutch so I can't do
better for the translated string.
The 'msgid' were updated to ensure gettext will find them, and the
translation have been updated too to stay in line with the message
(within the limits of google translate's help, because I don't speak all
these languages).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Currently, there is no WUtil.pc file so that WINGs developers can use
pkg-config to compile programs using the WINGs utility library. This patch
fixes this, and such a file is created by make, just like WINGs.pc and
wrlib.pc.
In addition, the WUtil.pc file has been added to .gitignore and included in
the debian libwings-dev package.
The original code would not provide correctly sized images in some cases
of ratios on the original image and on the requested size.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch is adding support for google webp image format, if you
don't know it a quick recap is to say that according to their tests
they claim it is better than png and jpeg.
Follow the link below for some more details:
https://developers.google.com/speed/webp/
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
- remove header not needed anymore
- revert change for XPM to stay in line with the standard
- changed PNG to stuck to the standard
- increased GIF to the complete signature
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
load.c is used to identify the file format, I rebased some checks on
what is defined in the spec of the format.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
- re-add PPM to the list of supported format, because legacy apps may
expect to find it from the list;
- add the copyright notice related to the code that have been borrowed
from Netpbm;
- fixed a small error in PPM format check in the loading procedure;
- and a few minor code improvements
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
This is actually taken from patch:
wrlib: Added support for webp image
in which it is out of place as it is unrelated
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
1) according to that checks i was able to see that some netpbm support
type are missing (exactly: ascii graymap (PGM files) and pixmap (PPM)
and ascii/binary bitmap (PBM))
See the link below for more details.
http://en.wikipedia.org/wiki/Netpbm_format
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
WUtil's file misc.c was including the private header for WINGs, where it
should have been using WUtil's public header instead.
Took opportunity to include the appropriate copyright header.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch is used to add support for syslog messaging implemented in
WINGs lib directly, so available from the lib itself and wmaker too.
I believe it will in a first time help to get some logging info
centralized in one point, and in a second time maybe add some info
level messages like wmaker is starting, stopping, restarting and else.
For now, it's built by default when the syslog support is found, maybe
we could also disable it by default.
As reported by Shawn W Dunn, the configuration strings written by wmgenmenu
in $HOME/GNUstep/Defaults/WMRootMenu were not reflecting his installation
directories choices.
Fix this by writing strings composed with PKGDATADIR.
This patch is replacing XRandR naming to RandR,
as XRandR is the name of the tool used in X11 and RandR
is the technology WMaker wants to support.
I had to update the info panel too, when many features are
activated not all of them can be displayed properly.
Because the existing nl locale files are ancient and obsolete,
the user interface stays mainly in English with hardly Dutch text
showing up.
This is a complete translation in 6 parts from scratch, using the
current (pl)menus and Makefile generated .pot files. These files
provide a very good, and thoroughly checked user interface in the
Dutch language.
Adding me as a maintainer in related README files; removing previous
maintainer (e-mail bounced).
Removing unnecessary executable bit from pt.po on the fly.
As this is a feature that is quite handy, let's document it there so more
people can benefit from it, we'll probably gain from less patches that
mail software broke or that are hard to review, because not inlined.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This explicit \n caused a bad line split in the displayed text; without it
the natural line splitting keeps the text on 2 lines instead of 3.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Some labels were truncated because the frame containing them was a bit
short; now the frames on the left have been shrunk (they had too much
margin) so the frames on the right can be expanded to have the labels fully
visible.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by David Maciejak, when libXinerama was not found on the
system, the configure script still tried to search for associated
header and stopped all with unclear message.
The proper behaviour that was expected is to accept silently that
libXinerama may not be missing, and only stop compilation when the lib
is present but not its header.
EWMH 1.3 added various window types: dropdown_menu, popup_menu, tooltip,
notification, combo, and dnd. We may as well set appropriate defaults if
these types get set on a window that isn't override-redirect.
I'm not terribly attached to these defaults, except that the ones for
'notification' are what I set manually for xfce4-notifyd before deciding
to patch wmaker.
While we don't provide compositing ourselves, add-on compositors such as
xcompmgr need us to copy the _NET_WM_WINDOW_OPACITY property from the
client window to the frame window.
We can do this easily enough.
This patch removes the unused code in screen.c.
The variable "dock_head", is defined, but not used. When we removes
the variable, then we can see that a lot of code is not used! so we
can remove it :-)
This patch also removes some extra curly brackets and join the variable
definition and the initialization in one line for some variables.
The arguments for SetButtonAlphaImage() are not used, so can be removed to
avoid compiler warnings.
This patch removes these arguments, so the code in SetButtonAlphaImage(), not
used, can be removed too.
A label was truncated because the widget next to it was too close for
historical reasons; now that the frame is wide enough the Pop-up Button
widget at its left is pushed further apart so the label for the position
of the workspace name can be fully visible.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Functions placing windows (PlaceWindow - src/placement.c, wMaximizeWindow -
src/actions.c) need to calculate area differently than ones placing icons
(PlaceIcon - src/placement.c, wArrangeIcons - src/actions.c).
So this patch puts adjustment code in wGetUsableAreaForHead which is
used by all those functions but called with different 'bool noicons'
argument depending on if it's called in window or icon placement
function.
Reported-by: Carlos R. Mafra <crmafra@gmail.com>
Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>
Although this is a modification of the public API, this patch does not
change the binary interface, and the attribute does not add a constraint
on the parameter for the caller so we do not break the source compatibility
either.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original code supposed that the library was available and ready to use.
As it may not be the case (modern distro generally don't include dev files
by defaults), add a the appropriate checks to make sure WRaster compilation
will not fail later.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not have
a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- checks also for header usability with the compiler;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The usual way to define a macro in is to name macro with 'USE_xxx' when
they are used to enable a feature 'xxx'.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not have
a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- checks also for header usability with the compiler;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The usual way to define a macro in is to name macro with 'USE_xxx' when
they are used to enable a feature 'xxx'.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not have
a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- checks also for header usability with the compiler;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support)
The feature is also no more disabled by default, as it is something user
would expect from a modern window manager.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The usual way to define a macro in is to name macro with 'USE_xxx' when
they are used to enable a feature 'xxx'
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- checks also for header usability with the compiler;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support)
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
There are a number of steps in the library check procedure that are a bit
repetitive, and have been placed into the macro WM_LIB_CHECK to make the
code simpler to write.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function can have a wider use, so moved then to a more neutral place
and gave them more generic names.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The usual way to define a macro in is to name macro with 'USE_xxx' when
they are used to enable a feature 'xxx'
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The default case should not arise because the tested variable is an enum,
but it the case someone would get a value wrong, we're safer with a little
warning and proper fallback than with unknown behaviour.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The other extension detection have the same situation and they already
have a variable for that, do not do differently for Xrandr because it adds
unnecessary complexity in the code.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When autoconf encounters a macro name that it does not know, it assumes
that it is just a standard keyword that needs to end up in the generated
configure script.
This patch teaches m4 about the syntax used for our macros so it can now
that it is supposed to be a macro, and then it can stop and report the
problem.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The macro 'AC_ISC_POSIX' is now said to be deprecated, so we use the new
recommended check instead as we make light use of the corresponding
function.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This attribute reports cases where GCC has to generate special code to
handle some C constructs, for which we'd prefer an explicit code instead
of counting on GCC doing the work.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
In the present case, it just led to a compiler warning on unused variable
when the XShape extension is disabled, but in general case it just adds
complexity by duplicating things; this should be reserved for complex
cases.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As suggested by GCC, this function is a good candidate. There is a little
constraint on how to do it however, as it is part of the public API so we
have to do it in a portable fashion (We can't rely on our "config.h" when
the file will have been installed).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This attribute helps the compiler checking the type matching between
arguments and the specification in a printf-like format string, to
avoid invalid output. This attribute is optional, but some compilers
can suggest functions that could have it.
This patch adds the appropriate compiler flags if they are supported when
the source is being compiled with DEBUG enabled.
The patch also introduces a new macro WM_CFLAGS_CHECK_FIRST because in some
cases AX_CFLAGS_GCC_OPTION is not really efficient and in present case it
does not fits the job.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
A previous commit broke the behaviour of the left and right arrow keys
in the switchpanel. Releasing either of the keys would correctly select
a new window but the panel would then close. The original, and desired,
behaviour is for the panel to remain open after selecting a new window
with the arrow keys.
Reported by Yury Tarasievich.
Fixes bug report:
The 0.80.2 and older versions allowed the miniwindows to lay down in
parallel with wmdock (if one wanted it that way) but newer Windowmaker
versions could only partially simulate such behavior and that includes
to either setting an option NoWindowOverDock to yes or to manually set
the dock to "Keep on Top".
Reported-by: Josip Deanovic <djosip+news@linuxpages.net>
Signed-off-by: Amadeusz Sławiński <amade@asmblr.net>
It is not only not very efficient, but in present case it also participates
in memory fragmentation.
This patch replaces this with a stack allocated buffer with a buffer which
is way too large.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As this label is being displayed in the window, it is a good idea to make
it translatable to the user's locale.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The parameters for the textures were stored all together in an array
which made its usage error prone; now there a struct to clearly identify
which string is what, so it is clear in the source what's being done.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The parameters for the theme colors were stored all together in an array
which made its usage error prone; now there a struct to clearly identify
which string is what, so it the source is clearer on what's being done.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The idea is that an array is easier to work with, when it's about to add,
remove or change entries, because all data end up stored in one place
instead of dispatched around the code.
It also makes code smaller as it avoids repetitions.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When animations are disabled, we still create the function but we make it
empty to keep the rest of the code simple. This patch does:
- mark the function inline, to increase the probability that the compiler
will not generate the function at all;
- mark arguments as unused to avoid some compilation warnings.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Allocating memory with 'malloc' has a cost and participate to memory
fragmentation, so for a temporary buffer that has a fixed size let's
prefer allocating it on the stack.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The check on length of string before comparing is not necessary
because this will be checked as part of strcmp; the check won't
save time and may actually cost.
As the number of element in the array is not going to change during
the loop, took the call to 'WMGetArrayItemCount' outside the loop
to be faster (and ease compiler's optimisation work).
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The idea is to:
- have a consistent looking file by using autoconf macros
- provide better feedback on improper option usage
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Some compiler support C11 standard by default, some need an explicit
option, and some don't support at all; this new macro only tries to enable
support if possible, the actual feature support being done by other checks
on case-by-case
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The previous code made a dumb change to CFLAGS which would cause user
supplied CFLAGS to be ignored, and possible compiler compatibility issues.
The new code does the same changes in a smoother way, so we do not totally
drop user specified or script detected compiler options.
When debug is enabled, we also include a reminder for the final CFLAGS
value.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The new macro 'wlengthof' from WUtil makes code easier to read than the
previous [sizeof() / sizeof([0]) ] construct.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The new macro 'wlengthof' from WUtil makes code easier to read than the
previous [sizeof() / sizeof([0]) ] construct.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The new macro 'wlengthof' from WUtil makes code easier to read than the
previous [sizeof() / sizeof([0]) ] construct.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The new macro 'wlengthof' from WUtil makes code easier to read than the
previous [sizeof() / sizeof([0]) ] construct.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The new macro 'wlengthof' returns the number of elements for which a local
array have been defined, which makes code easier to read than the
previous [sizeof() / sizeof([0]) ] construct.
The macro includes a static assertion to stop compilation if it is being
used on a pointer, for which we cannot know the size of the array, to
avoid generating dummy result. This can work only with C11 which
standardised the static assertions.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This patch makes all files appear together when listing files and
it is more clear about their function.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The two functions require an argument which is used when the libXpm is
used, which means having the same argument list for the builtin code to be
permutable.
This patch adds the appropriate stuff to avoid a false report from the
compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Now that the builtin XPM loading is separated from the saving code, the
use of the conditionals provided by autotools provides a better result
as it avoids unnecessary compilation.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Big files are harder to work with, so split load and save, which
happen to already be independent anyway.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Even if there is nothing to do for PPM at current time (all builtin
code), for consistency it is better placed with the other image
format checks.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support);
- makes uses of shell functions to keep generated configure smaller by
sharing reusable stuff;
- uses an automake conditional to avoid compiling the file is support is
not enabled
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support);
- makes uses of shell functions to keep generated configure smaller by
sharing reusable stuff;
- uses an automake conditional to avoid compiling the file is support is
not enabled
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support);
- makes uses of shell functions to keep generated configure smaller by
sharing reusable stuff;
- uses an automake conditional to avoid compiling the file is support is
not enabled
It includes a typo fixed by Amadeusz S-B³awiñski.-A
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support);
- makes uses of shell functions to keep generated configure smaller by
sharing reusable stuff;
- uses an automake conditional to avoid compiling the file is support is
not enabled
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function is a callback, which means having a fixed argument list.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The function did not use the argument 'mPtr', so removed it.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
As reported by Nicolas (nhs), compilation of wrlib is broken when
switching to gif_lib v5.
The API of gif_lib has known little change to provide thread-safe usage,
so we now detect this in configure and and use the functions as appropriate
in gif.c
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The use of the conditionals provided by autotools provides a better result
as they avoid unnecessary compilation.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The original check was not compliant with autoconf's syntax, did not
have a very good behaviour for user and was not easy to make evolve.
The new macro:
- uses as much as possible autoconf macros for portability and code
consistency;
- provides a consistent behaviour on yes/no/auto (if user explicitly
enables support, do not silently disable if not found; if library is found
but not the header, complain to let user install it or explicitly disable
support);
- makes uses of shell functions to keep generated configure smaller by
sharing reusable stuff;
- introduces a tracking of unsupported things to be reported to user
It includes a fixes from a problem found by Amadeusz S-B³awiñski, because-A
the function 'wm_fn_imgfmt_try_link' modified the variable LDFLAGS
instead of LIBS.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The previous rule was to ignore the whole directory, while still having
some files tracked inside.
Now we only explicitly ignore the files that have been placed there by the
autotools, so we can have a better view of it in Git when adding stuff.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
A recent patch removed the newlines from the default English legal text, as the
legal panel automatically wraps text so the newlines were unnecessary. In this
patch, they are removed from all .po files which contain a translation of the
legal text.
This patch uses two variables to set the width and the height for the
Info Panel. It also includes a margin space.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
This patch uses two variables to set the width and the height for the
Info Panel.
Now the panel has the same size than the window.
Finally, this patch moves the center definition to the beginning, and
removes the curly brackets.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The label containing the legal information in the legal panel was set to wrap.
However, there were redundant newlines in the text itself. In wider fonts, this
had the effect of the legal text not completely fitting in the legal panel.
This patch removes the unnecessary newlines.
In keeping with the spirit of free software, this patch removes several
Microsoft fonts (Trebuchet, Arial, and Verdana) from being the default fonts for
Window Maker and WINGs.
This patch has the added advantage of allowing users some control over the
appearance of WINGs widgets. Formerly, there was no way to control the font if
the Trebuchet MS font was installed. Now the WINGs font can be changed by
setting the sans-serif alias in fonts.conf.
These two files provides configuration panels that were never enabled
because they have not yet reached a usable state.
With this patch, we do not compile them anymore to avoid useless stuff in
the program.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The functions that create the different configuration panels were taking
the screen structure as argument, but it turns out that none of them
actually need it.
We just remove the argument to make code simpler and easier to maintain.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The screen argument was not used in 'InitDoubleTest', 'InitEditMenuItem'
and 'InitEditMenu', so remove it to make the code simpler and avoid
compiler warning.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The x coordinate is not needed by the function because menu are a vertical
stacking on items, which means only y coordinate matters.
This patch removes the x parameter to fix a compilation warning.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The event that triggered the function call is not used by the function,
so the parameter can be removed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Now that the code have been cleaned up on that matter, it is a good
idea to re-enable warning on these cases because they help keeping
the code as simple and maintainable as possible.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The timers proposed by WINGs toolkit are handled with a callback function,
which means having a fixed argument list for that application function.
It is then correct to not use the argument, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The WINGs toolkit dispatch events on widgets using callbacks, which means
having a fixed argument list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When the process receive a signal, a callback function is used, which
means having a fixed argument list for that application function.
It is then correct to not use the argument, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The WINGs toolkit dispatch events on widgets using callbacks, which means
having a fixed argument list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When an error occurs in X, the Xlib is using a callback mechanism to
execute application code to handle the problem, which means having a
fixed argument list for that application function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The mechanism of Notifications in the WINGs toolkit is relying on
callbacks to dispatch notifications, which means having a fixed argument
list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When an error occurs in X, the Xlib is using a callback mechanism to
execute application code to handle the problem, which means having a
fixed argument list for that application function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
If a window handles a ConfigureRequest which did not specify a move or
resize operation we should not call wWindowConfigure() and save the
window geometry.
Sergey Popov reported a scenario in which the old behaviour caused a bug:
* Start gvim with a server.
"gvim --servername qwe .gimvrc"
* Maximize gvim.
* (Re)open a file in the same window
"gvim --servername qwe --remote-silent .gvimrc"
* Now the window claims to be unmaximized and its old geometry is
forgotten.
The bug was that when the gvim window reread the file it generated a
ConfigureRequest without specifying a geometry change but we called
wWindowConfigure() and saved its geometry as though it had been
maximized.
All the non-Window Maker-specific icons have been removed from the default
configuration in WMWindowAttributes.
In addition, the icon for drawers has been added to WMWindowAttributes. Also,
it has been changed from a .png file to a .tiff and .xpm to bring it in line
with the other provided icons.
Update: in the original version of this patch, I neglected to modify the
appropriate Makefile.am to account for the drawer icon changes.
There are a few function in WING's API that take parameter for consistency
reason and for possible future evolution, but actually do not need the
argument.
As they are case we know about, this patch adds the appropriate stuff to
tell the compiler we are ok with this to avoid a false report.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As callback have a fixed prototype, it is correct to not use all the
arguments, so this patch adds the appropriate stuff to avoid a false
report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The WINGs toolkit dispatch events on widgets using callbacks, which means
having a fixed argument list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When an error occurs in X, the Xlib is using a callback mechanism to
execute application code to handle the problem, which means having a
fixed argument list for that application function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
WINGs dispatches window resize events using callback functions, which
means having a fixed argument list for that function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The drag-n-drop mechanism is managed by WINGs through callbacks, which
means having a fixed argument list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The mechanism of Notifications in the WINGs toolkit is relying on
callbacks to dispatch notifications, which means having a fixed argument
list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
There is a function in WUtil's API that take a parameter for consistency
reason, but actually does not need the argument.
As it is a case we know about, this patch adds the appropriate stuff to
tell the compiler we are ok with this to avoid a false report.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As callback have a fixed prototype, it can be correct to not use all the
arguments, so this patch adds the appropriate stuff to avoid a false
report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The code to update the "Other maximization" menu's labels was moved from
makeMaximizeMenu() to updateMaximizeMenu(), making the WMenuEntry
pointer in the former function redundant.
Ensure that keyboard shortcut labels for other entries in the window
menu are updated if the user changes the shortcut. Previously they were
only set when the menu was created and could get out of sync.
The shortcut labels for items in the "Other maximization" menu were set
at startup and not updated if the user changed the shortcut key. Thus
the labels shown could be wrong. They might even be missing if no
shortcut was assigned at startup but was subsequently set during the
session.
We now ensure that the shortcut labels are updated with the menu
whenever preferences are reloaded.
The window menu Unmaximize entry had the Maximize shortcut key as its
label. That's because the Maximize and Unmaximize menu options are in
fact the same single entry with different text depending on the window's
state.
It can, however, cause confusion if a window is maximized using one
of the "Other maximization" options such as Maximus. Selecting the
Unmaximize entry from the window menu would indeed unmaximize the window
but pressing the listed shortcut key would not.
We now dynamically update the shortcut label so that it shows a key
which will actually unmaximize the window. Thus the menu description
and shortcut action are now consistent.
Windows which enter fullscreen mode were not automatically given focus.
Usually that didn't matter because they already had focus when they
switched modes. An example of unexpected behaviour is opening a media
file in an already-running vlc from the commandline or via a file manager.
vlc would fullscreen mode but the launching application would retain focus.
Note that if vlc were not already running and it was launched as
described above, it would receive focus when it was opened and thus
retain focus going into fullscreen.
We now track which window had focus before a window enters fullscreen
mode and focus the original window afterwards. In the (usual) case
where the window going fullscreen already had focus, nothing changes.
In the rarer case where the window going fullscreen didn't
have focus, it will gain focus temporarily then yield to the originally
focussed window when it leaves fullscreen mode.
To reproduce:
* Launch vlc and configure it to switch to fullscreen when playing a
movie and to disallow multiple instances.
* Switch to a terminal and type 'vlc /media/funny_cats.mp4' or use a
file manager to open funny_cats.mp4 with vlc.
* Press space to pause the movie. Nothing happens because the
terminal/file manager still has focus.
There are a number of window maximization options which are currently
only available to the user via keyboard shortcut:
* maximize vertically/horizontally
* maximize left/right/top/bottom half
* maximize left top/right top/left bottom/right bottom corner
* maximus
In this patch, they become available through the window menu, under
the "Other maximization options" submenu.
When the process receive a signal, a callback function is used, which
means having a fixed argument list for that application function.
It is then correct to not use the argument, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When a sub-process terminate, the function to process that event is
defined using a callback mechanism, which means having a fixed argument
list for that application function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It is dangerous to use a function that does not use the same prototype as
expected by the callback, because that mean there is conversion performed
for the arguments, on which the compiler has no possibility to report
problems.
It is safer to create the function with the strict argument list, and
insert an explicit type conversion for which the compiler will be able
to perform compatibility checks, and include optional code when needed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When an error occurs in X, the Xlib is using a callback mechanism to
execute application code to handle the problem, which means having a
fixed argument list for that application function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The WINGs toolkit dispatch events on widgets using callbacks, which means
having a fixed argument list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It is dangerous to use a function that does not use the same prototype as
expected by the callback, because that mean there is conversion performed
for the arguments, on which the compiler has no possibility to report
problems.
It is safer to create the function with the strict argument list, and
insert an explicit type conversion for which the compiler will be able
to perform compatibility checks, and include optional code when needed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The mechanism of Notifications in the WINGs toolkit is relying on
callbacks to dispatch notifications, which means having a fixed argument
list for the handling function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The toolkit dispatches window close request events using callback
functions, which means having a fixed argument list for that function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The toolkit dispatches X events for MouseButton using callback functions,
which means having a fixed argument list for that function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The custom drawing code for windows is handled using callback functions,
which means having a fixed argument list for that function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
When an entry of a menu is selected, the appropriate action is triggered
using a callback, which means having a fixed argument list for that
function.
It is then correct to not use all the arguments, so this patch adds the
appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
To have an easy-to-maintain code, the configuration loading function is
using a lot off callbacks, which means having a fixed argument list.
It is then correct to not use all of them in all case, so this patch adds
the appropriate stuff to avoid a false report from compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
There are a few cases in the code where conditional compiling can
lead, when the corresponding block is not activated, to some arguments
not being used.
This patch adds the proper stuff to avoid a false report from the
compiler.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As a callback function has a fixed prototype, it is not an error to have
some arguments unused, but we need to let the compiler about it to avoid
spurious messages that would hide real cases.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The C language provides the macro 'sizeof' to handle this kind of
situation, so do not create locally a macro which can become a source
of problems.
Took opportunity to change the size of buffer to follow guidelines
from Inotify's manpage, and to remove the initial value that is useless.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
If the triggering condition occurred, the file handle for Inotify were
closed, but as the variable containing the handle was not updated it
could lead to multiple call to file close, which behaviour may be
problematic.
The file handle was passed as a parameter, which does not allow for a
clean value change, so now we use the variable directly as it is
available in the global namespace.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The XDND support is not enabled by default, and not simple to
enable (does not use the appropriate configure flag method).
With previous global namespace related patches, this file not being
enabled a change was missed. Now at least it compiles, although it
is still an experimental feature.
Instead of defining 2 separate data arrays, use a array of struct
to make it less bug prone (no dependancy on order of elements).
Took opportunity to de-CamelCase the variable name;
Took opportunity to add appropriate const qualifier.
Instead of defining 2 separate data arrays, use a array of struct
to make it less bug prone (no dependancy on order of elements).
Took opportunity to de-CamelCase the variable name;
Took opportunity to add appropriate const qualifier.
The default domains were originally defined in different global
variables in C files; This patches groups them in a single
structure placed in global namespace.
The variable keymove_tick is set, but is never used after, so can
be removed.
kix@osaka:~/src/wmaker/git/wmaker-crm/src$ grep keymove_tick *[ch]
moveres.c: scr->keymove_tick = 0;
screen.h: int keymove_tick;
kix@osaka:~/src/wmaker/git/wmaker-crm/src$
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
Setting a clip as autoattracting will disable an autoattracting drawer,
and reciprocally. Although not immediately obvious, it should be fairly
easy to figure out, and is totally reversible. So we'll probably never
opt to show that confirmation dialog box.
toggleAutoAttractCallback needs to check whether there are clips before iterating
through them (or it crashes WM!). The extra test is needed because now drawers can
auto-attract as well. Before drawers, only clips could auto-attract, so testing
wPreferences.flags.noclip was pointless.
The variable wapp_list is never set, then the comparison with
wapp is only true if wapp is NULL. wapp can not be NULL in the
function wApplicationDestroy() because is tested previously.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The variable session_state, defined in screen.h (WScreen), is moved to
the global variable w_global.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
This patch gets rid of the warning:
switchpanel.c: In function 'wSwitchPanelSelectFirst':
switchpanel.c:673:18: warning: variable 'tmpwin' set but not used [-Wunused-but-set-variable]
by using the body of the WM_ITERATE_ARRAY directly and avoiding the temporary
variable 'tmpwin' which ends up being used only on the LHS inside the macro.
Because the wClipIconPaint() function is specific to paint the clip,
it knows where is the clip (wks_info.clip_icon), so the argument is not
needed.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The appicon list is moved out of WScreen. The reason is because
the appicon list is used to create and remove icons (appicons, dock,
clip,...) on the screen, but these icons are not associated with the
WScreen. These icons are associated with the Workspace.
If we check the WWorkspace struct we can see that the Clip is inside
this struct. The dock, the background, the workspace name are other
items related to the Workspace, not with the screen.
So, we should take out the appicon from the WScreen. But, what is the
better place to hold it? The workspace? NO!, because the icon list
is common to the all workspaces. Probably the better place is hold
as independent list in the global namespace, so this is my option.
In the next patches we can see that this is the better option, because
we can create the icons, without think where we should paint them.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The WScreen arguments in the functions wWorkspaceMenuUpdate() and
makeMakeShortcutMenu() is not used now, so can be removed.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The clip, dock, workspaces and drawers menus are common for all
workspaces, so they should be included in the global namespace
instead of the screen struct.
The variable workspace_menu, that contains the workspace_menu
is moved to the global workspace properties.
Now the screen is not needed to know the workspace_menu.
This variable is moved to the global workspace struct because it is
related to the workspace system, and not to the screen.
The variable workspaces, that contains the list of workspaces
is moved to the global workspace properties.
Now the screen is not needed to know the workspaces.
The function getWindowState() doesn't need the WScreen argument.
The variable current_workspace, that contains the current workspace
is moved to the global workspace properties.
Now the screen is not needed to know the workspace_count.
The variable last_workspace, that contains the last used workspace
is moved to the global workspace properties.
Now the screen is not needed to know the workspace_count.
The variable name is changed to workspace.last_used because a similar
variable name is also in the WApplication struct.
The variable workspace_count, that contains the number of workspaces
is moved to the global workspace properties.
Now the screen is not needed to know the workspace_count.
The new workspace struct includes the global information
for the workspaces. All information related with the workspaces
should be included here.
The first variable moved is workspace_name_font, included in this
patch. This variable was included in the screen info, but the
screen and the font used in the workspace name don't have any
relationship.
This patch changes the keybinding check in cycling keyrelease.
Now, the variable binding contains the keypressed, so we can
check if the key pressed is the same than the keybinding. If
the keybinding is different (user press other key) then finish.
Without the loop, the code is faster.
Then, the keybinding variable is not used anymore, and can be removed.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The function StartWindozeCycle() uses now the variable WShortKey
to store the shortcut key.
The code is the same, but now the hasModifier variable is calculated
in only one place.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
We were using wwin as the temporary variable in WM_ITERATE_ARRAY() but
wwin was already set to the result of WMGetFromArray() and was due to be
returned from the function. As a result of this oversight, wwin was
always set to NULL.
The bug could cause a problem in the - admittedly rare - case where all
windows were minimised at startup, causing the first invocation of the
switchpanel to highlight the first minimised window but then fail to
raise it.
Reported by Yury Tarasievich.
This patch removes the error variable, because is not changed
after their initialization.
This patch also removes the if check to wfree the path variable,
because this variable never is NULL.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
This patch swaps the if and the else blocks for the
image value. Now, if the image doesn't exist, and because
error is always "0" (is not changed after the definition),
we can exist from the wIconChangeImageFile function.
Before this patch, if the image was NULL, the error variable
was set to "1", so we must return !error, then return "0".
Because the "path" variable is not NULL, we must wfree it.
Now, the "Set the new image" block is moved out the if block.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
This patch swaps the "if (path)" block to "if (!path)"
therefore, the true block is now "error = 1" and the
false block is the call to get_rimage_from_file function.
The second change is that the "error = 1" statement is
changed to "return 0", because if "error" is "1", then
only the "return !error" statement is used, then we can
change it for "return !1" that is "return 0".
Then, we can move the get_rimage_from_file call out of the
if block.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
A default cursor would make sense if we were talking about an API,
where we would propose a 'I-don't-know' choice to devs, but here
we're talking about WindowMaker's internal, where we're supposed
to provide a consistent interface to the user, so picking a
cursor is not an option (and 'Normal' is actually a decent
looking choice anyway).
Suggested-by: Rodolfo García Peñas <kix@kix.es>
As the cursor choice for each action is actually a user choice,
it is logical to put this into the structure, instead of a poorly
defined 'extern' in every file.
Having all the global variables at the same place will provide
better visibility on the code; grouping them in a structure
reduces the risk for name clash; it also offer the possibility
to organise them.
Another BIG benefit is that, when reading the code, an access to
a global variable will now be clearly visible, and distinguished
from a local variable use.
The name will be a problem with next change because it uses a too
common word which will be useful at a better place.
Took the opportunity to rewitre the check in the human logical
thinking order to make it easier to read.
This patch includes the function "NumLockMask", that checks if the
Alt Modifier key is "NumLock". Then we can use this function in the
calls to Capture the key sequence and don't use this modifier.
This patch includes the changes for the new debian package 0.95.5-1:
Changes are:
- New libwutil library, new version 3:
- Updated debian/control file, replacing libwutil2 with libwutil3.
- Files moved in debian folder:
- libwutil2.changelog-upstream -> libwutil3.changelog-upstream
- libwutil2.install -> libwutil3.install
- Removed file libwutil2.symbols
- New file libwutil3.symbols
- libtiff migrated from version 4 to version 5:
* "Build-Depends: libtiff5-dev" in packages wmaker and libwraster3-dev,
since libtiff-dev introduces dependency to libtiff4 which is in oldlibs.
- Small changes:
* Included the word "WindowMaker" in the wmaker package description,
to found it easily.
* When compiled on DragonFly+gcc 4.7.3, this out of bounds array element
initialization causes an important variable to be overwritten and a
subsequent WindowMaker crash
* It was sheer luck other compilers/environments didn't exhibit any obvious
issue so far
The flag "supports_tiff" doesn't belong to the wScreen, because all
screens has (or not) the same value. If wmaker supports (or not) tiff
files is global to all wmaker, not to the screens.
The optimize_for_speed was used to set the Alpha channel for jpeg.
The alpha channel for jpeg should be used always. The current CPUs/GPUs can do
it without problems and this is an old behaviour.
This patch removes some extra curly brackets, some empty lines,
extra spaces,...
This patch is not removing the goto calls. I am not sure if the code
is better without gotos.
This patch removes some extra curly brackets, some empty lines,
extra spaces,...
This source file had some empty if blocks. Now are removed.
This patch also changes some preprocessor macros for C code.
When an application steals focus and switches workspaces autonomously,
"ignore workspace change" does not get reset. Since I didn't want to
track down the exact case I thought it best to reset the flag in all
cases, as it probably should be.
> How did you discover this? What was the behavior you saw?
After saving files in avidemux, the "Done" dialog pops up, and with it the
entire application is moved to the current workspace. After this, using the
mousewheel on the root window or using the appropriate key combination does
not change the workspace anymore, only using the arrows on the clip works.
And, as mentioned on the mailing list, just starting emacs (22.3.1) also
provokes this bug. Using a Debian/Sid system btw.
modified: WPrefs.app/po/Makefile.am
Ommited .pot messages in Docks.c were recovered.
modified: WPrefs.app/po/ko.po
Update Korean Localization of WPrefs.app to 0.95.4
modified: po/ko.po
Update Korean Localization of WindowMaker to 0.95.4
The original type was not as safe as it looked like to be, and led
to more memory usage than necessary.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It is bad practice to define the prototype of an external function
inside a source file that wants to use it.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
It is dangerous to define an extern variable locally because the
compiler is not able to ensure proper type definition, but including
the header that defines it allows a safer situation.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The whole changeImage() function in switchpanel.c is a no-op if
wPreferences.swtileImage is set to None because the panel isn't actually
drawn in that case. As a consequence there are no images to change.
As well as being logically incorrect the existing code causes a segfault
if the user has the SwitchPanelImages preference set to None because
changeImage() would attempt to access the icons and images arrays which
are only initialised in wInitSwitchPanel() when swtileImage has a value.
Bug report and diagnosis by Juan Giordana, Gabriel Vlasiu and Christophe
Curis.
There is a dedicaded floating point modulo operator that does a
better job (faster, no precision loss), so we use it instead of the
complicated formula that uses type conversions.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
For this case we happen to already have the same variable as integer
type, so it is better to use them directly for computation.
Please note that the floating point variables do not contain more
precision in this use case, otherwise the change would not be
transparent.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Compilation will not work if the header is missing, so make sure
it is present and compilable first. This can highligh sooner if
a user forgot to install the corresponding dev package.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Took the opportunity to rewrite the check using autoconf macros
to generate a more compatible configure script.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The function contained an internal statement that would have
discarded the const attribute, but as this statement is actually
useless it have been simply removed.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Because we have to use some X functions marked as deprecated to
ensure portability, we prefer to hide the warnings from standard
users, and keep them only for dev team.
[crm: Add -Wno-deprecated-declarations too]
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
To ensure proper behaviour with X servers that do not support the
Xkb extension, implemented call to the fall-back legacy function
to the other places where it is being used, to ensure proper
behaviour in any case.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The original code was assuming that the Xkb extension is available
everywhere, which is not the case. This resulted in all modifiers
to be seen as NoSymbol, thus the empty list.
Now we initialize properly the Xkb extension at the start of the
program, and provide a work-around with the historical function in
the case the ext would not be available.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Apparently, there is some exceptional case when the 'Mouse Grab
Modifier' list can be empty, in which case the program would
crash.
This patch makes sure the program will not crash (that's a bad
behaviour), but does not solve the empty list case
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by clang, the format used was %hi which is expecting
a short-typed argument, but the parameters are actually defined
as ints in the structure, so use only %i.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
As reported by Iain Patterson, the clang compiler is (by default)
strictier on having clean C syntax. A few structure definition
did not comply, now they do.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
Using an unspecified prototype function for a callback is calling
for problem. Added an explicit prototype for the 2 callback
functions in WDefaultEntry, and fixed the functions used to match
it. The new code should be a lot safer.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
This attribute for functions was not standard until recently, so
there are a few different possible syntax. With this patch, the
configure script will search for the proper syntax, and define what
is needed to have the attribute compile correctly.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
If the function was called more than once with different usernames
it would always return the path for the user on the first call,
which is not what would be expected.
Furthermore, if the function succeeds it allocated memory to save
this path but it was never freed.
The good thing is that the use case for this function is so rare
that it is improbable it was ever called, which explains why it
was never seen.
The new code always behaves as expected, and does not allocate
memory anymore to avoid wasting time and memory for such small
things, which is acceptable because this function is local.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
The previous method was to use a custom install procedure, which
is a bit complex when automake can handle this for us, and as a
side effect it made 'distcheck' fail on uninstall procedure check.
The new method is to use a simple conditional and autoconf/automake
will do all the work for us.
Signed-off-by: Christophe CURIS <christophe.curis@free.fr>
wMaximizeWindow() was previously rewritten and simplified.
Unfortunately the rewrite introduced a regression with windows which
were horizontally maximized or Maximusized, whereby they would be too
small after the maximization.
The bug is easy to see by opening a standard 80x24 terminal and hitting
the maximize horizontally shortcut. The terminal shrinks to 21 lines.
Pressing the shortcut key again correctly restores it to 80x24.
The fix is to refrain from shrinking the final window height, which is
done to account for its frame border and titlebar, for the affected
maximization styles.
Explicitly match a Return keypress in the switchpanel, so that the
currently selected window is focused without the keypress event being
handled elsewhere.
Pressing Return to commit the window selection is intuitive but other
shortcuts may bind Return. Any keypress not explicitly handled in the
switchpanel code will be passed through to other shortcut processing.
If, for example, alt+return were mapped to Select Window and the
switchpanel is open, pressing return while holding alt would both commit
the window change AND select the window, which is not intuitive.
If the new preference SwitchPanelOnlyOpen is set to YES, pressing
alt-tab or similar shortcut will only open the panel without
automatically switching to the next window. Shortcuts will continue to
function normally once the panel is open.
Users can enable the new preference if they find themselves regularly
opening the switchpanel just to visualise which windows are open, or to
initiate same-class cycling.
The default value of the new preference is NO. Behaviour is unchanged
unless the preference is explicitly enabled with the following command:
$ wdwrite WindowMaker SwitchPanelOnlyOpen YES
Keep an array of flags to remember if icons are supposed to be selected
and/or dimmed, and don't redraw any which are already in the correct
state.
Originally I didn't think the benefit of doing this was worth the effort
but future commits will call changeImage() more often so cutting down on
unnecessary redraws suddenly became more useful.
The Workspaces entry in the main menu allows to rename workspaces by
clicking on the workspace name while pressing the Ctrl key.
However since commit 63219247c6 ("Added shortcut to switch to last used workspace")
there is one more entry before the workspace name list, and that leads to
picking a wrong name to rename -- clicking on the first workspace asks to
rename the second workspace and so forth. Trying to rename the last workspace
leads to a segfault.
This happens because there is an explicit offset in the current code (- 2)
to account for the "non-workspace" entries in the menu. If one adds one
more entry that offset should be changed too (this is a prime example
of fragile code leading to bugs).
The file funcs.h is removed. A new file osdep.h is created to hold
the definition for all osdep_*c files.
The files .c has been adjusted to include the right header files,
removing funcs.h, including osdep.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file switchmenu.h includes the function prototypes for switchmenu.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file winmenu.h includes the function prototypes for winmenu.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file monitor.h includes the function prototypes for monitor.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file event.h includes the function prototypes for event.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file cycling.h includes the function prototypes for cycling.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
The file startup.h includes the function prototypes for startup.c.
The prototypes included were moved from funcs.h.
Signed-off-by: Rodolfo García Peñas (kix) <kix@kix.es>
If one or more icons were dimmed in the switchpanel because the user
used GroupNext/PrevKey, then an icon was selected with the mouse or the
Home or End keys, dimmed icons remained dim. That could be unintuitive
if the selected window was of a different class.
Instead we now always redraw all icons when highlighting a different
icon.
When cycling through windows in the switchpanel using the GroupNextKey
or GroupPrevKey shortcuts, dim the icons of windows which are of a
different WM_CLASS.
A while loop in StartWindozeCycle() was checking the value of the panel
pointer and setting a flag to break out of the loop if it were NULL.
The current iteration of the loop was allowed to continue, however, with
the result that the null pointer could be passed to one of the
switchpanel functions and cause a segfault.
To reproduce, close all windows except one. Open the inspector and set
the window's "Do not show in the switch panel" flag. Then close the
inspector and press alt-tab to open the switchpanel. As there is only
one window and it is not allowed to appear in the switchpanel, a null
panel pointer is returned, then later passed to wSwitchPanelSelectNext()
causing wmaker to crash.
The fix is to break out of the loop immediately instead of setting the
done flag.
new window positions top left, top right, bottom left and bottom right
these new maximized positions are combinations of left, right, top and
bottom maximized positions
if a window was at x=0 or y=0, the original condition return
false and the window is not moved to original position,
but if width or height from old_geometry is set then wmaker
already saved old_geometry, and we can trust x=0 or y=0 is
original position.
- remove extern declaration in source file, use header instead
- add inclusion of header defining the functions of the file to
get the compiler to cross-check them
- marked static the functions that should not be visible ouside
their file
- remove extern declaration in source file, use header instead
- add inclusion of header defining the functions of the file to
get the compiler to cross-check them
- marked static the functions that should not be visible ouside
their file
This allows the compiler to warn if the definition in the file is
no in line with what is exposed to the users of the function through
the header definition.
The declarations have been split by source file, adding the usual
separation mark. Removed these 2 prototypes:
- W_SetFocusOfToplevel: case typo, the correct prototype already existed
in the file;
- W_TextWidth: function is not defined anywhere
When a function is used as a call-back, it is dangerous to have
arguments with a type different than what is expected by the
call-back definition.
This patch sets the argument list to match what is expected by
the call-back prototype and inserts an explicit type conversion
at the beginning of the function.
It is dangerous to let the compiler know about a function without
letting him know the arguments because he won't be able to report
invalid calls.
This patch concern the cases where adding the arguments led to
problems because the functions were used as call-back. As it is
dangerous to have parameter mismatchs in call-back, setup the
args as expected by prototype and added explicit conversion inside
the concerned function, so the compiler will know and be able to
do what may be necessary.
It is dangerous to let the compiler know about a function without
letting him know the arguments because he won't be able to report
invalid calls.
This patch concern the cases where adding the arguments did not
need other code change.
This is the correct way to tell that a function takes no
arguments, because an empty parameter list tells the compiler
that it is not yet defined, and is tolerated only for
compatibility with very old C compilers for whom prototypes
were not yet a defined language element.
The functions are now grouped by source file (groups in alphabetic
order) with the traditional separation mark. This makes the file
easier to work with.
It is not a good idea to multiply the number of header files,
specially in this case where 'double.h' defined so few things
(and lacked the usual copyright notice / include guards).
We probably don't want our users to have to endure them, so they
are enabled only when Debug is activated, because they tend to
help keeping the code safe.
The equality comparison (a == b) is known to be a dangerous trap
when floating-point arithmetics are involved. In the current case
the offending operation can be done with integers directly.
The equality comparison (a == b) is known to be a dangerous trap
when floating-point arithmetics are involved. This patch changes
all the cases which try to do this to a safer check.
The original code allowed to have 0.0, but this can generate
division by zero in WScrollView. As a value of 0.0 is not realistic
anyway, use a minimum constant instead.
The equality comparison (a == b) is known to be a dangerous trap
when floating-point arithmetics are involved. This patch changes
all the cases which try to do this to a safer check.
Some header were creating variable, this is a bad practice which
is likely to not behave as expected. This creates one distinct
variable in each object file that used the header, and:
- on well behaved compiler, this ends up in a link error (see
commit 39fdb451ba for an example)
- on bad behaving compiler, this can be linked as multiple local
variable, thus having strange effects when running program
- on insouciant compiler (who said gcc?) the variables are
silently merged, hiding portability issues
Autoconf provides the necessary stuff to detect if inline keyword
is supported, and to detect special syntaxes, so let's use this
and remove the multiple local definitions, this makes code simpler.
When using the formula [sizeof(array) / sizeof( x )] to get the number
of element in a static array, it is better to use array[0] for 'x'
instead of the base type of array:
- in case the base type would change someday;
- if the compiler were deciding to insert padding somewhere
Some compilation options are actually targetting the preprocessor
instead of the compiler; using the wrong variable can have some
subtile side effects, so let's get things right.
Considering the number of headers we have, it is a good idea to
avoid possible problems. For details, you may read:
http://en.wikipedia.org/wiki/Include_guard
All headers should be ok now.
The automake documentation states that using substitution inside the
list of SOURCES will not work and calls for not doing it. The use
of 'EXTRA_xxx' made things look like they worked but is probably not
enough for corner cases.
This patches switches to the conditional method which will be safe.
This patch includes info about XRandR extension in the info panel
dialog. If wmaker was compiled with xrandr support, then the dialog
show the XRandR info. The info includes if the X-Server supports or not
XRandR (wmaker could be compiled with XRandR support, but the X Server
may not include XRandR extension).
The string was separated in two by Christophe Curis to allow translation,
as suggested by Alexey I. Froloff.
To be able to do this in a clean way, it was necessary to add the
attribute also in PLData's ptr field, which is actually right
because none of the function changes its content.
The function that fills it from a file/pipe however needed small
changes to respect the const-ness of the field.
The history is actually loaded from a file into a property list that
is then converted to an array. The intermediate property list was
not freed, which led to memory leak.
It looks like it was a tentative of optimisation to avoid duplicating
an already allocated string and re-use the pointer instead, but this
means it is not possible to free the original container as it would
free the string too.
There is a better way to do this, but it requires an API change on the
WUtil library so it is left for a future improvment.
The previous code limited the number of entries that were read into
the history array, but the user is expecting a maximum on the
number of entries displayed. This can make a little difference in
two cases:
- if there are duplicate entries (dups are checked for and removed)
- if some entries are not strings (unlikely, but not impossible)
The new code just stops adding history entries when the user
specified count is reached.
When menus are read in the PropList format, they are loaded into a
temporary PropList object, which is parsed into the internal menu
structure, and the PropList object is no more used. There were two
cases where this temp object was not freed.
A common argument to all these functions is the name of the key to
operate on, and this name is never modified by the functions. Marking
it as const reflects this, and can allow compiler to generate better
results thanks to this info.
The addition of the const attributes is actually an API change, so
we have to reflect this for the next official release.
Because the change on 'wusergnusteppath' may impact users of the API,
we won't only change REVISION like it was done for WRaster lib.
The first optimisation is to compute only once the path, and then
always re-use the value which did not change anyway.
The second optimisation is to avoid a lot of excessive function
calls, including alloc+free that are not necessary and participate
in memory fragmentation.
According to the way its value is being used everywhere, that is
what would be expected, so let's make it official.
Please note that this may introduce warnings on user code using
this function ("...discard const...") but that's an opportunity
for them to check that their code is not doing anything wrong.
The internal function 'unescapestr' is used to transform strings which
may contain escape sequences (\x) into their plain representation.
There are a few cases where the function can misbehave (typically parse
after the end of string, thus writing past the end of the reserved
result area) which can be a source of problem later. The new code
should be safer.
The previous syntax used an explicit cast to remove the CONST
attribute, but the right way is to keep the attribute and store
the result in a variable which is defined properly.
This makes the WUtil API as much const-correct as possible for
the arguments being given to its functions.
This does not make it totally correct as it does not changes the
const-ness on returned values because the goal of this patch is
to make no visible change to existing program that would use this
library.
Note that the argument is also stored as-is in the PLData structure
but only for debugging purpose (warning display to user), hence the
choice to not duplicate it. As a side effect, it was 'const'-ified
too to reflect that.
As a side note, in 'wfindfileinlist' the first argument should be:
const char * const *path_list
However, due to limited support for const in plain C, that would
introduce warnings in user code. For compatibility issues, this
was not implemented.
A number of functions do not actually modify the strings given as
parameter, but only read or duplicate it. In this case it is a good
practice to mark that parameter as pointer-to-const to let the
compiler known about it, to be able to perform appropriate
optimisations.
Prototype of function changed in commit
d1e1c13fa3
but two of them were not updated in the source.
Thanks to Rodolfo García Peñas for taking the time to test and report.
The addition of 'const' to parameters of library's functions is an
API change, although this will break neither the binary interface
nor the compilation of program using the library; the other changes
also have no impact on the compiled library object.
All these functions expects agruments like color or list-of-points
that should not be modified (and are not) by the function; added
the corresponding qualifier to reflect that.
An enum is always a better idea as it allows the compiler to do
some checks, and as this info is internal only to the WRLib it
will not change the API.
The functions are declared in different files but they were called
in another file which re-declared the prototypes. This is dangerous
as it can lead to misaligned prototypes when functions changes.
They are now grouped in the library internal header 'imgformat.h'
The config-file-keyword was defined in an array, but the user displayed
string was defined at another place, thus requiring special care when
modifying the list, to keep the same number of entries, in the same
order, ...
This was an open call for subtle bugs, so the keyword and user string
are now at a single place, making updates on the list easier and the
GUI generation code smaller.
If the switchpanel was opened with either FocusNextKey or FocusPrevKey
shortcut, and the user presses GroupNextKey or GroupPrevKey, skip over
windows of a different class when cycling through windows in the
switchpanel.
In the case where the switchpanel was opened with FocusNextKey or
FocusPrevKey initially, the check can be skipped because all the
available choices are necessarily of the same class already.
The hasModifier flag was set if the FocusNextKey or FocusPrevKey
shortcuts had modifiers, depending on which shortcut was used to open
the switchpanel.
The switchpanel can also be opened with the GroupNextKey or GroupPrevKey
shortcuts, so we should account for those when setting hasModifier.
Window Maker's behaviour changes when StrictWindozeCycle is active. As
a rule we try not to set the default value of new options such that they
would change the behaviour expected by users.
In this case, however, the switchpanel was not working as intended.
Users who prefer the old method can set StrictWindozeCycle off with
wdwrite WindowMaker StrictWindozeCycle NO
As the name implies, StartWindozeCycle() cycles windows in the same way
that a popular commercially-available operating system does. However
Window Maker's handling of the shift key in the switchpanel does not
currently mirror that of its commercial counterpart.
In the popular operating system:
Holding alt and shift then pressing and releasing tab will highlight
the previous window in the switcher.
Releasing shift with alt still held will not close the switcher.
The window change is commited when alt is released.
In Window Maker:
Holding alt and shift then pressing and releasing tab will highlight
the previous window in the switchpanel.
Releasing shift with alt still held will close the switchpanel and commit
the window change.
This patch adds the StrictWindozeCycle boolean preference. When it is
set to YES the switchpanel will remain open as long as alt is held even
if shift is pressed and released.
The default case is moved to the bottom of the switch case.
The default case should be removed, because the icon has always
a right value, because the icon creation always uses a real value:
kix@debian:~/src/wmaker/wmaker-crm/src$ grep wAppIconCreateForDock *c
appicon.c:WAppIcon *wAppIconCreateForDock(WScreen *scr, char *command, char *wm_instance, char *wm_class, int tile)
dock.c: btn = wAppIconCreateForDock(scr, NULL, "Logo", "WMClip", TILE_CLIP);
dock.c: btn = wAppIconCreateForDock(scr, NULL, "Logo", "WMDock", TILE_NORMAL);
dock.c: btn = wAppIconCreateForDock(scr, NULL, name, "WMDrawer", TILE_DRAWER);
dock.c: aicon = wAppIconCreateForDock(scr, command, winstance, wclass, TILE_NORMAL);
(1)dock.c: aicon = wAppIconCreateForDock(dock->screen_ptr, NULL,
kix@debian:~/src/wmaker/wmaker-crm/src$
kix@debian:~/src/wmaker/wmaker-crm/src$ grep TILE_ *c | grep -v ICON_TILE_SIZE
***[2]appicon.c: tile = TILE_CLIP;
dock.c: btn = wAppIconCreateForDock(scr, NULL, "Logo", "WMClip", TILE_CLIP);
dock.c: btn = wAppIconCreateForDock(scr, NULL, "Logo", "WMDock", TILE_NORMAL);
dock.c: btn = wAppIconCreateForDock(scr, NULL, name, "WMDrawer", TILE_DRAWER);
dock.c: aicon = wAppIconCreateForDock(scr, command, winstance, wclass, TILE_NORMAL);
(2)dock.c: wm_instance, wm_class, TILE_NORMAL);
***[3]icon.c: icon->tile_type = TILE_NORMAL;
icon.c: case TILE_NORMAL:
icon.c: case TILE_CLIP:
icon.c: case TILE_DRAWER:
kix@debian:~/src/wmaker/wmaker-crm/src$ grep tile_type *c
icon.c: icon->tile_type = TILE_NORMAL;
***[1]icon.c: icon->tile_type = tile;
icon.c: switch (icon->tile_type) {
icon.c: wwarning("Unknown tile type: %d.\n", icon->tile_type);
kix@debian:~/src/wmaker/wmaker-crm/src$
There are only three cases without value (asterisk in the line start) set
as preprocessor variable. (1) and (2) is the same call. These are the three cases:
Case [1]:
-------------8<--------------
WIcon *icon_create_for_dock(WScreen *scr, char *command, char *wm_instance, char *wm_class, int tile)
{
WIcon *icon;
icon = icon_create_core(scr, 0, 0);
icon->tile_type = tile;
-------------8<--------------
Calls to icon_create_for_dock, is only call in appicon.c:
-------------8<--------------
kix@debian:~/src/wmaker/wmaker-crm/src$ grep icon_create_for_dock *c
appicon.c: aicon->icon = icon_create_for_dock(scr, command, wm_instance, wm_class, tile);
icon.c:WIcon *icon_create_for_dock(WScreen *scr, char *command, char *wm_instance, char *wm_class, int tile)
kix@debian:~/src/wmaker/wmaker-crm/src$
-------------8<--------------
The call:
-------------8<--------------
WAppIcon *wAppIconCreateForDock(WScreen *scr, char *command, char *wm_instance, char *wm_class, int tile)
{
[snip]
if (strcmp(wm_class, "WMDock") == 0 && wPreferences.flags.clip_merged_in_dock)
tile = TILE_CLIP;
aicon->icon = icon_create_for_dock(scr, command, wm_instance, wm_class, tile);
-------------8<--------------
And the calls to wAppIconCreateForDock() are checked before.
The case [2] is just the line:
-------------8<--------------
WAppIcon *wAppIconCreateForDock(WScreen *scr, char *command, char *wm_instance, char *wm_class, int tile)
{
[snip]
if (strcmp(wm_class, "WMDock") == 0 && wPreferences.flags.clip_merged_in_dock)
*** tile = TILE_CLIP;
aicon->icon = icon_create_for_dock(scr, command, wm_instance, wm_class, tile);
-------------8<--------------
Then, is sure too.
The case [3] is:
-------------8<--------------
WIcon *icon_create_for_wwindow(WWindow *wwin)
{
[snip]
icon->tile_type = TILE_NORMAL;
-------------8<--------------
All windows have TILE_NORMAL.
Then, all cases are secure.
WMMergePLDictionaries() and WMSubtractPLDictionaries() are declared to
return WMPropList * but are set to call the wassertr macro when their
arguments do not pass a sanity check. The wassertr macro eventually
calls return with no return value, triggering a compiler warning if
-Wreturn-type is used.
Change wassertr to wassertrv and force a return of NULL in the error
case.
When first using the wmdrawer panel of WPrefs the settings
ClipAutoexpandDelay
ClipAutocollapseDelay
ClipAutoraiseDelay
ClipAutolowerDelay
are not yet saved in the user's config file and 'value' ends up
being undefined in showData() leading to a crash in adjustButtonSelectionBasedOnValue().
This patch is a workaround to make it possible to choose the config values for the
first time using WPrefs.
The dock will have the up-right and down-left arrows to change workspaces and
the clip will be disabled. That is, if option ClipMergedInDock is set to yes in
GNUstep/Defaults/WindowMaker.
[not thoroughly tested]
No more ghost dock when switching sides: the real swap happens
immediately, you can still adjust vertically afterwards. Removed two
functions in superfluous that are no longer used
Drawers are horizontal docks, and they can themselves only live in the dock
To use them, right click on the dock or a docked appicon and select "Add
a drawer". Then move appicons into the drawer (drag them with the
mouse). You may change the icon of the drawer. By default, drawers
auto-expand and -collapse, and auto-raise/lower. This can be customized
in the same way as for the clip.
Set DisableDrawers to YES in G/D/WindowMaker if you do not want to see
the menu entry to add a drawer.
Just discovered this bug: the auto-attract icon functionality will not
work (to be precise, it crashes WM!) if the clip is disabled
(NoClip=YES). Will fix shortly, of course.
Behaves essentially the same, only a bit more consistently.
Known differences:
1. An AppIcon will now always end up undocked if moved while Mod1 is pressed.
2. Moving a docked AppIcon with Mod1 pressed (undocking it) used to auto-expand the clip,
as clip expansion happened first, while looking for a snapping position,
and the test on Mod1 being pressed happened only later
The options are ClipAutoraiseDelay, ClipAutolowerDelay, ClipAutoexpandDelay, ClipAutocollapseDelay
The default values are weird but merely represent the previously
hard-coded ones. They are repeated in Defaults/WindowMaker to avoid a
WPrefs crash (but it's a good idea to make them "visible", too)
WPrefs change coming up
When changing workspace, mapped windows are unmapped from top to bottom
(referring to their stacking order), causing mapped but obscured windows
to briefly appear when the obscuring window is unmapped and until they
are themselves unmapped. [This might not be visible on recent hardware].
The fix is to unmap windows in reverse stacking order when changing
workspace.
In handleIconMove, when moving an appicon near the clip, all clips were
systematically shown each time the mouse moved, causing some flickering
(perhaps only perceptible on a low-end machine?). This patch introduces
a showed_all_clips boolean to expand clips only once.
If btn->omnipresent, then btn->dock is precisely
scr->workspaces[scr->current_workspace]->clip (cf.
wWorkspaceForceChange() calling wClipUpdateForWorkspaceChange())
As proposed by Christian Schulte, when compiling WindowMaker on BSD
it uses the kvm library to gather information on processes (in the
os-dependant file), so we must properly link against the corresponding
library, which this patch adds.
The function wIconChangeTitle() now changes the icon title name
doing the full work (except painting it).
The function receives now the icon to change the name and the
wwindow with the new name. The function checks if icon and the
window exists.
Then, try to get the name using wNETWMGetIconName(), if not found
then try to read it from wGetIconName(). Then the icon has the new
name and the function returns.
This is better because:
1. We don't need a flag to know if the window got the name
using the wNETWMGetIconName function. Now call this function
always.
2. We do the same work in all calls to the wIconChangeTitle()
function.
The functions that uses wIconChangeTitle (at client.c, icon.c and
wmspec.c) uses always the value set by wNETWMGetIconName() first,
else, the value set by wGetIconName(). This is the reason for the
flag net_has_icon_title. Now the flag can be removed.
The call to wIconUpdate() can be changed by call to update_icon_pixmap(),
because the icon doesn't need to be changed.
Now, the icon pixmap is updated faster.
This change is more difficult to see, this is the explanation:
1. wIconUpdate() updates the icon for the applications, then call
update_icon_pixmap() to re-create the pixmap.
2. tileObserver() is used if the event WNIconTileSettingsChanged() is
launched. This event is used in the Notification for docks and wwindows:
WMAddNotificationObserver(tileObserver, icon, WNIconTileSettingsChanged, icon);
3. The function WNIconTileSettingsChanged() is used if the icon need refresh,
used in the wReadDefaults() function. See code below. This function is used
in the wmaker startup.
if (needs_refresh & REFRESH_ICON_TILE)
WMPostNotificationName(WNIconTileSettingsChanged, NULL, NULL);
4. Finally, the flag to refresh the icon tile is throw by setIconTile():
return (reset ? REFRESH_ICON_TILE : 0);
And this function only changes the the icon tile if reset is "1", that
happend if src->icon_tile:
if (scr->icon_tile) {
reset = 1;
RReleaseImage(scr->icon_tile);
XFreePixmap(dpy, scr->icon_tile_pixmap);
}
5. Then, we can drop the function wIconUpdate(), because the change is in the
icon_tile variable, used only in icon_update_pixmap(). This function is
only used in update_icon_pixmap() (not in wIconUpdate):
kix@kentin:~/src/wmaker/git/wmaker-crm/src$ grep icon_tile icon.c
tile = RCloneImage(scr->icon_tile);
XSetWindowBackgroundPixmap(dpy, icon->core->window, scr->icon_tile_pixmap);
kix@kentin:~/src/wmaker/git/wmaker-crm/src$
static void icon_update_pixmap(WIcon *icon, RImage *image)
{
[snip]
if (icon->tile_type == TILE_NORMAL) {
tile = RCloneImage(scr->icon_tile);
} else {
assert(scr->clip_tile);
tile = RCloneImage(scr->clip_tile);
}
[snip]
The XSetWindowBackgroundPixmap() call doesn't matter here.
The call to wIconUpdate() can be changed by call to update_icon_pixmap(),
because the icon doesn't need to be changed. This update is only for change
the icon pixmap to un-shadowed.
Now, the icon pixmap is updated faster.
The call to wIconUpdate() can be changed by call to update_icon_pixmap(),
because the icon doesn't need to be changed. This update is only for change
the icon pixmap to un-shadowed.
Now, the icon pixmap is updated faster.
The call to wIconUpdate() can be changed by call to update_icon_pixmap(),
because the icon doesn't need to be changed. This update is only for change
the icon pixmap to shadowed.
Now, the icon pixmap is updated faster.
The call to wIconUpdate() can be changed by call to update_icon_pixmap(),
because the icon doesn't need to be changed. This update is only for change
the icon pixmap to shadowed.
Now, the icon pixmap is updated faster.
The function set_icon_image_from_database() set the icon from the
database, but wIconUpdate() update the icon using different methods
, like for example get the image from X11. So, is better move the
wIconUpdate() to the function who call set_icon_image_from_database(),
to avoid understanding problems.
The function wIconChangeTitle() function calls wIconUpdate() or
wIconPaint() depending of changed. In both cases the icon->file_image
doesn't change, only the icon title, so we can use the same image
and we don't need update it. Because the variable changed is removed,
we can use wIconPaint() here.
Then the variable changed is not used, and these lines can be removed.
The function drawIconTitle() only draw the icon title square, without
the title. The title must be set later, because yet is not initialized,
so is NULL.
Then, the right function name should be drawIconTitleBackground.
This patch also removes the variable titled.
The function wIconChangeImageFile() now use set_icon_image_from_image()
then:
1. Don't need call unset_icon_image() because the memory is free in the
function set_icon_image_from_image().
2. After update the icon info, it must set the file name.
3. The pixmap image could be updated.
The new function set_icon_image_from_image() sets the icon image using
a image provided as argument.
This function will be used to avoid call wIconUpdate() with the image
argument, doing the code easier and faster.
This patch calls unset_icon_image(), to free the icon image, to avoid
lost memory. After this patch, in winspector.c, the wIconUpdate() call
don't free the memory before update it.
The definition and call for wGetRectForHead is moved inside the
if block where is used. This code adds a WScreen pointer to make
the line shorter.
If menu->frame is NULL, then was NULL before this patch, so this
code doesn't include an error about this.
This patch will used with next patch to move the code inside the
block to an specific function.
Added MoveToWorkspace*Key, MoveToNextWorkspaceKey,
MoveToPrevWorkspaceKey, MoveToLastWorkspaceKey,
MoveToNextWorkspaceLayerKey and MoveToPrevWorkspaceLayerKey to the
keyboard shortcut panel of WPrefs.
Added new keyboard shortcuts for moving windows between workspaces.
MoveToWorkspace1Key moves the active window directly to workspace 1.
Similarly for MoveToWorkspace2Key through MoveToWorkspace10Key.
MoveToNextWorkspaceKey moves the window to the next workspace,
MoveToPrevWorkspaceKey moves the window to the previous workspace.
Both keys respect the ws_advance and ws_cycle preferences.
MoveToNextWorkspaceLayerKey moves the window ten workspaces "forward" if
possible.
MoveToPrevWorkspaceLayerKey moves the window ten workspaces "back" if
possible.
A number of functions relating to the Workspaces menu assume that the
menu contains at least three entries: New, Destroy Last and Workspace1.
Because it is assumed that index 1 contains the Destroy Last entry and
index 2 contains the first workspace entry there are several occurrences
throughout workspace.c of magic numbers relating to those indices.
Replace the magic numbers with some #defines, making the code more
readable and allowing the menu to be expanded later.
The function getShortcutString() was defined statically in winmenu.c.
Replace it with the new function GetShortcutKey() which calls the
existing function GetShortcutString() so it can be used elsewhere and
cut down on code duplication.
A result of this change is that shortcuts are now labelled consistently.
Previously the format was different in generated menus, which used, for
example, M1 to refer to Mod1 whereas window menus used the full string
Mod1. Now both use the shorter form.
One could argue that the new function name is more consistent, as now
GetShortcutString() takes a char * argument and GetShortcutKey() takes a
WShortcutKey argument. That argument assumes that the original
intention behind the name of GetShortcutString() was not to hint that it
returns a String...
This code was used to create the window where the workspace name was showed
to the user. The code now use the standar balloon system, therefore this code
is not used anymore and can be removed.
Other toggle menu items such as Maximize and Shade change their text
labels according to the action which will be performed. A shaded window
has its Shade menu item text changed to Unshade, for instance.
As well as maintaining consistency with other menu items, changing the
Select menu entry's text to Deselect for current selected windows
provides another (the only) way of reporting that a window is selected
when its border has been disabled. Currently the only indication that a
window is selected is that its border colour will change.
Use the new preferences FrameBorderColor and FrameSelectedBorderColor to
set the border colour of frame windows and selected frame windows
respectively.
The function XClearWindow() is called twice, we can remove one.
See the asterisks (**):
Before the call to wIconPaint():
-----------------8<--------------------------
+++ b/src/icon.c
@@ -621,7 +621,6 @@ void update_icon_pixmap(WIcon *icon)
XSetWindowBackgroundPixmap(dpy, icon->core->window, icon->pixmap);
/* Paint it */
** XClearWindow(dpy, icon->core->window);
wIconPaint(icon);
}
-----------------8<--------------------------
First call in wIconPaint() function:
-----------------8<--------------------------
void wIconPaint(WIcon *icon)
{
WScreen *scr = icon->core->screen_ptr;
int x, l, w;
char *tmp;
** XClearWindow(dpy, icon->core->window);
-----------------8<--------------------------
The function wCoreCreateTopLevel() is used in two files (icon.c and
framewin.c), but after create the window, some attributes are changed.
This patch moves the change inside the wCoreCreateTopLevel(), avoiding to
call XChangeWindowAttributes() after the window creation. Now the window
is created in only one step, with all the final attributes.
Some details:
- The function wCoreCreateTopLevel() has now one argument more, the
border pixel color. This attribute was used always as the screen
frame_border_pixel, but in icon.c the attribute is changed to
white_pixel. Now the function wCoreCreateTopLevel() receives the
value frame_border_pixel in framewin.c and scr->white_pixel in
icon.c, as argument.
- The vmask and attribs variables and the call to XChangeWindowAttributes()
are removed in framewin.c and icon.c. The values CWSaveUnder for vmask and
attribs.save_under = True are used if wPreferences.use_saveunders is True.
- CWBorderPixel is not needed in icon.c, because was previously set in wcore.c!
The dock types NORMAL_DOCK and EXTENDED_DOCK are never used
so can be removed.
kix@debian:~/src/wmaker/git/wmaker-crm-base$ rgrep NORMAL_DOCK *
src/dock.h:#define NORMAL_DOCK WM_DOCK
kix@debian:~/src/wmaker/git/wmaker-crm-base$ rgrep EXTENDED_DOCK *
src/dock.h:#define EXTENDED_DOCK WM_CLIP
kix@debian:~/src/wmaker/git/wmaker-crm-base$
The function wSessionRestoreState() calls execCommand()
with the argument "host", but this argument is never used,
so we can remove the argument and the code to make that
argument, dead code at wSessionRestoreState.
From 35f19d77874d1b50de5505b6b1cb31531e1c055a Mon Sep 17 00:00:00 2001
From: Andreas Bierfert <andreas.bierfert@lowlatency.de>
Date: Thu, 31 Jan 2013 22:35:42 +0100
Subject: [PATCH 3/3] Teach WPrefs about OPEN_PLMENU
Add option to WPrefs to read, add and edit OPEN_PLMENU menu entries
This patch adds the OPEN_PLMENU options which behaves similar to OPEN_MENU but
can be used to parse command generated proplists. This can be used e.g. in
conjunction with wmmenugen like:
(
"Generated PL Submenu",
OPEN_PLMENU,
"|| find /usr/share/applications -type f -name '*desktop' | xargs
wmmenugen -parser:xdg"
)
v2: code cleanup
This functions reads a proplist from a pipe instead of a file (like
WMReadPropListFromFile does). It uses a call to popen to open the desired
command, reads data into a buffer till EOF and passes the data to getPropList
for parsing.
v2: code cleanup
There can be these 4 situations:
1. ignore flag + file set
2. no ignore flag + file set
3. ignore flag + no file set
4. no ignore flag + no file set
If the "ignore client supplied icon flag" is not set it means the user wants to use the client-provided icon,
so the eventual file name which might be in the entry field is ignored.
If the user sets the ignore flag and sets the icon name all is fine and Option 1 is done.
If the user tries to set an icon (ignore flag is set) but the file textbox is not set we show the
warning:
"Ignore client supplied icon is set, but icon filename textbox is empty. Using client supplied icon"
and use the client-supplied icon instead.
This patch removes all the appicon stuff from the application creation
to the appicon creation. Now, the application only calls one function
(create_appicon_for_application()) and this function do all the work.
The function do the same code than the code before this patch, but the
only change is that the "if" test to check if the appicon was found in
the docks now is negated, removing the return and doing the appicon_save
inside the function.
Finally, the old makeAppIconFor is now create_appicon_for_application().
The new function set_icon_image_from_database() removes the dup code
from these functions:
icon.c:icon_create_for_dock()
icon.c:icon_create_for_wwindow()
appicon.c:removeAppIconFor()
The only different change is that in the functions icon_create_for_dock()
and icon_create_for_wwindow(), the icon->tile_type assignment is done
before set the icon image filename and icon image, but this variable
is not used in these functions (is used in wIconUpdate function) but in both
functions the icon->tile_type assignment is done before wIconUpdate(), like
the code previous to this patch, so there is no problem moving icon->tile_type.
This patch sets the correct icon after kill a docked application.
To reproduce the problem (as Carlos reported to me), follow these
steps:
1. With a docked application, not running, set the icon image.
2. Launch the docked application.
3. Kill the application, using for example the app dock menu.
The icon is lost. But if wmaker is restarted, the icon backs again.
The problem is because when the application is closed, the function
removeAppIconFor() calls wIconUpdate(), but this function doesn't
find the icon in the file system again. The icon is loaded from the
filesystem only when the application is created.
This patch creates the icon image again, then calls wIconUpdate() to
set the icon image or the default icon image if the icon image was
not found, then wPaint it.
These lines comes from the create_appicon_from_* functions, used
to set the icon image for docks and applications. Are copy/pasted.
When the "Center" choice for window position is chosen, really enforce
the center position for all windows, instead of falling back to "Auto"
when there are already windows on the workspace.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The new function W_setconf_doubleClickDelay() sets the value for
WPreferences.W_setconf_doubleClickDelay(), therefore the private
data of WPreferences struct is not used.
This call is used at defaults.c to set the doubleClickDelay().
From the BALATON Zoltan comment:
If the P in WINGsP.h stands for Private then no files outside of WINGs
should include it. If WINGsConfiguration is an internal structure of
WINGs then only accessor methods should be used outside of WINGs.
This patch uses the new functions to read the WINGs configuration, not
using the internal library data.
These new functions:
unsigned W_getconf_mouseWheelUp(void);
unsigned W_getconf_mouseWheelDown(void);
returns the WINGs configuration for the Mouse Wheel Up and Down
values.
The new shutdown.h file includes the shutdown modes and the functions
to shutdown Window Maker.
The function definitios were moved from funcs.h and the struct WShutdownMode
from WindowMaker.h.
Now, externs are not needed, only include shutdown.h
This patch moves the XRANDR variable definitions to screen.h, because
xrandr is related to screen behavior. Then, the definition externs
can be removed.
This patch also changes (in event.c) the ConfigureNotify event processing
to avoid call the "if" block if XRANDR is not enabled. I chose the option of
move the if into the HAVE_XRANDR block instead of include the
ConfigureNotify case inside of the HAVE_XRANDR block (proposed by
BALATON Zoltan) to avoid call the default option if not needed.
Resizing windows with the mouse wheel was introduced in a063338175
("Mod+Wheel Window Resize") and it should be disabled by default
in order to not confuse users.
Users wanting this behavior can enable it through WPrefs.
Bug report: https://bugs.launchpad.net/ubuntu/+source/wmaker/+bug/1082879
These are the final changes for the new version 0.95.4. The main change
is set right the symbols for the new version, that include:
- Changes in debian/libwutil2.symbols from the git version to 0.95.4
- Changes in debian/changelog about the comments for the new version
On the other hand, the new debug schema is using hashes, therefore
the /usr/lib/debug/wmaker folder not longer exists. The new folder
is /usr/lib/debug/.build-id. This folder is auto-created and then
we don't need create it in debian/wmaker-dbg.dirs. Therefore, the file
debian/wmaker-dbg.dirs is empty and can be removed.
Finally, a little comment about folder in debian/README.Debian.
This code was commented out but enabling it leads to no issues AFAICS
and improves the usability of WINGs applications.
The motivation for this patch comes from the need of distinguishing
where the cursor is in the WINGs-based application which handles the
database of my comics collection.
This changes the behavior in all parts of wmaker where there is
a text field entry, e.g. in the settings panel of dockapps. There should
be no issues with a blinking cursor in such cases though...
For a icon, the icon image is stored in icon->file_image. If we repaint
the image, we we don't need search the icon again. We can continue using
icon->file_image.
Indeed, if launch the "window inspector" window (winspector), using a
saved icon and "ignore client icon" set, then the database holds the
name of the icon to draw on the screen. If we remove the icon info in
the winspector (removing the file name or un-setting the "ignore client
icon" flag, then the database losts the icon name and then the function
wDefaultUpdateIcons() cannot find the icon to draw for the icon.
In winspector, with this patch, the previous image set by "ApplySettings"
and stored at icon->file_image is used.
This patch avoids to save the icon name in the configuration file
if the flag "Ignore client supplied icon" is not selected.
First, when winspector tries to show the icon to the user, it must not search
the default icon. It should use the specific icon or the icon provided
by the client.
When we click in "save Settings", the function saveSettings reads the icon
specified by the user in the text box. If the checkbox "Ignore client supplied
icon" is not selected, then saves the icon too.
Using this behaviour we never recover the initial position (no icon selected,
no ignore client supplied icon flag set). This patch recovers the initial
state of the icon options.
When the user unsets the ignore client the filename provided is not saved in
the configuration file. Now, the configuration is like the initial status.
If the user checks the flag and set the file to use as icon, the file is used.
This patch sets a correct style on winspector.c. It removes some
empty lines, moves variable definitions to the function heads,
moves preprocessor definitions to the top of the file and removes
some extra curly brackets.
The function to create the winspector panel is so big, and is hard
to understand it. This patch split the function createInspectorForWindow
in some extra functions:
create_tab_window_attributes
create_tab_window_advanced
create_tab_icon_workspace
create_tab_app_specific
This functions don't do nothing extra, only the code is moved to them.
This patch avoid a crash when moving an icon without command.
To reproduce the problem:
1. Launch an application, for example xeyes, with appicon.
2. Move the appicon to the clip.
3. Close the application.
4. Edit the appicon in the clip, and empty the commands fields.
5. Move the appicon from the clip to the dock. -> Crash.
The crash happends because icon->icon->owner is NULL and then
wwin will be NULL. Then the call of wwin->client_win will crash.
This patch checks if icon->icon->owner is not null (application is
running) and then assign it to wwin. Then get the command from the
running application.
Added CycleIgnoreMinimized configuration option settable on Expert page in WPrefs.
When option is set, switch panel cycling ignores minimized (grayed) windows. They
are still visible and can be selected using left/right arrows or mouse click.
Please find attached a diff to add the missing .h files from src/Makefile.am
One way to find these:
$ git checkout next
$ ./autogen.sh
$ ./configure
$ make dist-gzip
$ tar xzf WindowMaker-0.95.3.tar.gz
$ cd WindowMaker-0.95.3
$ ./configure
$ make
The list of options in the Expert tab of WPrefs.app was filled using
hard-coded list of commands with static indexes and the related
usage were spread accross procedures. This was dangerous because
it could lead to mismatched indexes, the list length being also
variable, so it was hard to update.
This patch proposes to group all the options in a single place (a
static const array), so it is easy to add/remove/reorder the options,
and the code just parses this array based on its auto-calculated
size.
The function applySettings() set now the icons. Updated icons are:
- Appicon: Always, with or without always_user_icon
- Window: If the window is minimized, with or without always_user_icon.
Window must be minimized to see the icon change, else, wwin->icon
doesn't exist and therefore you won't see this change (no icon).
There is a problem if the window is not minimized and is minimezed,
because it will show the icon in the database, not the assigned icon.
This problem must be updated not here, in the wIconifyWindow()
function, at actions.c
The function wIconChangeImageFile set the file_image variable directly,
now the code is faster and the assigned image is used.
Before this patch, the assigned file couldn't be used, because wIconUpdate
serached the icon using different methods.
The variable wwin is only used in one block of the if, so should be moved
inside this block. OTOH, is better check if the variabl exists before
assign it. The code now is more stable and avoid crashes.
The dock applications has their icon image set in icon->file_image
to draw the icon if the application is not running. But if the
application is running, the icon mustn't be show. This patch don't
paint the icon for docks.
The function wIconUpdate can receive a image to setup as icon image.
If image is NULL, then use the original method, using different procedures
to get the image.
Two problems caused Boolean options which default to true to be shown
incorrectly when the option was not already present in the defaults database.
First a typo (Enabled instead of Selected) caused the switch button to stay
off then because GetBoolForKey key returns False also when the option
is not set it would have been turned off while it should stay on to
reflect the default value of the option.
Windows were drifting by FRAME_BORDER_WIDTH pixels when their
borders were toggled on or off.
Windows which had a border before we managed them were drifting
on shutdown and again at startup. It happened because the absolute
upper-left co-ordinates of a bordered window would in fact be the
upper-left co-ordinates of the border itself, whereas we consider
the client window to have no border and co-ordinates offset by the
titlebar and frame border.
The initial fix for the bug reported by Paul Seelig whereby windows
would drift on restart introduced two regressions.
New windows would place higher on the screen than intended, possibly
obscuring the bottoms of other windows with their titlebars, and all
windows would jump vertically at shutdown because we weren't restoring
them to where they were before they had a titlebar and border.
Bug report from Paul Seelig:
"Yet another rather strange glitch:
- open three terminal windows
- repeatedly restart wmaker
- all windows slowly drift to the left and up by just a few pixels
If i remember correctly, this is also a longstanding issue and nothing
new. It is no showstopper either, as one rarely restarts wmaker."
The slight drifting left and up seems to have been due to
wWindowConfigure() accounting for the window border when placing, which
was fixed in an earlier commit.
Windows could still shuffle down, however, because wWindowConfigure()
was moving the window down to make room for its window frame.
We now move it up by the titlebar height to cancel out that movement.
As reported by Amadeusz Sławiński, there were a number of debug-only
information printed by the menu parser, which had no reason to be
in a user package.
This patch removes all the hard-coded DEBUG definitions pointed
by Amadeusz and adds a new (standard-like) option '--enable-debug'
to the configure script, which activates debug stuff for the devs.
The default behaviour is now to not have them.
As a side effect, the option also disable optimisation, which are
generally annoying when trying to run a debugger.
Managed to squeeze two more options from the expert page to the
corresponding pages and also reorganised some widgets to avoid large
now unused spaces in one place and clipped text in the other.
As found by Rodolfo, it looks like there could be memory leak in the
function 'RSmoothScaleImage' because it reserveda given number of
memory blocs but used another count to free them after use.
This patch uses the same count for release as it seems this variable
is not modified in between. Took the opportunity as Rodolfo proposed
to convert a global variable to a local variable - this global
definition seems incorrect.
As reported by Amadeusz Sławiński, the support for (d)quoted string
in the menu file was not exactly the same as in previous versions
of WindowMaker, so we had a regression which is not acceptable.
This patch propose a consistent handling for (d)quoted text instead
of the previous on-the-title-only code; now all (d)quoted strings
are still kept as-is (as expected) but the (d)quotes are always
removed.
As a side note, it also improve the support for \escaped character
to work correctly as mentioned in the example of the default menu.
Note: This reverts b2e7620868 which
was an incomplete solution to the problem.
The function wIconValidateIconSize checks if the width size and height size are
less than the preference size (and left space for the border). If the width
size or height size is greater than the preference, then checks what is the
bigger size of them. Then resize it using the bigger value and holding the
aspect ratio.
Before this patch, wIconValidateIconSize didn't check if height was bigger
than width and always suppose that width was greater than height.
The function get_wwindow_image_from_wmhints returns a image from WM Hints,
but the image could be larger than the desired.
Then, the image can be resized using wIconValidateIconSize(). The resize
should be done in get_rimage_icon_from_wm_hints(), not in the function
get_wwindow_image_from_wmhints(). This is because the function
get_wwindow_image_from_wmhints() is used in wIconStore() too. If we resize
the image before save it to disk, then if we change the icon/dock size, then
the image saved will have a different size than the curren icon size. Is
better resize the image when is painted in the screen, not the image saved.
As reported by Rodolfo, there are some cases when working with icons
where a crash can occur, which is related to trying to re-release
some memory that have been already freed previously.
This patch adds a reset-to-NULL of the corresponding pointers so that
on next usage wmaker will know there's no more memory associated with
these pointers.
On 12.11.2012 Paul Seelig reported:
- open an application positioning an app icon on the bottom
- right click this app icon to show the context menu
- wonder yourself why there are two lines saying "Hide"
- first Hide entry does not do anything, second does
The reason for this curious behavior is the following.
The "Launch" entry was added in 8352c9ef60 ("Allow relaunch with shortcut key")
as the first one in the appicon menu, but this first position was hard-coded
in another part of wmaker's code in order to decide the menu entry text based
on the application's 'hidden' state in openApplicationMenu():
if (wapp->flags.hidden)
menu->entries[1]->text = _("Unhide");
else
menu->entries[1]->text = _("Hide");
But the "Launch" entry is before these "Hide/Unhide" entries and now the assumption
about entries[1] containing the relevant string for this hide/unhide decision is
no longer valid.
The simpler "fix" is to move the "Launch" entry below these "Hide/Unhide" games.
Move bounce options from expert prefs to its own box on ergonomic
prefs to have them at one place which makes them somewhat more clear.
The options previously occupying this place have been moved to other
pages where they better belong.
This patch only add some tabs the window.h (only window struct area).
The main change is that moves this block to switchpanel.c, because
is only used there:
((w)->wm_gnustep_attr->flags & GSWindowLevelAttr) && \
((w)->wm_gnustep_attr->window_level == WMMainMenuWindowLevel || \
(w)->wm_gnustep_attr->window_level == WMSubmenuWindowLevel))
The patch also include two comments in the window's image:
- WIcon icon is only used when the window is minimized, else is NULL
- RImage has the window image, used in switchpanel,...
The WIcon variable force_paint can be removed, because now is possible
update the icon directly in the function wAppIconPaint.
Now wAppIconPaint creates the RImage file if needed, and then paint it.
The function get_default_image_path has the WScreen argument that
is not used, so can be removed.
The function could be static, but perhaps is interesting leave it
as public.
The function get_default_icon_filename(), now get_icon_filename(),
was working in a bad way. Before this patch, the function always
searched the default icon in the second search:
- file_name = wDefaultGetIconFile(winstance, wclass, True);
+ file_name = wDefaultGetIconFile(winstance, wclass, False);
For this reason, the argument default_icon didn't work.
This patch change it. Now, if the default_icon is false, the function
can return NULL, then other functions can do the correct work. For
example, now wDefaultGetImage() doesn't need do nothing, because the
default_icon is set to True. get_icon_filename() checks if the
icon exists in the disk before returning it (except for default icon
in get_default_image_path(scr).
The line "icon_editing = 0" is the first line in the "if" sentence
and the first (and only) sentence in the "else" block, so can be
moved outside the if-else (and the empty else block can be removed).
Bug report from nikty:
"Window Maker does remember the size of windows, but for maximized windows it does not keep the position.
When a maximized window appears it is shifted to the right for a few pixels (firefox, thunar) or both to the right and to the bottom (openoffice, vlc, virtualbox)."
We weren't accounting for the window border when calculating whether the window would position correctly.
To reproduce:
* Launch some application which remembers its position. I used
Thunar as suggested in the bug report.
* Maximize!
* Kill the application.
* Launch it again.
* Maximize! The expected behaviour is that nothing would happen
because the window should have started right where it was before.
Observed behaviour is that it moves a few pixels.
After the patch we can verify that the maximize operation is
idempotent with regards to geometry.
Tested with and without Xinerama, with and without a panel strutting
one edge.
Bug report: http://www.kix.es/mantis/view.php?id=4
The function wIconSetHighlited() is now faster because it doesn't
re-create the icon before setting the Highlited status. Now it only
draws the icon pixmap.
The new function update_icon_pixmap() updates the pixmap for a icon task.
Now, wIconCreate() creates the image and sets the icon->file and
icon->file_image variables and update_icon_pixmap() updates the icon->pixmap.
This is interesting because we can update the pixmaps without creating the
image again.
The function wIconUpdate now create the rimage for the icon, using
different functions. When the icon is created, then the pixmap is
generated using this rimage.
The function get_pixmap_icon_from_icon_win(), before this patch,
creates the pixmap icon for dockapps (docks with a mini application
inside). The function did three steps:
1. Create the (r)image
2. Create the pixmap using the image
3. Put the application inside the image
Now these three steps are three functions:
step 1: static void get_rimage_icon_from_icon_win(WIcon *icon);
step 2: static void get_pixmap_icon_from_icon_win(WIcon *icon);
step 3: static void set_dockapp_in_icon(WIcon *icon);
These functions contains the same code, that before the changes.
The original function get_pixmap_icon_from_icon_win() includes the
calls to the new functions get_rimage_icon_from_icon_win() and
set_dockapp_in_icon()
There is a bug in wmaker with icon files set in config files, but
that doesn't exist in the disk. For example, if the config files have:
etc/WindowMaker/WMWindowAttributes:
xcalc = {AlwaysUserIcon = Yes;Icon = "HP-16C-48.xpm";};
share/WindowMaker/IconSets/Default.iconset:
xcalc = {AlwaysUserIcon = Yes;Icon = "HP-16C-48.xpm";};
But the icon "HP-16C-48.xpm" doesn't exist in the disk, wmaker does:
1. Load the config file in memory, in a database
2. When the application is launched, wmaker tries to find their icon in
the database (using wDefaultGetIconFile), the icon is found: HP-16C-48.xpm
3. When WindowMaker try to find the full path for the icon, using FindImage(),
WindowMaker cannot find the icon, and returns NULL.
Even, if the user set the default_icon boolean variable to True, wmaker
finds in the database the wrong icon.
This patch checks that the icon exists both in the database AND in the disk. If the
icon doesn't exist in the disk and the default_icon variable is set to True,
then windowmaker loads the default icon using the function get_default_image()
because this function searches the default icon directly.
The function get_default_image() is moved from icon.c to wdefaults.c because
it is now used in both places. This function is now splitted, to find the file
(get_default_image_path) path and the file image (get_default_image)
The function get_pixmap_icon_from_icon_win() is now rewritten.
This function now creates the icon using the function get_window_image_from_x11(),
using itself as Window and updates the icon using icon_update_pixmap().
At this point the Pixmap is only the icon background. Now we must put inside the
dockapp, in the correct position. Then the funcion get the dock application size,
(the max size is wPreferences.icon_size, because get_window_image_from_x11() checks
the valid size), checks if the icon has title and then reparents the dockapp to put
it inside the icon.
This patch includes a lot of comments to better understand the code.
The function get_wwindow_image_from_x11() is renamed to get_window_image_from_x11()
and does the same work, but now its argument is a Window struct instead of a
WWindow.
This change is better because it allows objects with Windows (but without
WWindows) to call this function.
The function now is not static to allow its use in other parts of the code.
A new functions get_pixmap_icon_from_x11 and get_rimage_icon_from_x11
are included in icon.h
1. get_pixmap_icon_from_x11, set the rimage at icon->file_image
The code sets net_icon_image in icon->file_image
2. The get_rimage_icon_from_x11 function, only converts
the icon->file_image in icon->pixmap
The function get_pixmap_icon_from_wm_hints() is splitted in two
functions now:
1. get_rimage_icon_from_wm_hints, set the rimage at icon->image
The code comes from the function get_pixmap_icon_from_default_icon()
2. The get_pixmap_icon_from_wm_hints() function, but now only converts
the icon->file_image in icon->pixmap
The function get_pixmap_icon_from_user_icon() is splitted in two
functions now:
1. get_rimage_icon_from_user_icon, set the rimage at icon->image
The code comes from the function get_pixmap_icon_from_default_icon()
2. The get_pixmap_icon_from_user_icon() function, but now only converts
the icon->file_image in icon->pixmap
The function get_pixmap_icon_from_default_icon() is splitted in two
functions now:
1. get_rimage_icon_from_default_icon, set the rimage at icon->image
The code comes from the function get_pixmap_icon_from_default_icon()
2. The get_pixmap_icon_from_default_icon(), but now only converts
the icon->file_image in icon->pixmap
The function get_pixmap_icon_from_wm_hints() is now rewritten using
the functions get_wwindow_image_from_wmhints() and icon_update_pixmap().
Now, the function creates a new RImage using get_wwindow_image_from_wmhints()
and then updates the Pixmap using the function icon_update_pixmap().
The CachedPixmap folder is added to the icon list. This is because
if one icon is created, but the user change it and set other icon, the first
icon cannot be selected anymore using the settings window. The user
needs modify the configuration files by hand.
Some extra paths are added to the debian default config file.
This patch changes the behavior of wIconChangeImageFile.
Before apply this patch the behavior was:
The image (file_image) is set to NULL, then the function tried to find
a new image. If the new image was not found, the variable file_image contains
NULL when the function returns, and the return code was 1.
With this patch the image is changed only if a new image is found.
The function wIconChangeImageFile mustn't update the icon if
no file is set.
This is a bug. To reproduce it:
1. Right click on dock icon. Select "Settings"
2. Erase the Icon Image contents (set empty)
3. Click "OK" -> Exit, no changes... no?
4. Right click on the same dock icon. Select "Settings"
5. Erase the Icon Image contents (set empty)
6. Click "OK" -> Crash
If wmaker don't crash, watch the file ~/GNUstep/Defaults/WMWindowAttributes,
the default icon ("*" has trash in the name). Probably something is not fine
at wIconUpdate.
Respect _NET_WM_STRUT and _NET_WM_STRUT_PARTIAL from all windows, not
just WWindows. This will allow us to respect struts from app icons
and dock icons.
Brief comment on what wAppIconFor() and wWindowFor() do, as it
may not be immediately obvious.
Given a raw X11 Window they will find the AppIcon (or WWindow,
respectively) associated with that window. Thus they can map an
X11 object to a Window Maker internal object.
Create wAppIconFor() and wAppIconTouchesHead() to match wWindowFor()
and wWindowTouchesHead(). These functions will allow us to locate the
window and head associated with a particular app icon or dock icon.
Window Maker already supports the _NET_WM_STRUT property as described
in the EWMH spec. We respect client-provided struts and avoid placing
or maximizing windows over those areas. An example is that we don't
try to place or maximize windows where they would be obscured by an
always-on-top gnome-panel.
_NET_WM_STRUT is now deprecated and redefined as a special case of
_NET_WM_STRUT_PARTIAL, which allows variable strut widths. A panel at
the bottom of the screen, for example, does not have to reserve the
whole width as a strut if it does not fill 100% of the screen width.
By default the XFCE bottom panel does not extend the whole width of
the screen, for instance.
Our method for restricting parts of the screen from placement doesn't
have a way to account for struts which are not 100% tall or 100% wide,
so until now we have ignored partial struts. In the case of the XFCE
panel mentioned above, the result is that a window may maximize
underneath the panel and be obscured.
As a partial hackaround we now query windows for _NET_WM_STRUT_PARTIAL
but throw away the start and end co-ordinates, assuming instead that
the struts are full-width/full-height. This trades off a small amount
of wasted placement space to avoid the case where windows can be
partially obscured by panels, which can be particularly annoying if
the panel is at the top and the victim's titlebar becomes hidden.
This patch removes the extra commas in the configuration file
WindowMaker.
This bug was created in the commit 0e995bbf6e
but no debian packages were released.
The function remove_wwindowstate is now remove_wwindowstate. This was
proposed by Christophe and Carlos:
- - -
On 2012-10-09 01:09, Carlos R. Mafra wrote:
> On Tue, 9 Oct 2012 at 0:58:19 +0200, Christophe wrote:
> you name the function "remove_*", which makes think that it would
> remove the entry from the list, but the function does not do that, it
> just frees the memory, so it may have been better called "free_" (or
> maybe "release_")?
>
Good point, release_ makes more sense to me.
- - -
The function get_default_icon_rimage name is not correct, because
the function doesn't return the default icon rimage, returns the
rimage from a file name. So the correct function name should be
get_rimage_from_file.
The new helper function get_pixmap_icon_from_default_icon search
the default icon in the disk and return it.
Now get_pixmap_icon_from_user_icon() returns do the work about
search the user icon, and the work about default icon is splitted
in the function get_pixmap_icon_from_user_icon.
The function makeIcon is only used in icon_update_pixmap(), so
both functions can be joined. Now the function icon_update_pixmap()
includes the contents of makeIcon.
This patch replaces the Pixmap icons from screen.h with only one
RImage. This image is not processed yet, therefore could be used
for icons with title or without it (replacing def_ticon_pixmap and
def_icon_pixmap variables).
Now the code is better because the Pixmap is generated and saved in
the icon structure.
See that before this patch, the icon->pixmap was set to None!!
With this patch, the icon->pixmap is saved, using the common
method used in the other functions to create Pixmaps, the function
icon_update_pixmap().
The new function get_default_image creates a RImage with the default
icon. Now the function get_pixmap_icon_from_user_icon() is more clear
and show that, first try to get the user selected image, if not set,
then it gets the default icon.
The functions:
void get_pixmap_icon_from_icon_win(WIcon *icon);
int get_pixmap_icon_from_wm_hints(WIcon *icon);
void get_pixmap_icon_from_user_icon(WIcon *icon);
should be static, because are only used in icon.c
The function icon_update_pixmap updates the WIcon's pixmap,
using the values in the WIcon struct.
This function provices a clear function to update the icon,
using only the icon, and the new image to set. This option is
better than using "makeIcon" function, with a lot of arguments.
This patch also changes the variable name "icon" to "image",
because is type rimage and then both functions uses the same name.
The function get_pixmap_icon_from_wm_hints has two extra arguments
(WScreen and WWindow) not needed, because these arguments can be
accessed using the argument WIcon.
This function removes them in the function definition and provides
them in the function code.
The function get_pixmap_icon_from_user_icon uses the argument
scr (WScreen) that is not needed, because the argument is in the
struct WIcon, the other function argument.
This patch removes the argument WScreen and provides it in the function
using WIcon.
This patch do these changes:
1. Removes the extra curly brackets
2. Add curly brackets if needed (code style)
3. Removes spaces in the function prototypes/declaration
I think this is the best moment to do it, all together
in a style clean patch.
4. Moves variable definition with the same type to the same line
The function wNETWMCheckClientHintChange returns a Bool,
but this value is never used, son can be removed.
This function now is void.
kix@osaka:~/src/wmaker/wmaker-crm-tosnd/src$ grep wNETWMCheckClientHintChange *.[ch]
client.c: wNETWMCheckClientHintChange(wwin, event);
wmspec.c:Bool wNETWMCheckClientHintChange(WWindow * wwin, XPropertyEvent * event)
wmspec.h:Bool wNETWMCheckClientHintChange(WWindow *wwin, XPropertyEvent *event);
The function wNETWMCheckClientHints returns a Boolean, but
this value is never used, so can be removed.
This patch removes the boolean returned, the variable hasState,
and so some parts of the function changes.
The patch also moves some variable definition to the same line:
one type, one line.
The variable "check" is always False, so the if (!False) is always
True and the function wNETWMCheckClientHints is always called.
The variable check is not used anymore, so the variable check
can be removed.
The function wNETWMCheckInitialClientState returns a Bool,
but this value is never used, so the function can be void.
kix@osaka:~/src/wmaker/wmaker-crm-tosnd/src$ grep wNETWMCheckInitialClientState *.[ch]
window.c: wNETWMCheckInitialClientState(wwin);
wmspec.c:Bool wNETWMCheckInitialClientState(WWindow * wwin)
wmspec.c: wmessage("wNETWMCheckInitialClientState");
wmspec.h:Bool wNETWMCheckInitialClientState(WWindow *wwin);
kix@osaka:~/src/wmaker/wmaker-crm-tosnd/src$
The wmessage function returns "enter" in two functions, so is
not possible to know the function that sends the message.
It is better to replace it with the function name.
This reverts commit 824255b1ae.
According to Iain Patterson:
This patch breaks the frame extents stuff I submitted recently. With
the patch and compton set to draw semi-opaque frame extents what
actually happens is no window decorations are drawn and a bunch of
BadDrawable errors are spewed to the console.
After reverting the patch everything works again.
The code was setting some values in the window's attribute structure
which were not being used (missing the corresponding vmask flag) and
was setting some vmask bits without setting the corresponding
value in the structure.
The new function 'unpaint_app_icon' checks its argument 'wapp' aginst NULL,
however it does dereference the pointer beforehand, which will end up in
a crash if the null-pointer case occurs.
As found by Rodolfo, a crash could happen when changing an icon to
a non-existent file. From his report:
Steps To Reproduce
1. Open the Settings box, for example for WMDock. (Right click on WMDock, Settings).
2. In the "Icon Image" field, write some characters. For example "asdfasdfasd".
3. Click in "OK". You get "Could not open specified icon file: GNUstep.tiffasdfasdfasd"
4. Hit in "OK"
5. Repeat 2,3,4 some times (sometimes 1, sometimes 2, sometimes 10).
The problem came from freeing the previous icon but not updating
the pointer, so if the new file did not exist it ended up with a
fake-looking icon being used, hence possible crash.
BTS: http://www.kix.es/mantis/view.php?id=1
This patch avoids the icon creation in winspector.c and adds the
ability of creating + paiting and unpainting instead of destroying the icon.
Now the icon is always created by wApplicationCreate and the icon
exists while the application is runnning. If the user doesn't want
an appicon the winspector.c will not remove the icon, it will only
not paint it on the screen. But the icon is still created.
Probably the most difficult part in this code is how to handle the
icons in the iconlist. We must include the icon in the iconlist when
it is painted, not when it is created. And it must be removed when it
is unpainted.
We can check if the icon is in the iconlist if icon->next AND icon->prev
are null, else it is on the applist. If it is included we must not
paint it again because the function PlaceIcon() will calculate a new
icon place in the screen including the icon!, then a hole is painted.
This patch set the vmask in one line.
- vmask = CWBorderPixel | CWCursor | CWEventMask | CWOverrideRedirect;
- vmask |= CWColormap;
Is similar to:
+ vmask = CWBorderPixel | CWCursor | CWEventMask | CWOverrideRedirect | CWColormap;
And vmask is not used between the two original lines, therefore we can do the
initialization in only one line.
The contents for searching the icon in wAppIconCreateForDock()
are moved to icon_create_for_dock().
Now wAppIconCreateForDock() and wAppIconCreate() do the same work and
icon_create_for_dock() and icon_create_for_wwindow() do the same work too.
The procedure to create the icon is similar in both functions.
Now wAppIconCreateForDock, needs more arguments (command, instance and class).
These functiosn were renamed:
wIconCreateCore to icon_create_core
wIconCreateWithIconFile to icon_create_for_dock
wIconCreate to icon_create_for_wwindow
The variable "dicon" is now named "aicon", the same name that
the WAppIcon variable used in wAppIconCreate.
Some clean code more, like spaces in function arguments.
Remove the call to wIconUpdate() in wAppIconCreate() and place it in
makeAppIconFor() since wAppIconCreate() is only used in that function
and icon update or icon painting is not icon creation.
Now the function wAppIconCreate() only creates the app_icon.
The function wIconCreate search the icon image calling wDefaultGetIconFile.
wDefaultGetIconFile search first the icon file name and then search the
image using the icon file name. wDefaultGetIconFile returns the icon image
to wIconCreate, then wIconCreate search the file name (again).
Is better that wIconCreate search first the file name, and then get
the image using the file name.
Code in the function wIconCreateWithIconFile does the same task as
get_default_icon_rimage(), therefore replace it with a call to that
function.
It is the code to get the image file (rimage) using the file name.
This patch reviews the code style in dock.c, removing some
curly brackets not needed and adding other where needed.
This patch removes some lines commented and some extra lines for
better understanding.
After we exit there are no window decorations therefore frame extents
are meaningless.
We could be left with parts of the window being the wrong opacity if the
property is left intact and a compositing manager configured to draw
decorations with a different opacity to the rest of the window is still
running.
Remove the _NET_FRAME_EXTENTS property from all windows when we shut
down to prevent windows from being drawn incorrectly after we're gone.
This is not a final version, is only a version for developers.
This version includes compatibility with debconf 9.
This version will be modified to change the libwutil2 symbols
to a new upstream version, probably 0.95.4.
Using window-supplied depth, visual and colormap information has the
side effect of causing window borders to be draw using inconsistent
colormap entries. Allocate entries from each window's colormap when
drawing its border.
Force setting the border when the window is first created so it's
guaranteed to be drawn in a consistent state.
Recalculate frame extents when the titlebar, resize bar or border are
enabled/disabled.
Account for border when calculating top and bottom frame extents.
Quoth I,
> I've just seen that _NET_FRAME_EXTENTS isn't updated when
> disabling or enabling the titlebar, resizebar and border of a window,
> so that needs to be fixed.
The attached patch fixes _NET_FRAME_EXTENTS not updating when using
the inspector to disable or enable the titlebar, resizebar or border.
It also fixes not taking the border width into account when
calculating the top and bottom extents.
With the patch the window's border is drawn for 32bpp urxvt windows
with compton. The border still isn't black, however. That's because
the border is taken from the screen's colormap rather than the
window's. I'll have a fix for that soon.
Accept windows' depth, visual and colormap instead of always using those
of the root window. Internal windows such as menus behave as before.
In conjunction with a compositing manager on a display supporting the
RENDER extension windows can now manage their own opacity.
I wrote the patch after reading the FAQ for urxvt, which says,
regarding transparency support:
"3. Use an ARGB visual:
urxvt -depth 32 -fg grey90 -bg rgba:0000/0000/4444/cccc
This requires XFT support, and the support of your X-server. If that
doesn't work for you, blame Xorg and Keith Packard. ARGB visuals
aren't there yet, no matter what they claim. Rxvt-Unicode contains the
necessary bugfixes and workarounds for Xft and Xlib to make it work,
but that doesn't mean that your WM has the required kludges in place."
In conjunction with a compositing manager (I tested compton) it does
work and urxvt draws a semi-transparent background with fully opaque
foreground text and scrollbars; much prettier than applying a blanket
transparency value over the whole window with the compositing manager.
Other application windows I tested were, as expected, drawn the same
as before.
I verified that urxvt is drawn in the same way when using xfwm4
(with builtin compositing). Since Window Maker doesn't (at time of
writing) have its own compositing manager I should clarify that one is
required to see any benefit from this patch.
Whether or not this feature is useful for any application other than
urxvt I don't know, though I assume that an application which chose an
ARGB visual in the same way would be able to draw itself prettily.
On a display without RENDER things work just as they do without the
patch. I have, however, only been able to test on a fairly standard
TrueColor display supporting multiple colour depths with 24bpp being
the default. Testing with more ... exotic ... display types would
probably be advisable.
This patch do this changes:
1. Change (add/remove) blank lines
2. Join some if's in only one
3. Change c++ comments to c comments (// -> /* */)
4. Add / remove curly brackets if needed
5. Change spaces (add/remove) & set correct indentation
When generating the full path+name of file to search for a file
being #included, it was generated in a buffer that's supposedly
large enough (MAXLINE > 2*PATH_MAX). However, this limit has a few
issues (PATH_MAX seem to be able to be bigger, and worse: we can't
be sure we're given longer args).
The code was rewrote to natively include boundary checks so we're
sure we won't overflow the buffer. A few strncpy have been removed
because in this case they tend to make things harder to write.
When a #include is used, the file is searched in many places; when
the file was searched in the ':'-separated list of path it could
led to infinite loop if the list contained more than one path and
that the file was not found in the first path, the ':' separator
was not properly passed over.
The Debian script for WindowMaker generates some warnings on startup
due to 'find' being cautious about its options, updated them to be
in line with expectations.
The functions wDefaultGetIconFile(), get_default_icon_filename() and
get_generic_value() use the argument noDefault in order to avoid searching the
default icon. This double negation is difficult to read though. This patch
changes it to be True if the default icon should be included or false if not.
This patch changes the noDefault argument to default_icon, then the
True is now False and False is True.
The main change is at get_generic_value():
- /* Search the default icon name - See noDefault argument! */
- if (!value && !noDefault) {
+ /* Search the default icon name - See default_icon argument! */
+ if (!value && default_icon) {
Because the functions wDefaultGetIconFile() and get_default_icon_filename()
mainly forwards the noDefault argument to get_generic_value().
The icon size should be set when the icon is created. Therefore the icon size
for net_icon_image should be set at get_wwindow_image_from_x11(), function that
creates the image.
The function scaleDownIfNeeded's code is duplicated with the code of
wIconValidateIconSize(), then this function can be removed. The icon
size in the switchpanel should be ~48 pixels, to allow the frame around
the icon. Then, we always should resize the icon to this size. The standard
icon size is specified in WPreferences.icon_size (usually 64 pixels).
It helps to understand why a file is present in this file, which
in turn can help to understand where a file comes from and it makes
updates to the file easier
The icons should be scaled if their size is "max_size - ~3 pixels".
Tipically, max_size is 64 pixels. With the current wIconValidateIconSize
the max size for an icon is "64 pixels", then the icon don't have border
and is ugly. This problem is in the dock icons and in the switchpanel.
The new wIconValidateIconSize function don't use the fixel value of 64
pixels, uses the argument max_size to get the final icon size. The icon
is scaled holding the aspect ratio and reserve ~2 or ~3 pixels to hold
left space to the icon border. Now the icon is inside the icon space,
with border.
This patch removes the preprocessor option of DONT_SCALE_ICONS, because
all the icons should be scaled if needed, to hold a beatiful interface.
The default icon (defIcon) is never used because the function wDefaultGetImage
always returns a icon (it searches the default icon). So we can remove the
defIcon variable, its creation and remove code.
The function create_default_icon sets the panel's default icon. This
is interesting to have a "cache" icon and don't search the default
icon more than one time if the icons don't have icon associated.
The code of create_default_icon was included in the addIconForWindow()
function, this patch only moves it to get a better (clear) code.
Now that the built-in parser has support for all the feature of CPP
being used by WindowMaker's default menu, we can remove the stuff
related to calling CPP:
- code for preparing and running CPP;
- compile-time option to de-activate the call to CPP;
- command-line option
It is now possible to use #ifdef/#ifndef to exclude some part of the
file. The implementation uses a stack to track conditionals so it is
possible to use nested constructs.
A number of macros are pre-defined by WindowMaker for CPP in the
function 'MakeCPPArgs', they are now available in the internal
parser too. CPP also had some predefined macros, a subset of them
have been added.
The definition have been split in two parts:
- the macro that are dependant on WindowMaker parameters are
defined by WindowMaker (src/rootmenu.c)
- those that are independant, which can be defined by the parser
itself (WINGs/menuparser_macros.c)
This adds support for defining new macros, with or without parameters, which
when found afterwards in the text are replaced by their definition.
The complex analysis for arguments replacement is done at macro definition
time, so it is done only once and the macro expansion will be fast.
The macro-related functions have been placed in their own file because it is
quite a complex task and we do not want filesize to explode, it is always
better to keep things human-sized.
The parser is prepared to handle '#' directives, starting with file
inclusion. The search path for the file are taken from what was
actually given to CPP. There is an arbitrary limit to the inclusion
nesting, which is actually not a design limitation but a security
to avoid infinite include loops.
Wrote a new parsing code that is able to skip over comments.
Note that because CPP is still active, you will not see any
effect unless you disable CPP pre-processing.
The default function used so far provides informations not so useful
to user, like wmaker's source file, line number and function; it
also cannot provide the line number from the parsed file because cpp
messes this information.
With this dedicated function we try to provide useful things which
are being tracked by the parser internally, like valid line number
and the name of the file being read (which can be convenient in the
case of #include, for which we may also be able to provide the
inclusion tree!)
From caller point of view, the two function have been merged into a
single function in the API. This will be needed by the advanced
parser that will have to not separate the concept of a 'line' and
the concept of 'content' (due to empty/comment lines, multi-line
comments, long lines split with '\')
All the information related to the file being parsed are stored in
a single place. The content of this structure is not visible to
caller to avoid messing the content; the parsing will be handled as
methods to this object.
Please note that all functions visible as part of the parser's API
are using the CamelCaseNotation to be consistent with the rest of
the API; however all internal functions use the non_camel_case_syntax
to follow the coding style set by Carlos for the project.
Due to the tasks to take in charge, the internal parser will grow in
size to support basic CPP feature, so it is a good idea to start by
moving the current functions into a dedicated file.
The code to find the icon in the function getWindowMakerIconImage()
is duplicated, because it is the same code as in get_default_icon_filename()
This patch includes the prototypes of get_default_icon_filename()
and get_default_icon_rimage() in defaults.h, so these functions
can be used in other files.
The function wDefaultGetImage() is splitted in two:
1. get_default_icon_filename(): This function returns the full
path of an icon. The function searches the icon in the database
using instance and class.
2. get_default_icon_rimage(): This function returns the RImage for
a given image path (full path). This function validates the icon
size, so the icon is fully usable.
The function get_default_icon_filename() now adds the .app icons in the
search (using wApplicationExtractDirPackIcon()). To do it, the command
should be included, because this function searches '"command".app' icons.
Setting the command to NULL, this case is not used.
To do it we need the function wApplicationExtractDirPackIcon() defined
at appicon.c, so we need set the function as non-static and provide their
prototype in appicon.h.
This patch also includes an extra pointer check at wDefaultGetStartWorkspace
to make sure that WDWindowAttributes exists.
This patch adds support for the _NET_FRAME_EXTENTS property as
described in the EWMH spec. With it I was able to use the compton
compositing manager to draw fully opaque windows with semi-transparent
titlebars and resizebars.
Set the _NET_FRAME_EXTENTS property based on border widths and
titlebar/resizebar heights.
The EWMH spec says:
"_NET_FRAME_EXTENTS, left, right, top, bottom, CARDINAL[4]/32
The Window Manager MUST set _NET_FRAME_EXTENTS to the extents of the
window's frame. left, right, top and bottom are widths of the
respective borders added by the Window Manager."
Update the saved X co-ordinate of a window which was moved when
maximized (only) vertically so that unmaximizing the window restores
its dimensions without warping it back to its previous X position.
Similarly update the saved Y co-ordinate of a window which was
moved when maximized (only) horizontally.
Handle Maximus as a special case. We remember the Maximusized X and Y
co-ordinates then adjust the restored co-ordinates relative to the
delta between the window's position just after Maximusizing and its
position just before restoring. So for example if a window is
Maximusized, moved 100 pixels to the left and restored, it will end
up 100 pixels left of its original geometry.
Also fix "jumping window" bug reported by Christian Wittmer:
This "jumping window" happens only when you
1) open new xterm (STRG +n)
2) current position is (+64, +0)
3) maximize window vertically
4) undo maximizing
5) move window to the right or left (up or down as you like)
I moved right to (+450, +0)
6) maximize vertically again
7) and undo maximizing
8) window jumps back to (+64, ..) position
If you move a new opened window to a new position (e.g. +200, +200)
and then start with "3)". window is jumping back "+200, +200"
The variables waiting_save_ack, dragged_while_fmaximized,
buttons_dont_fit, rebuild_texture and needs_full_repaint are not used,
so they can be removed.
Even though net_state_from_client was never being set, it was
being tested inside a
if (!wwin->flags.net_state_from_client)
which according to testing was always being true (probably because gcc
initializes it to zero). But this situation is confusing, so it's
better to remove the if() test altogether as that is the intention
if net_state_from_client is not explicitly set (which it isn't in
the current wmaker source).
This situation is analogous to:
[mafra@Pilar:c]$ cat init.c
int main(void)
{
int a;
if (!a)
printf("No a = %d\n", a);
return 0;
}
[mafra@Pilar:c]$ ./init
No a = 0
The functions related to colormap are moved from funcs.h to the new
file colormap.h. These files are included where needed.
The functions wColormapInstallRoot and wColormapUninstallRoot are
removed, because they are not used.
New functions
static void add_to_appicon_list()
static void remove_from_appicon_list()
to add or remove appicons from the app_icon_list, making the code easier
to follow.
LITE was removed in fe736e849c ("Remove LITE config option"), but
the check for it in wmaker.inst was missed.
Furthermore, remove confusing message about wmaker being configured for KDE.
The functions of main.c should be included in main.h, not in funcs.h.
This patch adds the main.h file and moves the function prototypes to
this file.
The not needed "include funcs.h" are removed.
The contents of the rootmenu.h file are not used and can be removed,
but the prototypes of rootmenu.c are in funcs.h and should be moved
to the correct (rootmenu.h) file.
Making all in src
CC actions.o
actions.c: In function ‘wMaximizeWindow’:
actions.c:421:14: warning: ‘maximus_height’ may be used uninitialized in this function [-Wuninitialized]
actions.c:440:29: warning: ‘maximus_width’ may be used uninitialized in this function [-Wuninitialized]
actions.c:442:35: warning: ‘maximus_y’ may be used uninitialized in this function [-Wuninitialized]
actions.c:454:18: warning: ‘maximus_x’ may be used uninitialized in this function [-Wuninitialized]
CCLD wmaker
These warnings were harmless because they were at a point where
find_Maximus_geometry() had already been called. So let's simply initialize
them to zero to silence gcc.
There are some code duplication in winmenu.c. Two new functions,
- open_window_menu_core
- prepare_menu_position
Join the common code and then the duplicated code can be removed.
Create a header file dockedapp.h to address a long-standing warning
which pollutes the compilation output with --enable-silent-rules:
Making all in src
CC dock.o
dock.c: In function ‘wDockDetach’:
dock.c:2118:3: warning: call to function ‘DestroyDockAppSettingsPanel’ without a real prototype [-Wunprototyped-calls]
dock.c:65:13: note: ‘DestroyDockAppSettingsPanel’ was declared here
CCLD wmaker
The superfluous.[ch] files have a few minor issues, this patch addresses some of them:
- Move the defines to the top of the superfluous.c file
- Include the ifdef NORMAL_ICON_KABOOM inside the DoKaboom() function
because the DoKaboom is used without the NORMAL_ICON_KABOOM ifdef in
other files.
- Include the ifdef WINDOW_BIRTH_ZOOM inside the DoWindowBirth() function,
therefore the function don't needs to be defined twice (with and whitout
WINDOW_BIRTH_ZOOM define.
- Now the functions are defined in superfluous.h and the externs are not
needed.
- We need include the dock.h in the superflous.h because is used by the
definition of MakeGhostDock().
- We need include the superfluous.h in window.c (removing the extern)
Fons Adriaensen reported that:
> WM's 'unshade' seems to use multiple Expose events [...] even when
> all animations and 'superfluous effects' are disabled.
Typo in shade_animate() could cause windows to be animated when the
no_animations preference was set. With this patch I see fewer
XMoveWindow() calls and quicker shading.
We want to avoid having to consider the position of an appicon whose
application has no_appicon set. The appicon is not "painted" so it
does not appear on the screen. But if the appicon is still added to
the list of application icons, the wArrangeIcons() function gets
confused and ends up creating holes in the icon positions which
correspond to the no_appicon apps.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
Having two similarly named functions for doing the same thing is confusing.
In order to account for the extra check done by the late wAppIconSave(),
add an argument "dock".
Now there's only save_appicon().
And as pointed out by Rodolfo kix García, the save_appicon_core() function
can be removed too.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
Function names are important and should not do more than their
names imply. In this case, create_appicon_from_dock() should only
try to get an icon from the dock (or clip).
If the icon is not there, do not try to make an icon from scratch.
You were told to create it from the dock!
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
Last group from a desktop file is always added to menu regardless of
NoDisplay property. Properly handle the NoDisplay and Hidden properties
at the end of file.
Signed-off-by: Alexey I. Froloff <raorn@raorn.name>
The icon painting is moved to the function makeAppIconFor()
including the check for no_appicon.
wAppIconCreate is now static because is only used in makeAppIconFor()
The path of the CacheIcons is defined at CACHE_ICON_PATH. The path
~/GNUstep/Library/WindowMaker/CachedPixmaps can be changed in a future
and the comment is incorrect.
The function extractIcon is removed, and its behaviour is moved to
save_app_icon, because this function is always called.
To do that, save_app_icon doesn't need to check if the icon is saved (it's
done at wIconStore). The icon now is saved always (if it doesn't exist),
like extractIcon() used to do.
When the application is created, the WAppIcon now is created always,
but it is only painted if the flag is not set.
The icon initialization to NULL can be done now at app_icon_create_from_docks
because it is always called.
The function app_icon_create_from_docks is moved to appicon.c, because
it is a function to create application icons.
The static function findDockIconFor is moved too because it is only used
at app_icon_create_from_docks().
The function wAppIconSave is splitted in two functions:
wAppIconSave() + save_app_icon_core()
The function save_app_icon_core will be used in other functions as
common code.
The Bool value is not used, therefore it suffices to return void.
wAppIconChangeImage is removed from appicon.h, because it is not implemented.
Small code cleanup for wAppIconCreateForDock at appicon.h.
The icon functions wApplicationExtractDirPackIcon() wApplicationSaveIconPathFor()
are moved to appicon.c from application.c.
This functions are Application Icon related and should be included in appicon.
The function get_name_for_icon is renamed to get_name_for_wwin.
The previous contents of get_name_for_icon are now at the new
function get_name_for_instance_class.
Center strategy: try to put window at the center of the usable area. If
window would overlap with existing windows, fall back to "Auto"
placement strategy.
It's very useful for fresh workplaces.
The function get_name_for_icon returns now the name of the icon,
without the full icon path and without extension (.xpm). Now is
not static.
The full path, including the folder creation, is done now by
wmkdirhier() (WINGs). This function is much better, because
supports "infinite" folders, not like the old get_name_for_icon
which could only create the specific folder for Cache Icons.
Using this functions, wIconStore() do the same work, but in a better
way, more clear.
This patch creates some functions:
1. Rename getnameforicon() to get_name_for_icon()
2. New function get_icon_cache_path, to get the icon cache folder
($HOME + GNUstep/Library/WindowMaker/CachedPixmaps). This folder
is defined now in the preprocessor. Not used yet, in next commits.
3. New function get_wwindow_image_from_wmhints to read the image from
X11 wmhints. Previous code at wIconStore() now changed.
The function wDefaultFillAttributes can be changed a lot:
1. Initialitation to NULL: If the pointers are initialized to NULL
then, the "if's" don't need the else block:
WMPropList *value, *dw, *dc, *dn, *da;
dw = dc = dn = da = NULL;
if's:
= if (instance)
= key2 = WMCreatePLString(instance);
- else
- key2 = NULL;
2. Added StrConcatDot in the class + instance block:
= if (class && instance) {
+ buffer = StrConcatDot(instance, class);
- buffer = wmalloc(strlen(class) + strlen(instance) + 2);
- sprintf(buffer, "%s.%s", instance, class);
3. init_wdefaults(scr); moved above. This function is used only
to load the default value "AnyWindow" (value "*"), to search
the default value. Can be moved above without problems.
4. Preprocessor code of APPLY_VAL moved to the top of the file.
5. New function get_value_from_instanceclass() to do the rest of
the (repetitive) code. This function is called to create the
proplist, search the value, and return the proplist.
EXTRA:
1. Added StrConcatDot (like dot 2) in wDefaultChangeIcon()
2. Added a comment in get_value()
The argument "src" is removed in the function get_generic_value()
because it is not used.
The function wDefaultGetIconFile() is moved close to
wDefaultGetStartWorkspace() because both are similar.
I noticed that I can have different background with wmmsetbg for each
desktop and wanted to use it, but it was missing option to fill the
screen with proportionally scaled image, so here it is.
The function removeAppIconFor() does exactly the same thing as the code inside
wApplicationDestroy(), including the last call to wArrangeIcons(). So let's
simply remove the code and call the function instead.
This patch checks if the no_appicon flag is set, then, the icon
is not created, therefore cannot be painted neither saved.
Some comments added to clarify the code.
The functions makeAppIconFor and removeAppIconFor are moved from
winspector.c to appicon.c, and now are not static.
The reason to move these functions is because are used to create
and remove app_icons, therefore should be included in this file.
Finally, this functions will be updated and splitted in next patches.
The new function app_icon_create_from_docks do all the needed stuff to create
the application app_icon using a previous image in docks.
The contents of the app_icon_create_from_docks were moved from the
wApplicationCreate function. Now this function is more clear.
New function to save the icon "save_app_icon". This function is
created in appicon.c, because is related to app_icons. The contents
are from application.c. No important modifications are included in
this function. Removed the includes not needed.
Due to variable size definitions in the structure, the 'for' loops may
actually try to work on wrong data. Instead of using hard-coded value,
we now simply let the compiler give us the number of elements.
This may (or may not) fix a crash reported by Rodolfo kix Garcia.
readMenuPipe() was calling freeline() on stuff which might be uninitialised, causing
a crash if no valid input was read.
readMenuPipe() was trying to snprintf() on an uninitialised pointer. We now use a
fixed-length buffer like the other readMenu*() functions.
Various memory leaks in readMenu*() functions have been fixed. There
are still some lurking around but most have been removed.
The original report from Charles Philip Chan <cpchan@bell.net> on 22 May 2012 says
that:
"Window Maker crashes when I try to open a sub-menu auto-generated by using:
xdg_menu --format WindowMaker --charset UTF-8"
There was also a report from Amadeusz Sławiński <amade@asmblr.net> on 24 May 2012
stating that wmaker crashes using:
% cat test.sh
cat << EOF
Test MENU
stuff EXEC true
Test END
EOF
% grep test GNUstep/Defaults/WMRootMenu
("Generated Submenu", OPEN_MENU, "|| /home/amade/test.sh")
Error is:
wmaker(MonitorLoop(monitor.c:134)): warning: Window Maker exited due to a crash (signal 11) and will be restarted.
This patch includes the changes to create the package for
debian.
Changes are included in changelog:
* New upstream release 0.95.3
* Removed debian/clean file. Upstream removes now the files.
* Bumped to standard version 3.9.3
* Important!: Removed symbol "W_DraggingInfo" in libwutil2 and
libwings2, because the struct W_DraggingInfo is now declared
as "typedef", therefore the struct is not included.
This change is included in upstream, see line 126 of the file
WINGs/WINGs/WINGsP.h
* Patch 53_Debian_WMState.diff changed, because the WMState file
in upstream is now different. Now, the dock launch WPrefs.
* Removed /etc/X11/WindowMaker files.
* Removed from debian/wmaker-common.dirs
* Removed (duplicated) files in debian/wmaker-common.install
* New path for menu.hook: /usr/share/WindowMaker
* Changed in the menu-method files (wmaker and wmappearance.
* Removed menu.posthook and menu.prehook.
* Changed the menu behaviour. Applications/* contents moved to
the root menu.
The test suite was added, but it wasn't included in the list of
directories to be included in the tarball, so ./configure fails:
config.status: error: cannot find input file: `test/Makefile.in'
The following patch adds the missing test suite.
This diff fixes a huge amount of issues that could be triggered by using
the old-fashioned configuration files, i.e menu. They can be activated
by using a line like "/path/to/menu". On most systems, this file will be
parsed with cpp and the result sent through a pipe to WindowMaker.
Anyway, in the code path, memory leaks and buffer overruns await. I have
tried to fix these parts, but in the end it is more or less a rewrite,
whereas I used WINGs whenever possible.
Difference to previous implementation, beside of bugfixes:
- You are free to use single quotes and double quotes in configuration
- Linewrapping is allowed for other pipe'd data, too
- In general, line reading is unified throughout that file
Bugfixes:
- Avoid buffer overrun on lines which have line wrappings, that will happen
ALWAYS, because wtrimspace-usage was erroneously.
- Avoid memory leak, also happened through wtrimspace usage.
- Avoid buffer overrun in separateline if a line ending happens early.
- Actually handle ferror() instead of only feof(), that would result in
endless spinning
- A line wrapping at the end of a file is an error.
The function wtrimspace returns a wmalloced memory region, therefore
it must be wfreed after usage again.
Please take note that this also happens with every single line
of menu files, which will get fixed in a separate commit.
In parseMenuCommand, title[300] might get filled with a string of length
300. The string is copied with strcpy, therefore the size would have to be
301 or -- as I propose -- the fixed value 300 gets replaced with
"sizeof(title) - 1". This shows also that the size 300 belongs to title
and it will already be replaced during compile-time into 299.
memset is the last function call in wmalloc, just before it returns the
newly allocated memory. Therefore it is not needed to call it again
after wmalloc call. Although I would prefer to switch wmalloc to a
calloc-based wcalloc function, the compatibility of WINGs for old apps
should be kept.
Avoid some pitfalls with window maximization and make it behave more
intuitively. We now treat a window's vertical and horizontal
maximization as separate properties and only remember its original
geometry in a particular direction when it actually changes. We also
deliberately do not remember a window's geometry when it changes from
one maximized state to another. As a result windows can be more
reliably restored to their original size.
For example the "Maximize active window" hotkey followed by the
"Maximize active window vertically" hotkey will now result in the window
being maximized horizontally only, whereas previously the second hotkey
would have no effect because the window was already maximized
vertically. In addition selecting the Unmaximize window menu in the
same example will now result in the window being restored to its
original size. Previously the unmaximize attempt would have no effect
because the vertical maximization would have remembered the window's
"original" geometry when it was fully maximized.
Maximus is handled separately. The "Maximus" hotkey will now toggle
Maximus mode regardless of the window's current maximization state. For
example if two unmaximized windows are on screen and one is Maximusized
it will fill the space left by the second window, as before. But if the
first window is maximized and the "Maximus" hotkey is pressed the window
will now fill the same space as if it were Maximusized from its original
size. Previously the window would not change size from its fully
maximized state because the Maximus algorithm would consider
fully-maximized to be a valid Maximus size.
Xrandr support is currently very primitive and stupid, but people
do not seem to be aware of its problems and still enable it in distros.
Try to scare them to not do that.
This reverts commit cbe2f4e61a.
The reason for the revert is the regression reported here:
http://marc.info/?l=openbsd-ports&m=133151145814675&w=2
I've been able to reliably reproduce the double-spawning of
xterms I've been seeing. The bad news is that it still happens
even with a fresh configuration directory.
To reproduce:
0) Create a keyboard shortcut if one does not already exist (I'm
using <ctrl>+u set to spawn an xterm)
1) Open WindowMaker's preferences.
2) Click on "Applications Menu Definition" (it's the icon
immediately to the left of the keyboard icon)
3) Click save, and then close.
4) Keyboard shortcuts will spawn two applications.
I use keyboard shortcuts to spawn xterms, browsers, and filemanagers.
All of them spawn twice after performing the above. To fix it,
<F12> or <right-click> -> Exit -> Restart.
If I repeat steps 1-3, three windows will spawn. Repeat again and
four windows will spawn etc. (I stopped after getting five to
spawn at once).
There is a bug which happens when one modifies the "Initial Workspace" in the
"Attributes..." menu. Upon saving the change wmaker segfaults.
Commit ad373ef0dd ("WindowMaker: New function wGetWorkspaceNumber")
used wfree() in variables which were not wmalloc'ed. Fix this.
The declaration of the structure actually also created an unused variable. This
variable was not used anywhere, and lead to symbol defined in multiple objects.
These symbol are silently merged by GCC, thus no problem was reported.
The issue was raised by Yves de Champlain when trying to compile with LLVM/clang
which is a bit stricter there.
The new function wIconCreateCore do the common work to create a WIcon.
This code was before in wIconCreateWithIconFile() and wIconCreateWithIconFile().
The new function wGetWorkspaceNumber returns the workspace number
for a given workspace name.
The code of this function is already used in session.c and wdefaults.c
and now is moved to workspace.c
In wSessionRestoreLastWorkspace the char value is checked before
calling the function, because without string, the function don't do
nothing.
This patch removes the unneeded curly brackets in multiple files.
It also add some comments in the code. In usermenu.c removes some
variables not used.
Changes to appicon.h:
- Spaces replaced by tabs in the struct
- Removed unused variables:
- client_machine, never assigned
- gnustep_app, never used
- Functions sort by return type
Then, in dock.c, the block with client_machine is never true,
then can be removed. The host variable then is never used.
Use the gcc-specific noreturn attribute to advise the compiler that
Exit() doesn't return. This is safe because we already use preprocessor
rules to hide __attribute__ from other compilers.
Suggested independently by Christophe Curis and Brad Jorsch.
main.c: In function ‘RelaunchWindow’:
main.c:461:1: warning: control reaches end of non-void function
[-Wreturn-type]
Actually we can't reach the end of the function because the three
possible cases are Exit(-1), return False and return True. Of course if
Exit() were ever changed the above statement might become incorrect.
Some compiler food silences the warning.
Zoltan Balaton points out that Control + Doubleclick on docked app will
launch a new instance. For consistency with that behaviour we now allow
Control + Doubleclick on an undocked appicon to invoke the new
relaunching functionality.
We also now restrict doubleclick handling to the left mouse button in
order to avoid relaunching the application twice when the middle button
is used.
Use the WindowRelaunchKey shortcut to examine the WM_COMMAND property of
the active application's main window and launch a new instance of the
application using the retrieved command line.
The function canReceiveFocus should check if the window is no_focusable first.
If the window is not focusable but is miniaturized, the window is shown in the
switchpanel.
How to reproduce the problem:
- Open an application
- Open the window properties, advanced options
- Set that the application can not get the focus and save
- Test the switchpanel (Alt+Tab) the window doesn't appear
- Minimize the window
- Test the switchpanel (Alt+Tab), the window appears
- If the window is selected (restored), and test again the switchpanel, the window doesn't appear!
A new function is created to get the image from the X11 window.
The function updateIconImage is splitted in two blocks, one is moved
to get_wwindow_image_from_x11 with the X11 code, and other is used
to update the Application and Window icons (kept in the same function).
Doug Barton reported that using "Technical drawing-like" for "Size display"
under WPrefs "Miscellaneous Ergonomic Preferences" together with opaque
resize would lead to artifacts left on screen when resizing.
This patch fixes it (and it also fixes some minor coding style issues
in the surroundings).
The file "clean", which contains "WindowMaker/Defaults/WMRootMenu"
can be removed, because this file was included to be cleaned in
the commit 01907f9983
The Application/* items are moved to the root menu, removing the
"Applications" submenu. The files from /etc/X11/WindowMaker moved
to /etc/GNUstep/Defaults
(Window|Menu)Title(Min|Max)Height defaults options allow to set
minimum and maximum titlebar height.
For example, to force all titlebars to 24 pixels execute
following commands:
$ wdwrite WindowMaker WindowTitleMinHeight 24
$ wdwrite WindowMaker WindowTitleMaxHeight 24
$ wdwrite WindowMaker MenuTitleMinHeight 24
$ wdwrite WindowMaker MenuTitleMaxHeight 24
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
It was removed on 67a8a82670 with the assumption that
nothing was using it. But that was not really the case - FSViewer
used it.
I've just tested it. After a trivial fix regarding the change in
the function definition of WMWritePropListToFile(), FSViewer
compiles and even seems to work (didn't test much though).
So let's not be unfair with FSViewer and put wmlib back. FSViewer
might even be used for educational purposes for people wanting to
write apps using WINGs etc.
The path /etc/X11/WindowMaker is not needed now. The files were
deleted or moved to /usr/share/WindowMaker. Some files were the
same in /etc/X11/WindowMaker and /usr/share/WindowMaker.
The configuration file menu.hook is moved to /etc/GNUstep/Defaults
Duplicated:
- wmmacros
- appearence.menu (copied & generated).
- background.menu
Removed:
- menu.prehook
- menu.posthook
Moved:
- menu.hook (generated).
I have noticed that in WindowMaker 0.95.2 application icons bounce
when some actions are done, like starting a program, etc.
Disabling Superfluous or Animations also disabled the minimizing
animation which I am used to, so doing that didn't do the job.
Everyone has got his or her own tastes, but I did not like the new
behavior and didn't find any way to disable it without affecting other
things.
So I made a patch to fix just that. It adds a new option in WPrefs.app's
Expert Tab called "Do not make AppIcons bounce" which when enabled,
disables any type of bouncing for Application Icons, restoring the
old behavior.
Bouncing stays the new default behavior.
The only place where this function is called is from a double click
in the first icon of the dock, and only if there's no program already
associated with it.
This is a bit superfluous and most people have defined the first
icon to call WPrefs instead and end up never seeing that panel.
And since the last commit ("Change behaviour of the GNUstep dockapp"),
this is now also the default behaviour of Window Maker.
Furthermore, the panel itself is not accurate. Window Maker is not part
of the GNUstep project.
Make it launch WPrefs instead of displaying GNUstep information.
GNUstep information under a dockbutton is unneeded and it's commonly replaced
by WPrefs. This also is giving one extra space for another dockapp.
WMCreateFont() was calling exit() if it could not create the
font, and was trying too hard not to return NULL.
Just return NULL if the font could not be created instead of exit()ing
and let callers decide what to do upon failure.
Thanks to Christian <chris@computersalat.de> for reporting this.
Icons in the switchpanel are constrained to the value of the IconSize
preference but the grid in which they are arranged is fixed at 64 pixels.
If IconSize is less than 64x64 the panel will show smaller icons with a
wide spacing, which looks pretty stupid.
Fix it by forcing the switchpanel to attempt to load images at the size
it's going to use. The icon it actually gets may of course still be
smaller.
Display a standard Window Maker balloon with the workspace name when the
mouse enters the clip area.
This complements commit 4954d4df23 ("clip: Do not display balloon with
workspace name") because now the balloon which appears on the clip is the
same as in other parts of wmaker.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
When the mouse passes over the clip, wmaker would display a "odd" balloon
text with the workspace name, but the balloon itself was covered by the
clip icon!
So if the workspace name was short enough ("Internet" is, by my testing here)
the user wouldn't see anything, the balloon is completely under the clip icon.
I found this issue because one of my workspaces is called "Beyonder" and I
noticed a small "r" under the clip one day.
Instead of trying to fix this, I just removed the whole thing about displaying
the balloon because it is superfluous and I haven't seen any bug reports about
this yet, so it probably means most people are not even aware of it.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
GRADIENT_CLIP_ARROW was never defined anywhere and having fancier clip
arrows is not something particularly interesting, so let's simply remove
the code.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
Replace the ARROWLESS_KBD #define with the ViKeyMenus preference.
When ViKeyMenus is TRUE, users can type h/j/k/l to scroll around menus.
Since ARROWLESS_KBD was previously undefined by default, ViKeyMenus is
FALSE by default.
As reported by Paul Seelig, it used to be the case that
getstyle -t ~/GNUstep/Library/WindowMaker/Themes/somefile
would save the current theme, and all old menus (WMRootMenu) were
relying on this.
The problem was that the following piece (from commit 6bf79945)
if (style_file && !make_pack)
print_help(0, 1);
would allow a style_file to be specified and saved to only with the
option -p (which implies make_pack), therefore saving a theme with
-t like the root menu used to do no longer worked.
Now things work fine:
[mafra@Pilar:~]$ ls GNUstep/
Applications/ Apps/ Defaults/ Library/
[mafra@Pilar:~]$ getstyle -t ~/GNUstep/theme
[mafra@Pilar:~]$ ls GNUstep/
Applications/ Apps/ Defaults/ Library/ theme
But also note that trying to save a theme outside of GNUstep/
is not allowed and it prints no error message - perhaps it should...
In commit e1453087f57... the function copy_file() was renamed to
wcopy_file(), then the symbols file needs to be updated.
The changelog includes a new debian bug closed.
...in order to avoid clashes that happen during compilation of
wmakerconf.
This is a new function in WINGs, so renaming it at this point is
not a big deal.
Thanks to Rodolfo García for the heads up.
This upload includes the debian changes of 0.95.0 and 0.95.1
- Debian stuff (themes, menus,...) is moved to different folder.
- Many changes in debian/rules file
- New files to avoid lines in debian/rules (new menu files)
- This is a little extract form the debian/changelog file:
* New upstream version 0.95.1
* The WINGs's file proplist-compat.h is removed in upstream.
- Removed the line in debian/libwings-dev.install
* Updated debian/libwutil2.symbols with new symbol.
* libpng12-dev dependencies changed to libpng-dev. [Closes: #648123]
* wterm package suggestion removed.
* Menu shows "Run..." option. [Closes: #165075]
Thanks to Andreas Tscharner for their patch.
* Menu shows the background files [Closes: #655122]
* Added patch 54_Debian_wmmacros.diff.
Based on changelog: Marcelo E. Magallon Tue, 17 Nov 1998
* Xterm and WMPrefs are now Debian specific.
* Added patch 53_Debian_WMState.diff.
Based on changelog: Marcelo E. Magallon Sun, 26 Nov 2000
* Fix wmaker-common dependencies. [Closes: #654668]
* Manpages moved from wmaker-common to wmaker (Lintian problem).
* Removed old stuff in wmaker.post* and wmaker-common.post* about
update-alternatives.
* Fix to the FTBFS. [Closes: #654524]
* New debian/watch file
* New upstream version 0.95.0, now from git. [Closes: #401900]
[Closes: #514438, #607550, #218110, #583734, #105351, #549157]
[Closes: #283610, #311563, #310285, #329783, #280819, #284048]
[Closes: #292391, #361241, #364290, #148370, #287459, #122076]
[Closes: #175503, #79598, #78088, #68381, #38184, #41434, #41434]
[Closes: #94960, #39543, #63265, #69499, #94446, #77488, #329783]
Thanks to Andreas Tscharner for their bug revision.
* This new version is based in wmaker-crm a wmaker fork, because
wmaker (original) is not updated.
* New debian/rules file. [Closes: #590244]
* Many many changes
* /usr/lib/WindowMaker/WindowMaker is now /usr/lib/WindowMaker/wmaker
* wmaker script launch now /usr/lib/WindowMaker/wmaker
* New maintainer. [Closes: #632875]
* New package wmaker-common (arch independent files).
* Removed the asclock diversions from the wmaker install scripts
wmaker.postrm and wmaker.preinst because asclock binary is not
included in wmaker package (see asclock package).
* New package wmaker-common with the arch independent files.
* debian/patches are now DEP-3.
* debian/copyright is now DEP-5.
* Bumped Standars-Version 3.9.2.
* Manpages moved to upstream.
* Solved problems with .la files (lintian clean).
* libwmaker0-dev isn't included, because was removed in upstream.
With the fonts I use here, the memory information line in the "Info panel"
did not fit inside the window. So make that line a bit shorter by displaying
Total memory allocated: 2600 kB (in use: 2182 kB)
instead of
Total allocated memory: 2600 kB. Total memory in use: 2182 kB.
Furthermore, the surrounding code was a bit convoluted to display either
"Additional support for: WMSPEC"
or
"Additional support for: WMSPEC and MW"
As WMSPEC is always present and it doesn't seem we are adding more stuff
to support, one can do the same with a shorter code which reads a little
bit better.
Despite having the inotify mechanism compiled in my WM, everytime I
changed a keyboard shortcut definition in the WMRootMenu via WPrefs
I was required to open the WMRootMenu (eg with the F12 key) for the
change to take effect...annoying.
So explicitly reload the key bindings when a modification inside
~/GNUstep/Defaults/ is detected. As the inotify mechanism calls the
function wDefaultsCheckDomains() when that happens, let's add a call
to rebind_key_grabs() in there.
Now it works fine and changes are automatically in effect once WPrefs
writes the menu.
PS: rebindKeygrabs() renamed to rebind_key_grabs()...
The function captureShortcut() is implemented in both files KeyboardShortcuts.c
and Menu.c, with just a minor difference regarding the conversion to upper case.
To unify them, define a new function which includes a new boolean paramenter to
dictate whether the upper case conversion should be done or not.
There's no need to have a private function while there's one in WINGs.
Besides that, it does not remove trailing whitespaces appropriately as I
just tested by adding trailing space in the shortcut captured by WPrefs.
It is not trimmed before saving it:
[mafra@Pilar:Defaults]$ grep CloseKey WindowMaker
CloseKey = "Mod1+C ";
Using wtrimspace() fixes that and even saves 208 bytes of code:
[mafra@Pilar:WPrefs.app]$ size KeyboardShortcuts.o.*
text data bss dec hex filename
7703 0 0 7703 1e17 KeyboardShortcuts.o.new
7911 0 0 7911 1ee7 KeyboardShortcuts.o.old
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
There is code duplication with the cropline() function, so get rid
of it and use WINGs wtrimspace() instead.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
On my PowerPC Debian Squeeze System, the icons are colored
incorrectly. This patch removes the swapping of the data on Big Endian
systems, thus causing the icons to be colored correctly.
The data appears to already be in the native endian format.
After using WINGs wInputDialog() to handle renaming the workspace
upon Ctrl+left click on the workspace menu, all functions in src/text.c
become unused.
As pointed out in the comments in
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=304480
it's not possible to rename workspaces with non-ascii characters -- more
precisely characters which require deadkeys with your particular keyboard.
According to the investigation made by Rodolfo Peñas:
http://lists.windowmaker.org/dev/msg02529.html
the fundamental issue behind this bug is the fact that XLookupString can't
handle deadkeys.
So either we do something radical or live with the bug.
However -- as Rodolfo also pointed out -- renaming the workspaces via the
Clip works with non-ascii chars. The reason is that the Clip uses a input dialog
function from WINGs to handle the renaming.
So instead of doing the low-level handling of text inside the menu in order
to rename workspaces, just fire up the same WINGs input dialog to handle it.
It works and the old function editEntry() can be removed.
Furthermore, it makes the whole set of functions in src/wtext.c useless, and
they are removed in the next patch.
Overall, 600 lines of code are gone now.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The summary now looks like:
Window Maker was configured as follows:
Installation path prefix : /usr/local
Installation path for binaries : /usr/local/bin
Installation path for libraries : /usr/lib64
Installation path for WPrefs.app : /usr/local
Supported graphic format libraries : XPM PNG JPEG TIFF builtin-PPM
Antialiased text support in WINGs : yes
Xinerama extension support : yes
XRandR extension support : yes
Translated message files to install : None
I want to see the library line in order to avoid forgetting that
I should put them in /usr/lib64 (and not in the defaul /usr/local/lib)
There are some problems in the alpha channel support, as is
reported at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=72917
This patch add a new RCombineAlpha function, based on Gimp. This
function is called when needed in the raster.c functions.
This patch is based on the Brad Jorsch <anomie@users.sourceforge.net>
patch for the 0.62.1-0.1 version.
[crmafra: v1 was sent by Rodolfo kix Garcia <kix@kix.es>]
The statement under "if (style_file)" in the else branch of
"if (make_pack)" will never be executed because getstyle already
prints out a help message and exits before that.
So remove that code.
Instead of using a temporary buffer to store a "reason" string
which is later printf()'ed by abortar(), use wwarning() directly
and do the small cleanup done by abortar() on the spot.
As this was the only call site for abortar() it can now be removed.
The resulting object code gets smaller as a side effect.
Thanks to Christophe <christophe.curis@free.fr> for finding a mistake
in the first version of this patch.
This is essentially the fetchFile() from wcolorpanel.c from the last
commit, but renamed to a better name.
This patch just adds the function to the lib. Nobody uses it yet.
The idea is to use the fetchFile() in getstyle.c and in wcolorpanel.c
instead of using two very similar functions.
In order to do that, let's move the most generic one (fetchFile()) to
libWUtils, and this is the first step.
Dario Niedermann reported on 27.11.2011:
Selecting "Save Theme" from WMaker's root menu results in dialog box
asking for theme name, but then nothing is saved.
This feature used to work in 0.92.0.
and the reason for it (if he uses a WMRootMenu generated by wmgenmenu) is
the incorrect use of the option "-t" instead of "-p" to getstyle.
Furthermore, let the invocation of getstyle reflect the new default location
for the themes.
When invoked as
getstyle -p NewTheme
Instead of saving the theme inside the topdir ~/GNUstep let's save it
insinde their proper directory ~/GNUstep/Library/WindowMaker/Themes.
There is a problem with getstyle invoked with no arguments:
[mafra@Pilar:util]$ ./getstyle
Usage: getstyle [-t] [-p] [-h] [-v] [file]
or with -t:
[mafra@Pilar:util]$ ./getstyle -t
Usage: getstyle [-t] [-p] [-h] [-v] [file]
In both cases it is supposed to write to the standard output:
[mafra@Pilar:wmaker.git]$ getstyle -h
Usage: getstyle [-t] [-p] [-h] [-v] [file]
Retrieves style/theme configuration and output to FILE or to stdout
This regression was caused by commit 6bf7994520 ("style Stuff up").
When that commit did
argc -= optind;
if (argc != 1)
print_help(0,1);
it excluded the output to stdout as valid, because in this case
argc - optind is zero (see the manpage of getopt(3)).
The correct handling is to set the style_file only when there is
one non-option ARGV-element (argc - optind == 1) and print the
help message when there are more than one non-option element.
Applying patches without needing to fix them first helps me save time,
and this file (stolen from the linux kernel documentation) contains some
valuable tips to configure email clients to not mangle the patches you send.
The goal is that patches are sent in the body of the email (to ease review)
and be applied directly with 'git am'.
I personally use mutt as mail client and xjed as an editor. I insert the patch
in the body of the email using xjed's "Insert file" menu operation. That
works fine, but this documentation file helps you with more choices.
I use xrandr to decrease the resolution of my display when I connect
it to my TV. When I change the resolution back to my monitor, a few
of the icons at the bottom of my dock are deleted.
This happens because wmaker computes the maximum number of dockapps
which the dock can hold based on the screen resolution:
icon_count = scr->scr_height / wPreferences.icon_size;
and drops the dockapps above that number (in wDockRestoreState()).
But now the resolution can change via xrandr, so the above computation
can lead to dockapps being dropped when the new resolution is smaller
than it used to be.
To fix this it's enough to have a resolution-invariant number of allowed
dockapps.
If you delete the file ~GNUStep/Defaults/WMRootMenu and run WMPrefs
and go to the Menu Edit tab, it will show a dialog about keeping the menu or
discarding it. If you discard it then WPrefs shows an error:
"Could not open default menu from
'/home/user/GNUStep/Library/WindowMaker/plmenu"
The problem is at WPrefs.app/Menu.c:1424, and it happens because it tries
to load the default menu from the user's GNUstep home, but plmenu is not
copied there during the installation.
Of course, the file doesn't need to be copied to the users home, so
WPrefs should try to read it from /usr/share/WindowMaker/plmenu (Debian)
or /usr/local/share/WindowMaker/plmenu (upstream).
* New maintainer [Closes: #632875]
* New upstream version, now from git. [Closes: #556677]
* New package wmaker-common (arch independent files)
* removed the asclock diversions from the wmaker scripts
* new package wmaker-common with the arch independent files
* debian/patches are now DEP-3
* debian/copyright is now DEP-5
* Standard-Version is now 3.9.2
* manpages moved to upstream
* solved some problems with .la files (lintian clean)
This reverts commit f4890b17e6.
It turns out that I needed some functions from wtext.c to develop
a WINGs front-end to my comic book collection MySQL database.
Conflicts:
WINGs/Makefile.am
WINGs/WINGs/WINGs.h
With the addition of an extra button to the 'Advanced Options' window,
the vertical space of the widget no longer suffices. Increase it from
350 to 360 and adjust other parameters accordingly.
And let's write the button positions as 'PHEIGHT - 40' instead of using
a hardcoded value, avoiding the need to change it everytime PHEIGHT is
changed.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
The flag to not let windows be minimized was already defined in
WM (it's called 'no_miniaturizable') and the minimize functions
respect it. However this flag can not currently be set manually.
As can be read in the NEWS file
"NotMiniaturizable option changed to NoMiniaturizeButton"
it seems that the "non miniaturizable" property was a first-class citizen
in ancient times and probably configurable (those changes do not appear in
the old CVS history).
Let's make this property be user-configurable through the "Advanced Options"
panel.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
Before this change some lines in the "Advanced Options" window would not
fit into one line and would screw up the others. This depends on the fonts
used, but increasing the width of the panel a bit does not hurt.
The one in WPrefs.app/Appearance.c is indirect: the function might
be inlined and might return an uninitialized value, which causes gcc to
complain that the caller might use that uninitialized return value.
The problem with f65b99e615 was that it removed the call to
newWindow() along with the unused variable, without considering that
newWindow() has side effects that need to happen even if the return
value is ignored. So as long as we keep the newWindow() call, we can
safely remove the (unused) variable.
There are were a few uses of 'strncpy' that could lead to a missing NUL,
resulting in possible garbage being displayed. As suggested by Tamas,
use 'wstrlcpy' instead
Instead of using _old_ and hard-coded predefined menus as the
source for the user's WMRootMenu during installation, generate
the menu automatically using wmgenmenu.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
wmagnify stopped working after f65b99e615 ("Remove warnings").
So revert that change in util/wmagnify.c, as it turns out that the
gcc warning was bogus there.
This may ease Carlos' work slightly when dealing with people like me that could not find it and thus submitting the patch directly to him, the only address found
PropGetWMClass()
- XAllocClassHint()s a struct for class hint data
- This is filled by XGetClassHint(), which in turn uses Xlib to allocate
some more space for XClassHint members
- Upon XGetClassHint() failure, "default" is libc malloc'd (via strdup),
and is returned to the caller
- Upon XGetClassHint() success, XClassHint members are returned raw --
these members must be freed with XFree() (see XAllocClassHint(3))
- Thus it's up to PropGetWMClass() callers to decide (based upon the return
value) which method (libc free() or XFree()) to use to free res_name
and res_class. This was done nowhere, thus leaking some memory
on every failed PropGetWMClass() call.
- So just strdup the successful res_name/res_class members, XFree() them
while still in PropGetWMClass(), and allow callers to unconditionally
libc free() whatever PropGetWMClass() returns.
By turning M'bert's d6c134 around a bit and adapting the surroundings,
allow _NET_ACTIVE_WINDOW only if fulfilling it doesn't cause annoying
unwanted changes in the workspace. This is now the default behaviour;
unconditional focus stealing can be enabled on a per-client basis in the
Advanced Options window menu ("Focus across workspaces").
With Xinerama, on heads other than the primary, certain parts (menus,
scrollbars) of certain types of clients (Qt) are incorrectly or not at
all drawn.
The fix follows other window managers in completely ignoring the
_NET_WORKAREA property.
Problem originally poked at by Ambrus Szabo, correct keyword and the
eventually implemented solution suggested by Lucius Windschuh, typed
up by me.
From the original report by Tamas Teves:
"i'm having (and have, for a long time) problems with openoffice and
java-based (perhaps only netbeans-based?) apps.
for openoffice, any menu opens in the far right edge (with xinerama,
on the far right edge of the rightmost display) of the display. it
then can be clicked and operated corretly where it appears.
as for netbeans (and jswat, which also uses netbeans platform as a
base), it's a bit more complicated, but let me try to describe. start
netbeans so that it is not maximized and not in the upper left of the
display. it works fine. move it around, still works fine, resize,
still works fine. now if i maximize it (not by moving to the upper
left corner and resizing, but by the maximize shortcut), then the
menus start acting weird. it seems that the actual position of the
mouse pointer and the hot spot of the mouse pointer get "out of sync"
so to speak, as if the hot spot stayed where the window was before
maximization.
i've made a capture which hopefully better shows what i'm trying to
say. this is a netbeans window maximized, while i'm trying to click
tools->options. the left mouse button must always be held down, or
else it instantly loses any interaction with the menus (in the capture
i released it once when i completely lost track of where i might be).
the capture is available at http://dawn.dev.hu/~ice/tmp/wmjava.avi
i'm not sure whether java stuff other than those based on netbeans
platform exhibit this behaviour. i've seen this with at least one
other nb platform stuff as well."
The fix is to call wWindowSynthConfigureNotify() after wWindowConfigure()
* Remove assigned but not used variables (GCC 4.6)
* Bump _XOPEN_SOURCE to 600, ridding of FreeBSD warnings (this probably need
to be tweaked on a per-implementation basis as problems arise)
Some applications running in my machine are only background
windows, e.g.: screenlets showing CPU usage.
In current wmaker version all these applications pollute my
switchpanel, so I wrote this patch (thanks Carlos for the helping me).
It includes an additional advanced option for windows "Do not
include in switchpanel" which, if set, allows applications to
not appear in the switchpanel.
Signed-off-by: Haroldo Santos <haroldo.santos@gmail.com>
The window menu displays the shortcut key for operations with a
shortcut, but does not display the modifiers. This reduces the utility
of the display as it's hard to know whether "h" means "Mod4+h" or
"Ctrl+Mod4+h" or something else.
This patch prefixes those shortcut displays with the modifier names,
e.g. Ctrl+Mod4+h, and adds a preference ModifierKeyLabels to allow
overriding this, e.g. to "⌃◇h". It doesn't add this preference to
WPrefs, if someone else wants to do that feel free.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
The previous patch "Fix path substitutions" accidentally caused autoconf
to overwrite automake's pkgdatadir. Since configure.ac does absolutely
nothing interesting with pkgdatadir anyway, just remove it completely.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Patch "Fix path substitutions" moved generation of pkgconfig files from
./configure to Makefiles. However the generation is not triggered since
the pkgconfig files are not listed as dependency. Fix by conversion to a
straightforward automake rule.
Acked-by: Brad Jorsch <anomie@users.sourceforge.net>
Signed-off-by: Andreas Metzler <ametzler@downhill.at.eu.org>
If the app is destroyed before the bounce finishes, the icon may be left
out of position. Fix that by explicitly resetting the position when the
bounce is complete for whatever reason, not just when the bounce ends
normally.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Recognize the freedesktop.org-mandated Main Categories in .desktop files
and put menu entries in the appropriate categories.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Add optional interface parsers may choose to implement that, when
a directory is being scanned for files to parse, may, based on
the file name, decide whether or not said file should be parsed.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Add fileInPath, which determines whether or not a given file exists
in $PATH (some heuristics apply).
Make the Wconfig parser use it.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
As far as i can tell this finishes the Wmconfig parser (all bits we are
interested in/can use are parsed and converted).
The PropList writer gained ability to properly react to and handle various
flags passed by parsers.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
There is little point caching a pixmap for an app that isn't in the
dock. This patch creates a function wAppIconSave that saves only if the
app icon is docked, and adds calls to that function in all the places
where an appicon can transition from undocked to docked.
It also "adds" a function wApplicationSaveIconPathFor that saves an icon
path to the configuration plist; the function already existed, it was
just static before.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
This patch changes the logic for choosing the icon, and gets rid of
wIconChangeImage in favor of wIconChangeImageFile.
Old logic:
* On create, load either NET_WM_ICON or the file from disk.
* On update, choose (in order):
1. WM_HINTS icon_window
2. WM_HINTS icon_pixmap
3. Whatever was loaded on creation, unless wIconChangeImage or
wIconChangeImageFile was called.
4. Default icon.
New logic:
* On update, choose (in order):
1. WM_HINTS icon_window
2. NET_WM_ICON
3. WM_HINTS icon_pixmap
4. Icon file from disk.
5. Default icon.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
NET_WM_ICON is a property change, not a message.
updateIconImage doesn't use scr, remove it.
updateIconImage should clear out the net_icon_image if there isn't one
anymore (i.e. XGetWindowProperty fails or the icon cannot be extracted).
When the icon is changed, we need to call wIconUpdate.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Non-obvious fixes:
WINGs/wfilepanel.c: Cast to void to avoid an unused calculated value
warning.
WINGs/wtabview.c: Test tab<0 to avoid a warning from the next condition
about signed overflow in an inlined invocation of the function.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
- add WMFindInTreeWithDepthLimit, which is like WMFindInTree, but
does not descend down more than a set limit.
- add WMTreeWalk, which will walk a WMTreeNode, running a callback
function on each node.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
When applications are embedded in firefox via mozplugger, e.g. ooffice
for viewing .doc files or xpdf for viewing pdfs, and the embedded
application opens a popup window, we end up with a shared appicon for
the embedded app that remains until Window Maker is restarted.
The underlying cause is that mozplugger winds up reparenting the
embedded app's original leader window to a subwindow of itself, so
wmaker's SubstructureNotifyMask on the root window no longer picks up
the DestroyNotify when that original leader is destroyed.
It seems easy to fix once you track down the problem: when fixing
several other properties in fixLeaderProperties(), add
StructureNotifyMask so we get DestroyNotify no matter where the leader
gets reparented to. A more targeted fix would only add
StructureNotifyMask when the fake group's origLeader is set, but I don't
think a few extra structure notifications are going to cause us any real
problems.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
When an app hides its window but doesn't exit (e.g. apps that "minimize"
to the system tray), the highlighting was not being removed.
When alt-tabbing, the tabbed-to app's icon was not getting highlighted.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Autoconf uses multiple levels of variables when defining paths. For
example, ${datadir} by default is ${datarootdir}, which by default is
${prefix}/share, which by default is /usr/local. Substituting from
./configure, as is done by AC_DEFINE or AC_DEFINE_UNQUOTED, does not
expand all these variables. This was causing some of our defines to have
garbage like "${prefix}/share/pixmaps" rather than the intended
"/usr/local/share/pixmaps".
The solution is to generate the files needing these paths from the
Makefile rather than from ./configure, because make does fully expand
all those levels.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
How to use:
- Install LCOV (http://ltp.sourceforge.net/coverage/lcov.php)
- AFAICT, compiler must be GCC.
- ./configure --prefix=/some/path --enable-lcov=/var/www
This will configure the sources so that generated coverage information
(HTML pages) will end up in /var/www/coverage-report. Choose a directory
you have write access to.
- make
No installation necessary (although it probably doesn't hurt to have
WM installed in ${prefix} so it can find resources and stuff, but the
build tree must be kept intact anyway).
- Put ${srcroot}/src, ${srcroot}/WPrefs.app and ${srcroot}/util first in
your path.
- Arrange for your .xinitrc or DM to start ${srcroot}/src/wmaker as your
window manager.
- Use it as you normally would, possibly for an extended period of time.
- make coverage.
This will create a tree of web pages under /var/www/coverage-report
- Further `make coverage's (after some more use) will re-generate the
report (deleting the directory first).
- `make coverage-reset' will reset execution counts to zero (see lcov(1)).
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
- Change the wusleep abomination to be a simple wrapper around
nanosleep (man says it's been POSIX for almost a decade)
- Remove autoconf tests that became unnecessary along the way
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
- on_exit is nowhere to be found in recent written history
- so there haven't for long been a system with !HAVE_ATEXIT
- so there is no need for either #ifdef HAVE_ATEXITs or configure
checks for atexit
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
An appicon is created for each group of windows, as specified by the
WM_HINTS window_group or the WM_CLIENT_LEADER. The "shared_appicon"
feature ignores the group leader specified by the application, replacing
it with a dummy leader matching the window's WM_CLASS. This causes
issues for dockapps, since each instance of a dockapp needs its own
appicon to display the different icon_windows, so shared_appicon is
automatically disabled for dockapp windows.
If the application creates some dockapp windows (no shared_appicon) and
some regular windows (with shared_appicon) with the same leader, it can
unexpectedly end up with two different appicons: one for the real group
leader it set and one for the fake leader created for shared_appicons.
Both of these appicons will try to use the same icon_window, which may
cause the dockapp window to suddenly lose its contents as they are moved
to the fake leader's appicon.
There is a simple fix: if a WApplication already exists (and has an
appicon) for the app-specified group leader window, disable
shared_appicon.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
wsyserrorwithcode - Not used, no point either.
wsyserror->werror - qualifying "error" with a "type" hardly makes
sense if there are not at least two "type"s. There are not. Safe trip.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
s/enviroment/environment/ both in the .c file and in all .po files that
already contain a translation for the string. The latter prevents
wrongly marking the translation as fuzzy.
The expansion of the macro wApplication{Dea,A}ctivate triggers an
implicit declaration warning for wIconSetHighlited. Add function
declararion to fix this.
On the one hand, libWINGs wasn't linking against -lX11 when it should
have been. And on the other, only libWINGs needs Xft, only wmaker needs
Xrandr, only wmaker and wmsetbg need Xinerama, only libwraster needs
Xmu, and -lpng may not need -lz.
Cleaning this up can help distributions get their dependencies correct,
and might even avoid loading the unused libraries at runtime, so we may
as well do it.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
On BSD systems, strlcat and strlcpy are included in the C library and
nothing special is needed. On Linux systems they are not, but libbsd may
be available to provide them. Use it if so.
This also adds wstrlcat and wstrlcpy instead of trying to maybe-provide
strlcat and strlcpy themselves, as that way there is no risk of symbol
conflicts. Not bumping the library version at this time, that should be
done (if necessary) before release.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Nice idea, x86 assembly is so 1980s.
tests/testgrad.c could probably have been saved, but seriously, who's
going to miss it?
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
The comment in WINGs/memory.c:wfree() pretty much explains the current
situation. There's an incredible amount of mixing the wmalloc/wfree
wrappers with native mallocs/frees on the other side, and a good several
cases of misusing external libraries' APIs. Until this is thoroughly
cleaned, WM with --enable-boehm-gc will hardly even start.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Turns unhelpful
wmaker: warning: wrong option value for key "NewStyle". Should be one of new, old, next
into helpful
wmaker: warning: wrong option value for key "NewStyle". Got "YES", should be one of "new", "old", "next"
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Commit acc868d91d tried to change
--with-appspath to --with-gnustepdir, but either it wasn't done
correctly (and no one noticed) or old versions of AC_ARG_WITH worked
completely differently (and in an extremely odd way) because the option
remained --with-appspath and just the help text was changed. Then commit
bde572f673, which claims to have been
fixing some sort of issue with path assignment, broke it completely (i
guess) because the author there trusted the help text.
This completes the change from commit
acc868d91d that makes the option really be
--with-gnustepdir.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
If we Restart() directly, the windows with titlebars come back N pixels
below their original positions before the xrandr-induced restart, where
N is the titlebar height.
To avoid this issue, let's do the proper restart preparation before
actually calling Restart().
Let's also grab ConfigureNotify in the event loop as in the patch
here: http://lists.kde.org/?l=kwin&m=116429907520188&w=2
(thanks to Tamas for pointing it out).
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
If you link against a library A that itself links against a library B,
it may or may not work to use symbols from library B in your executable;
this is known as "indirect linking". GNU ld does indirect linking by
default (but it can be disabled using --no-add-needed), but the new
experimental GNU gold linker does not do this. It's an easy fix for us,
as the tests are all already done in ./configure, we just need to tell
the Makefile.ams to use the results.
This should fix Debian bug #556677, if they ever start using this
branch.
[crmafra: Folded Andreas Metzler patch to update debian/changelog]
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Remove local strcasecmp implementation; whatever system doesn't have it
is off the table for now.
memcmp is used by WINGs; add autoconf check
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Prevent WPrefs (menu panel) segfaulting upon coming by an invalid command in
the root menu.
Reported and first patch version by Bento Loewenstein.
Signed-off-by: Tamas TEVESZ <ice@extreme.hu>
Some people using nvidia and/or playing some games (e.g. Einstein)
are reporting some issues with the restart induced by the
resolution changing. Let's disable it by default.
People without nvidia cards etc can still enable it with
--enable-xrandr.
Signed-off-by: Carlos R. Mafra <crmafra@gmail.com>
If the appicon is not in the dock/clip, or the dock/clip is not set to
"Keep on Top", there's a good chance you won't actually see the bouncing
because some other window is covering the appicon.
Besides adding the option to raise bouncing windows, this patch adds a
utility method to move a window back into its correct stacking position
after it has been messed with using XRaiseWindow.
Signed-off-by: Brad Jorsch <anomie@users.sourceforge.net>
Since bouncing respects DisableAnimations option, there's no need
in #ifdef's.
Acked-by: Carlos R. Mafra <crmafra@gmail.com>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
I have my dock on the right edge of the screen, and my clip right next
to it. So the first icon underneath the clip is 64 pixels from the top
and 64 from the right. It seems silly for it to bounce downwards instead
of sideways like every other icon under the clip.
So make the choice of vertical or horizontal for docked icons be
relative to the dock or clip rather than absolute screen position.
Left/right or up/down is still chosen by screen position, as are
situations where the icon is a chess bishop's move from the clip.
I just spent an hour trying to find out why the latest wmaker-crm was
ignoring a number of the preferences in my configuration file. It turns
out the problem was that commit f41db5 added a value into the middle of
the enum in src/keybind.h, and the setting of
AUTOMAKE_OPTIONS = no-dependencies in all the Makefile.ams meant that
make wasn't smart enough to rebuild the files depending on it.
Would anyone mind if we just remove no-dependencies from
AUTOMAKE_OPTIONS? People can still run ./configure with
--disable-dependency-tracking to disable it at configure time.
On Wed, Sep 15, 2010 at 04:40:52PM +0200, Gilbert wrote:
>
> "." = {NoAppIcon = Yes;};
> "*" = {NoAppIcon = Yes;};
Oh, I see. It lets you override NoAppIcon generically instead of for
each app.
> Exactly, having 'DockApp' as the class name makes them behave more
> consistently internally. This might be fixed in some other way
> internally, but I couldn't figure it out...
Except it doesn't, the only thing it changes inside wmaker is that it
avoids displaying the name as "DockApp" on the settings window.
A "dockapp" is just an application that sets icon_window in WM_HINTS
(which causes a window to be displayed as the app icon) and
initial_state to WithdrawnState (which causes the main window to not
actually appear when the window is mapped). The class has nothing to do
with it.
The problems I mentioned stem from gtk+ not allowing you to set
initial_state to WithdrawnState. Openbox and fluxbox (and maybe others)
don't take the icon_window into their "slit" unless initial_state is
WithdrawnState, so the whole thing completely fails. Wmaker always uses
the icon_window for the app icon, so that part works fine, but the
"main" app window still flashes onscreen for a split second before the
app can call gdk_window_withdraw(), and wmaker applies SharedAppIcon to
it which winds up screwing up the "Kill" menu option.
That said, I've attached a patch to have wmaker treat any window with
class DockApp as if it had initial_state = WithdrawnState, and hopefully
openbox, fluxbox, and the like will pick up the idea too so gtk+
hacks[1] will no longer be needed. If this patch is accpeted, I'll poke
fluxbox and openbox to suggest the idea to them.
[1] E.g. http://wmudmount.svn.sourceforge.net/viewvc/wmudmount/dock.c?r1=8&r2=7
Subject: [PATCH] Improve dockapp recognition
Dockapps are traditionally recognized by having initial_state =
WithdrawnState in WM_HINTS. But some toolkits (e.g. gtk+) will not allow
setting initial_state in this way. So we offer an alternative: any
window with the res_class portion of WM_CLASS set as "DockApp" will be
treated as if it had initial_state = WithdrawnState.
This patch highlites appicon of a currently focused window.
Original-patch-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
When starting application (or opening transient dialog) bounce
its appicon.
Original-patch-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
When doing right-click in a window list menu open window menu.
Can be useful for sending here window from another workspace.
Original-patch-by: Pavel S. Khmelinsky <hmepas@yauza.ru>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
MinimizeAllKey - minimize all windows on current workspace.
Original-patch-by: Pavel S. Khmelinsky <hmepas@yauza.ru>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
[crmafra: AllMinimizeKey --> MinimizeAllKey]
When generating menus, lookup translations for menu entries in
specified textdomain. Often used by linux distributions.
Original-patch-by: Alexey Voinov <voins@altlinux.ru>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
When unhiding window by clicking its appicon, unconditionaly
raise its frame.
Original-patch-by: Artem Delendik <u2u@nm.ru>
Signed-off-by: Alexey I. Froloff <raorn@altlinux.org>
Keep libtool's -version-info arguments for the three libraries
(libwraster, libWINGs and libWUtil) centralized in configure.ac
instead of in the specific Makefile.am.
If AC_INIT is invoked without tarname argument PACKAGE_TARNAME and
PACKAGE will be sanitized (lowercased, and all characters other than
alphanumerics and underscores are changed to ‘-’.)
For windowmaker this has the sideeffect of changing install directories,
util/wmaker.inst installs stuff to #pkgdatadir# which suddenly is
/usr/share/windowmaker instead of /usr/share/WindowMaker.
Supplement AC_INIT invokation with tarname (and while we are at it url)
argument to fix this.
The code was added in 2001 in commit acdc0e3d6b for
"backwards compat". This motivation no longer exists in 2010,
so there is no reason to keep a private implementation of
a standard function from the C library.
Instead of using the --disable-verbose-compile hack, let's
use the standard option --enable-silent-rules (or 'make V=0' directly)
available with autoconf 1.11 and higher.
This is enabled in autoconf by using 'silent-rules' in AM_INIT_AUTOMAKE.
The verbosity of compilation is reduced in a similar manner as using
the --disable-verbose-compile option, so we can remove all that ugly
hackery.
According to the autoconf manual, using AS_HELP_STRING is recommended
to avoid having to manually adjust the alignment of help text
in ./configure --help
WUtil.h:212: warning: function declaration isn’t a prototype
WUtil.h:259: warning: function declaration isn’t a prototype
WUtil.h:300: warning: function declaration isn’t a prototype
WINGsP.h:593: warning: function declaration isn’t a prototype
WINGsP.h:599: warning: function declaration isn’t a prototype
WINGs.h:616: warning: function declaration isn’t a prototype
raster.c:295: warning: no previous prototype for ‘calculateCombineArea’
This patch makes wmaker automatically Restart() to take into account
resolution changes done by xrandr.
The motivation to write this patch is that when I start my laptop connected
to an external monitor (VGA1), the resolution in the monitor is initially
wrong (perhaps this is a bug in XOrg or OpenSuse, but anyway), so I always
do this after starting wmaker:
xrandr --output VGA1 --preferred
The resolution in the external monitor gets fixed, but wmaker does not
recognize the change automatically and I have to manually restart it.
So this patch avoids the last step by making wmaker be XRandR-aware.
Wmaker now restarts automatically when it receives a RRChangeScreenNotify
event and I don't have to do the last part of the silly procedure described
above.
I am not sure if restarting wmaker is the most elegant solution, but
it at least solves part of my issue with wmaker + xrandr.
Implement the switch from SelectWindowsMouseButton et al. to
MouseLeftButtonAction = SelectWindows from WindowMaker 0.65 in Debian's
customized defaults. Closes: #116963
lintian says
W: wmaker source: debian-rules-calls-debhelper-in-odd-order
dh_makeshlibs
and indeed invoking dh_installdeb (which install the maintainerscripts)
before dh_makeshlibs can add the ldconfig call to them causes these
errors:
E: libwutil1: postinst-must-call-ldconfig usr/lib/libWUtil.so.1.0.2
E: libwings2: postinst-must-call-ldconfig usr/lib/libWINGs.so.2.0.1
Shared libraries need to be shipped in separate packages with shlibs
information. Add libwutil1 and libwings2 packages and let libwings-dev
depend on them.
Instead of calling dh_installman multiple times just call it once. List
the manpages that should be installed in package specific
debian/package.manpages.
Starting with debhelper v3 mode every file in etc is flagged as a
conffile by dh_installdeb. - Drop (commented) code for doing this
manually from debian/rules.
-isp is default behaviour of dpkg-gencontrol since 1.13.16 (18 Feb
2006)
The code for generating debian/shlibs.local is not used, remove it.
dpkg-shlibdeps already searches in debian/libwraster3/usr/lib/ by
default.
Use dpkg-buildpackage -rfakeroot option to only run the parts of the
package generation process which require (fake)root privileges under
fakeroot. (e.g. debian/rules build should not be run as root.)
Simplify debian/rules by using dpkg-source v3 (quilt) instead of the
home-grown patch system in debian/rules.
Change debian/nightly_build.sh to build a non-native package.
Use dh_prep instead of "dh_clean -k" (dh-clean-k-is-deprecated). However
do not replace instances of dh_clean without -k option. (Debian
bug590244). Build-Depend on debhelper 7.
Correct floppy mountpoint location to match FHS.
Quoting FHS <http://www.pathname.com/fhs/pub/fhs-2.3.html>:
/media : Mount point for removeable media
Purpose
This directory contains subdirectories which are used as mount points for removeable media such as floppy disks, cdroms and zip disks.
[...]
The following directories, or symbolic links to directories, must be in /media, if the corresponding subsystem is installed:
Directory Description
floppy Floppy drive (optional)
[...]
libWINGs is using mathematical functions (atan, sqrt, etc.). Check
whether we need to lik against libm to use these and link libWINGs
against it if necessary.
libWINGs is using libwraster functions (e.g. RScaleImage) and therefore
needs to be linked against it.
ametzler@argenau:/tmp/WMAKER/wmaker-crm$ env LD_LIBRARY_PATH=debian/wmaker/usr/lib ldd -r debian/wmaker/usr/lib/libWINGs.so.2.0.1 2>&1 | grep RScaleIma
undefined symbol: RScaleImage (debian/wmaker/usr/lib/libWINGs.so.2.0.1)
libWINGs is using libWUtil functions (e.g. WMAddIdleHandler). Reorder
lib_LTLIBRARIES to build libWUtil before libWINGs. Link libWINGs against
libWUtil.
In C, dividing two integers automatically rounds towards zero, so ceil(a
/ b) is useless as the result is truncated before ceil ever sees it. The
correct result for positive integers is obtained by (a + b - 1) / b.
I personally cannot stand stealing Ctrl+Wheel for resizing windows, so I
turned it off. But wmaker is still preventing Ctrl+Wheel from reaching
apps, which is certainly not the behavior I desire.
We grab Mod+Button events for our own purposes, and swallow them using
XAllowEvents(AsyncPointer) to prevent the client window from seeing
them. But if events are coming in fast enough (e.g. via fast wheel
scrolling) so that multiple grabbed events are in the queue, the second
queued event would be allowed to leak through to the client by the
unconditional XAllowEvents(ReplayPointer) a few lines later.
One misuse of CARD32 was missed back when someone fixed the 64-bit
issues (background: X has 8-bit, 16-bit, and "32-bit stored in a long"
data formats; on a 64-bit system, long is 64 bits). This was causing
minimized windows to be restored as shaded, and possibly occasional
crashes, when wmaker was restarted.
Also, throw in a safety check that 10 items really were returned before
trying to access them.
The "make dist" command can be used to generate the release tarball, but
only if the Makefile.am correctly specifies EXTRA_DIST, specifies
headers in foo_SOURCES, and so on.
Initialized variables that are conditionally set. In particular, this
construct is dangerous:
void *p;
if (something) p = couldReturnNull();
if (!p) p = fallbackFunction();
C99 defines new strict aliasing rules to allow compilers to make certain
optimizations. These rules prohibit converting an XEvent to an event
struct (e.g. XShapeEvent) that is not already in the XEvent union using
pointer type punning (e.g. "(XShapeEvent *)&ev"), and vice versa. The
canonical fix seems to be to create a union between XEvent and the
extension event struct to make the aliasing explicit, so do that.
A few internationalized format strings have only one %s in the default
while two strings are being passed. Change those defaults to have two
%s's, which means we also have to update the .po files to match. May as
well throw in the extra %s in the translated version while we're at it.
If a window has the skip_window_list flag set wmaker currently not only skips
this window from the window list (F11) but also from the switchpanel (TAB).
This patch makes skip_window_list-windows appear in the switchpanel and be
skipped only in the window list, strictly honoring what the name of the flag
is supposed to mean.
The motivation for this is that I have lots of workspaces and almost all of
them has a xterm, so that I set the skip_window_list for the xterms not to
overcrowd the window list. But I still want to alt+TAB to xterms on the current
workspace.
In addition to that, now the "internal window" flag is not checked before
appearing in the switchpanel.
gcc-4.4.1 was complaining:
fontconv.c: In function ‘mapWeightToName’:
fontconv.c:114: warning: field precision should have type ‘int’, but argument 4 has type ‘size_t’
fontconv.c:114: warning: field precision should have type ‘int’, but argument 4 has type ‘size_t’
fontconv.c: In function ‘xlfdToFc’:
fontconv.c:157: warning: field precision should have type ‘int’, but argument 4 has type ‘size_t’
fontconv.c:157: warning: field precision should have type ‘int’, but argument 4 has type ‘size_t’
- Kill unused vars in wrlib
- Add missing initializers to defaults.c:staticOptionList
- Re-format it slightly
- Kill some dead code
- Kill stupid "if (const op var)"
I noticed a bug today in menu workspace positioning (in the next tree):
if I right-click and hold the button down at the left edge of the
screen, the menu appears (correctly) half off the edge but will *not*
slide back onto the screen. Compare the behavior to right-clicking at
the right edge of the screen: the menu appears half off the edge, but
sliding the mouse to the edge causes the menu to slide until it is fully
visible.
Also, opening a submenu in this state positions the submenu as if the
menu were fully on the screen, leaving a gap between the menu and the
submenu.
If the menu happens to also go off the bottom of the screen, moving the
mouse to the bottom edge causes the issue to be magically fixed as soon
as the sliding upwards begins.
I also note inconsistent behavior when simply right-clicking (without
holding) to bring up the menu: at the right or bottom edge the menu
appears in the correct partially off the edge position, but at the left
edge it immediately jumps to be fully on-screen, almost as if
WrapMenus = YES were set for the left edge only.
I bisected it to d316260395. As far as I
can tell the "fix" there was to position the menu at the correct
negative X position but then lie to wmaker so it thought the menu was at
X=0. Presumably the "WrapMenus" behavior was the intended result.
Since (AFAIK) the window menu is the only one with this problem, why
don't we just check if x is too negative in OpenWindowMenu()?
When using Xinerama, make it possible only to cycle between those
windows that are on the currently active head - AFAICT this means
windows that have the majority of their area on the same head as
the pointer.
Patch complete with WPrefs integration, alas this being quite
unusable - for a lack of a better idea, I tucked the option on the
expert panel, making the item list too tall - making it scrollable
would be a nice addition.
The default for the new option (CycleActiveHeadOnly) should be NO,
that is, no change in default behaviour.
- change some variable names to better reflect their purpose, do a little
write-up how a menu is built
- auto-detect what terminal to use for apps in need of a terminal
- fix the chunk that adds the terminal-based apps
- convert to getopt
- tidy xsendevent
- modify hackPathInTexture() so that:
- it properly figures where the style file is
- if it's under the user's home directory, that part of the path
is changed to `~' - this makes your wm configuration usable on
systems with a network home, used on systems where home is /home,
or /u, or /export/home, or whatever.
XXX: there's a note WINGs/findfile.c:wexpandpath() stating that
`too many functions handle a return value of NULL incorrectly'. this is
so true. so if anyone is looking for a small project to sink their
teeth in...
- according to the automake manual, `acinclude.m4' is the old style
of doing stuff, putting local macros in their own directory is the
way to go, so move acincluce.m4 to m4/windowmaker.m4
- reflect this in autogen.sh and Makefile.am
- while there, add a `conditionally set cflags' macro from the
autoconf macro archive
- use this to slightly pump warning levels up if we are on gcc
- convert hand-rolled arg parsers to getopt_long
- sort headers, get rid of some duplicates in the process
- in fontconv, replace magic numbers with constants (there's
little practical value to this here, but it's much nicer)
- slightly redo fontconv.c:strToInt()
The rationale is the same as in commit c7f2a189c4
("Fix the call to XChangeProperty() in 64-bit mode"), because we are
calling XChangeProperty() with format 32, which requires the type of
'data' to be 'long'.
- Does away with the O_BINARY abomination
- as a byproduct, plugs an fd leak in wcolorpanel.c:fetchFile()
- sprinkle some fsync()s to files that have been written to (this
needs to be done everywhere)
+ fix brown paper bag thinko in configure.ac
* Remove directly include of <libintl.h> and include src/wconfig.h, that
already include libintl.h and declare _() function
* Replace some gettext() calls to _()
These changes fixes build with LINGUAS is unset, and removes I18N
support in this case.
was this ever used for anything? i checked a couple hundred dockapps,
no joy; there's nothing in debian (afaict) that build-depends on it,
so let's try doing away with it.
- tested on Net, Free, Open and DragonFly
- fix up the linux version (terminate argv with a null ptr)
- add error handling to file ops in the linux version
- add a stub version for platforms not supported
As the README used to say, it was a unfinished work. If someone
wants to finish it after all these years it can be recovered
from the logs. But meanwhile let's not carry dead code around.
- add new wglobaldefaultspathfordomain() to wings (replaces several
hand-rolled individual implementations in utils/)
- make all of 'em handle -h|--help, -v|--version
- try making them not to nothing silently
- change various ways of knowing thyselves to using __progname
- generally try to make them feel similar (NOT right, similar --
right is a completely different matter)
In the wmaker-dev list, Iain noted that:
"I can see the benefit of this. I know of at least two
applications[1] which don't bother to set a class and instance. Early
versions of my OS X cycling patch for Window Maker caused a crash when
faced with such windows because I forgot to check that these
properties existed. Nowadays it just refuses to do the cycling.
Also Window Maker won't let you save attributes on such windows from
the Inspector.
[1] Interestingly they are competiting commercial applications: No
Machine and Exceed onDemand."
[crmafra: edit commit log]
Since getstyle now saves the theme pack in $GUR instead of
the cwd, there's a slight possibility you can specify a theme
pack name that will do Bad Things to your existing stuff.
To that end, specified theme pack name is "sanitised" a bit, as in
forward slashes in it are replaced by an underscore.
Oh yes, i forgot to mention. getstyle now saves the theme pack
in $GNUSTEP_USER_ROOT, not in cwd.
- make it use wings functions, remove duplicated code from getstyle
- de-static necessary functions in wings
- add new wrmdirhier to wings
- rename WMMkDirHier to wmkdirhier (fits better)
- remove calling shell from getstyle (what were they thinking?)
i couldn't quite test getstyle (no idea about themes), but it still
basically works.
do back your ~/G dir up... wrmdirhier might eat it!
definitely needs testing, especially by people who have any idea
how themes work.
Some more getstyle
- missed a shell invocation
- maybe copyFile should be in wutils too...?
[crmafra: Folded second patch into the first]
By defining a struct containing the coordinates of the
windows (top, bottom, left, right, width, height) and
using a helper function to set up these coordinates, the
whole function becomes much cleaner and easier to read.
make wtrimspace() use wings' own function for a task
semantics change: it used to segfault given null, now it returns null.
this doesn't affect any current use (there's exactly one..), and i see
no harm in this behaviour, and perceive this to be more natural.
The result is not much more convoluted than the original was, but much shorter.
Several vararg macros were added -- no idea what !gcc compilers make of this.
The messages sent to these functions are inconsistent across the source tree.
I have now decided that the logging function will add the final newline -
messages will need to be modified accordingly.
I have no idea why the original implementation was as complicated and ugly as
it was. My guess is that it was anticipated that these are be called from
sighandlers, but why no snprintf when they are all stuffed up with vsnprintf...
Still, the result is not worse in this regard either.
It addresses this warning
dialog.c: In function ‘LoadHistory’:
dialog.c:209: warning: passing argument 2 of ‘WMFindInArray’ from incompatible pointer type
../WINGs/WINGs/WUtil.h:455: note: expected ‘int (*)(void *, void *)’ but argument is of type ‘int (*)(const void *, const void *)’
but induces others in other places. One of them was this one
window.c: In function ‘wManageWindow’:
window.c:782: warning: passing argument 2 of ‘WMFindInArray’ from incompatible pointer type
../WINGs/WINGs/WUtil.h:455: note: expected ‘int (*)(const void *, const void *)’ but argument is of type ‘int (*)(void *, void *)’
which is fixed by constifying the arguments of matchIdentifier(). The other warnings are
fixed similarly.
WindowMaker.h already #includes wconfig.h, which checks exactly
the same thing. So remove it from WindowMaker.h.
However, it would be better if WindowMaker.h didn't include
wconfig.h at all. The only reason I see for the inclusion is
the check for KEEP_XKB_LOCK_STATUS, defined in wconfig.h.
But that can be probably cleaned up too (I guess modern X's
enable XKB stuff that by default, by I haven't checked)
But anyway.
remove the choice of atomic/non-atomic writes. firstly, the only users
of non-atomic writes were getstyle and geticonset; secondly, who in their
right minds would ever want non-atomic writes; thirdly, the file system
will screw you anyway *G*.
...which creates directory hierarchies under $GNUSTEP_USER_ROOT
- make WMWritePropListToFile() use this
readily solves the "run dialog history vs nonexistent state dir" case.
if other consumers are found for this, just de-static and move the
prototype to WUtil.h or some other appropriate place.
Taking the border width into account was making it hard to follow the
algorithm, and complicating things usually lead to mistakes.
In fact, in trying to account for this +/- 1 pixel issue the maximus
geometry for the requested window would sometimes overlap another window
by one or two pixels (depending if it was on the right or left). Therefore,
applying the maximus geometry to this other window in sequence would
maximize it completely instead of finding the correct tiled geometry.
So that always bugged me. Now I got tired of it and decided to kill the
whole consideration of the border width, and simply put some "+ 1"
inconditionally to be on the safe side.
Now it can happen that the windows will stay away one or two pixels,
but trying to be 'good enough' is better than trying to be 'perfect'
and getting it wrong for corner cases.
And the matter is worse because the size of xterms are 'quantized', so
changing the position of the other windows by a few pixels result sometimes
in a overlap of +1 pixel, or a distance of 1 or 2 pixels. Furthermore,
the bottom of xcalc behaves differently from the bottom of a xterm or
xjed (for xcalc the ad hoc correction for bottom_0 must be -8 in total).
But now my tests show that things are OK, and I can use maximus
for all the windows in the screen and no overlap will occur (except
if one of the windows is a xcalc, I don't know why).
It wasn't defined anywhere and when I defined it, the compilation
did not finish:
appicon.c: In function drawCorner:
appicon.c:206: error: WScreen has no member named focused_texture
appicon.c:208: error: WScreen has no member named unfocused_texture
appicon.c: At top level:
appicon.c:214: error: conflicting types for drawCorner
appicon.c:193: note: previous definition of drawCorner was here
make[2]: *** [appicon.o] Fehler 1
make[1]: *** [all] Fehler 2
make: *** [all-recursive] Fehler 1
So let's simply remove those #ifdefs.
This patch only calls XGrabButton for CTRL+Button4 and CTRL+Button5.
This leaves CTRL+Button1-3 to the application.
This then removes the functionality of moving a window between
workspaces with CTRL+Button1 and CTRL+Button3
This patch constrains MOD+Wheel to vertical resize, and adds
CTRL+Wheel horizontal resize. Two resize in both directions, you
have to use CTRL+MOD+Wheel.
To enable this functionality I have to grab all CTRL+Mousebutton
events in wmaker, which stops them from reaching the application.
This definitely hurts application functionality in some apps, for
example the "VT Fonts" (CTRL+Button3) menu in xterm is no longer
accessible. To stop this from happening use the "Do not bind mouse
clicks" window attribute for the apps in which you want to disable
this.
Because wmaker now controls all CTRL+Mousebutton events, I also
added CTRL+Button1 and CTRL+Button3 shortcuts that will move a
window back and forth through your workspaces without changing its
position or size.
The non-gpl warnings in WPrefs.app/tiff/README
and WPrefs.app/README were misleading, because
they were refering to Marco van Hylckama Vlieg's
work which was licensed with OpenContent License.
But as Alfredo Kojima wrote in the Changelog in
commit f65c549814 from 30.03.2010,
Marco's icons were licensed with the GPL.
Furthermore, in an email to wmaker-dev on 10.03.2010,
Marco confirmed:
"Anyway, I have no problem with the artwork being
licensed under GPL or whatsoever."
The autocomplete feature in the "Run" dialog now adds a trailing "/"
when autocompleting directory names. This does three things:
1) shows you that you are completing a directory and not a filename,
2) saves one character of typing, and
3) makes it behave more consistent with shell (bash) autocompletion
I've found a buffer overflow problem in RSmoothScaleImage. There are
some scaling calculations involving floats which are finally converted
to integers. Since such conversion does not round the number, just
truncates the decimal part, sometimes the number is smaller than it
should be. As a result, smaller buffer is allocated for picture
scaling and thus buffer overflow occurs.
Strange thing is that this bug has not appeared earlier so it probably
has something to do with newer gcc or glibc (I switch from
"prehistoric" Fedora Core 5 to Fedora 12).
<What about the symptoms?>
There were several ones, probably depending on application version and
compilation flags. First, it just stopped responding. Looking at the
process with strace I saw it locked in some FUTEX wait (unfortunately
I don't have the logs). Second, it just crashed. And last I got
complaint from glibc about double free or corrupted heap before
malloc. I've found the bug through wmweather+ dockapp, versions 2.9
and 2.11 (http://sourceforge.net/projects/wmweatherplus/), I've never
encountered it in WindowMaker itself.
The inspiration comes from Geir T. Kristiansen's 'genmenu' shell script
(http://gtk.no/genmenu) which I have been hapilly using for
the last 9 years or so.
That script generates the Window Maker menu by asking a few questions
and checks whether the applications in a predefined list exist in
the user's $PATH, so that the menu contains only the applications
which are guaranteed to exist and which the user cares about.
However I always thought it was a bit slow to finish, even
when having a file containing the answers to the questions to
be piped in.
And as Kristiansen states in his webpage, his script does not
support internationalization:
"If you want internationalization I really suggest you rewrite
genmenu from scratch anyway in a more sensible language and remove
other limitations while you are at it."
so I decided to rewrite in C and make it support internationalization
(only English and German so far). While I am not sure yet if I removed
"other limitations while you are at it" I definitely made it finish faster.
This C program does not make you questions though.
But in the same spirit as the author of 'genmenu', who says:
"Genmenu is a hack, deal with it, it just happens to serve my needs."
I present you 'wmgenmenu', which you can use like this:
wmgenmenu > /home/mafra/GNUstep/Defaults/WMRootMenu
and the menu is automatically generated from the list of
predefined apps which are contained in the C sources (wmgenmenu.h).
And when you generate the WMRootMenu as above, wmaker will automatically
load the new menu for you (due to the 'inotify' mechanism in wmaker-crm).
As for 'wmgenmenu' being faster than 'genmenu':
[mafra@Pilar:~]$ time ./genmenu.sh < answers.txt &> /dev/null
real 0m3.626s
user 0m0.968s
sys 0m1.830s
[mafra@Pilar:~]$ time wmgenmenu &> /dev/null
real 0m0.020s
user 0m0.006s
sys 0m0.013s
Once upon a time (< 2005) the CachedPixmaps directory was located
at ~/GNUstep/.AppInfo/ and that was later moved to
~/GNUstep/Library/WindowMaker. So Dan Pascu introduced this function
in 24519b6292 to make the convertion
automatically to users back then.
As it is highly unlikely that there is an old-timer wmaker user still
running a pre-2005 wmaker which suddenly decides to switch to wmaker-crm
and runs into trouble with his CachedPixmaps folder, let's simply remove it
to make defaults.o 1.8% smaller (596 bytes).
No new translations were added. The .po file was recreated with
make WPrefs.pot
msgmerge de.po WPrefs.pot > de.po.new
(remove old strings from it)
mv de.po.new de.po
It was fun trying to translate them, but even after several
minutes of deep thoughts about the different cases etc I managed
to translate them wrongly.
So I thank Martin Dietze for kindly fixing them!
I wanted to learn some German and got curious to read the Window Maker
translations. After noticing that some strings with translations
were already removed from the main source I decided to recreate the
German .po file to get rid of those messages.
Even the option to enable "virtual desktop" in configure.ac was
commented out...and I would never intend to use it anyway.
So let's just remove the ~800 lines of #ifdef'ed code to have a
cleaner code base to read when bored.
The raising and lowering of the clip is already taken care of
by the "ClipRaiseLowerKey" shortcut.
wmaker gets a bit smaller for free:
text data bss dec hex filename
449483 17384 8208 475075 73fc3 src/.libs/wmaker.old
449307 17256 8192 474755 73e83 src/.libs/wmaker.new
This patch adds the DockRaiseLowerKey shortcut, which raises/lowers
the dock depending on whether the dock is lowered/raised.
[crmafra: Reformatted Brad's patch against git repo and removed the
DockRaiseKey and DockLowerKey shortcuts ]
I don't want the (small) overhead of the window birth zoom special effects
because I like apps popping up instantly when I open them (specially now
with the SSD).
Another way to not get the birth animation is to unset "animations" in
WPrefs, but that makes the other animations which I care (ie they don't have
the effect of making my computer seem slower than it really is) go away.
Paul Harris reported that using the mouse wheel over a miniwindow
would deiconify it to a different workspace than the original one
where it was iconified.
This happens because after the window begins to be deiconified the
"residual" mouse wheel scrolling hits the workspace background, and
Window Maker changes workspace with wWorkspaceRelativeChange().
But if it all happens fast enough (so the deiconification animation
did not finish yet) the workspace will have changed before the
window reaches its final deiconified destination, leading to
the situation that Paul described in the link below.
So to avoid this, let's set a 'ignore_wks_change' variable
from wDeiconifyWindow() and make wWorkspaceRelativeChange() respect it.
Original report: http://lists.windowmaker.info/dev/msg00821.html
Since a single default visual ID cannot be used for multiple screens, thus
Window Maker refused to start. There is now a global function for getting the
default visual ID. The command line argument --visual-id can be a comma
separated list of visual IDs for each screen. A default value is only set for
the first screen.
Part of the menu on Debian systems is stored in
/etc/X11/Windowmaker/menu.hook. This patch will add this path to
src/wconfig.h.in when building the Debian wmaker package. This is
what is done in the Debian version of wmaker 0.92.0.
This patch fixes a minor bug in Maximus: the new window size didn't take the
border into account. This bug was particularly visible with the
"do not cover dock" option turned on.
New windows will only get focused if the mouse is on the same screen.
The code doesn't handle multiple heads but since it's just one screen
the function works as usual. Checking the head and assign the focus only
if the window is on the same head might be a good idea. Making the whole
stuff optional might be even better.
This patch fixes a problem with restarting Window Maker in multi-screen
environments. The code for setting the background by calling wmsetbg
changes the environment. In multi-screen setups the DISPLAY variable
becomes :0.0 instead of :0 (for example). The restarted Window Maker
process therefore only manages one screen.
This patch fixes a small problem which only rarely occurs. If a
window is opened only for a very short time and closed right away, in
some circumstances the frame around it stays opened. It looks like
Window Maker doesn't get the destroy message. The XGrabServer call was
commented in the code so I activated it again which seems to prevent this
from happening. I think it actually make sense since the ungrab calls are
used anyway in the following code.
As pointed out by Nicolas Bonifas, wWindowCheckAttributeSanity() is
currently unused. Removing it saves 200 bytes:
text data bss dec hex filename
23476 0 8 23484 5bbc src/window.o.new
23676 0 8 23684 5c84 src/window.o.old
If menu is NULL we would have a null pointer dereference when initializing scr.
> Ie, why not move the assignment of 'src' to after the test?
I thought about this, but I checked the 3 different calls to
updateWorkspaceMenu (all in winmenu.c). They must all call
updateWorkspaceMenu with non-NULL pointers, so if we have a NULL
'menu' pointer then we obviously have a major problem and it is
probably better to crash here than to silently return and fail a
little latter.
[crmafra: edit changelog]
This animation is not as cool as the NORMAL_ICON_KABOON, so I don't
think having a choice here is justified. Let's remove this option
(it was not defined in wconfig.h.in by default) and keep using
the nicer NORMAL_ICON_KABOON.
This code makes wmaker a little bit bigger for no gain at all. I could
undefine SILLYNESS and forget it, but I actually also want a clean
source code to read. And the less code the better.
So now I won't get a different Info dialog on Christmas, but
my wmaker will also not carry around that code in the other 364 days
of the year.
As a result, wmaker gets ~1.7 % smaller
text data bss dec hex filename
448043 17424 8200 473667 73a43 wmaker.new
455340 18360 8328 482028 75aec wmaker.old
Thanks to Iains patch which showed me how to access the window hints, I
now propose the following: The resize increment is set to the closest
multiple of the size hints larger than wPreferences.resize_increment.
This should fix Carlos' complaint about fixed-increment windows resizing
too "slowly", and it also fixes my complaint about the blank space below
the last line of an xterm.
Thanks to:
Iain Patterson <wm@iain.cx>
Use height and width increment when wheel resizing if size hints are
set on a window and meaningful height and width increments are
specified.
Windows which don't care about their resize increments will have height
and width increments set to 1. For these windows - and windows without
resize hints at all - use the setting configured with ResizeIncrement.
Let's remove the configuration options about "sound" from
WPrefs (as it was never fully functionall anyway) and from
configuration files.
Remove also the xpm and tiff icons.
This patch adds the ability to resize windows with the mouse wheel
while holding the Mod key. This currently ignores wWindowConstrainSize
until I can figure out a way to repeatably resize windows with
fixed size increments (like xterm) using this method.
This also adds a slider to WPrefs to choose the increment with which
the wheel will resize a window.
WINDOW_BIRTH_ZOOM was a bit lame and it makes sense to not have
it at all. So remove the trailing "2" of WINDOW_BIRTH_ZOOM2 and
make it the only choice (you can #define it in src/wconfig.h.in).
Note that it uses the same algorithm as the animation to resize,
ie zoom, twist or flip. You can choose them in WPrefs.
Two reasons for removing it:
1) I won't ever want to hear useless sounds
2) The sound support is a bit of a joke. The code is there but you have
to hunt it somewhere else (not in any repository that I know of).
In my 10 years of using wmaker, I never used it for this reason.
Now I consider having no sound in Window Maker a feature, and I like
it that way. So there is no point in carrying useless code around.
PS: There is still the code in WPrefs to be removed.
The problem was the following. While alt-tabbing from one
window to a xterm, wait for the xterm to be raised but
keep the alt key pressed. Now move the mouse cursor
over the xterm and release the alt key.
The result is a xterm in a zoombie focused state; its titlebar
has the focused color but xterm itself is not focused and does
not accept any input.
Fix this by reinstating the check for a NULL pointer from
wSwitchPanelHandleEvent() before changing focus. That function
detects if the mouse cursor moved over to the same window which
is currently being pointed out by the switchpanel, and returns
NULL in this case (the check for panel->current != focus fails).
Thanks to Paul Harris for reporting it!
If we compute the maximus geometry in only one pass through
the window list, the order in which the windows appear in
the list can affect the outcome.
That is because before computing the intersections in the
y-projections we update the y coordinates from whatever window
which happened to change the new_y coordinate during the
previous x-intersection computations, but that may not be _the_
blocking window which decides the final positions in the y axis.
Therefore we may find that this "intermediate window state"
has a non-vanishing y-intersection with another one -- and
be blocked by it -- even though that should not be the case for
the final outcome.
So to avoid that we first scan through all the windows to decide
the final maximumized coordinates in the y axis. Only after that we
compute the x-coordinates.
This patch modifies the linking of the WINGs libraries
to create a shared library. wmaker used to do this, but
it was dropped around wmaker-0.90/0.91.
The shared .so library is needed when compiling and running
the wdm display manager and any other programs which link
to libWINGs.
Submitted by: Gilbert Ashley
Origin: ALT/Sisyphus Linux
When maximumizing a window which has the full_maximize attribute
set we have to consider two possibilities:
1. If the new y coordinate is zero it means that no other
window blocked its maximumization up to the top of the
screen, so we want to put the titlebar outside.
2. If the new y coordinate is not zero it means that another
window in the current workspace blocked its way to y = 0.
In this case we do nothing with the titlebar.
Note that there is another possible fine tunning which is not
addressed in this patch, which is to consider the resize bar.
Let's avoid checking the existence of a border every time
we want to adjust the widths. So we'll always subtract the correction
factor, which is zero in case the window has no border.
It is easier to read the computations if the variables
have shorter names. At the same time it avoids reading
the struct array elements repeatedly, so as a bonus
misc.o gets 64 bytes (0.5%) smaller
[mafra@Pilar:wmaker.git]$ size src/misc.o.*
text data bss dec hex filename
11125 0 4104 15229 3b7d src/misc.o.new
11189 0 4104 15293 3bbd src/misc.o.old
When changing workspaces, after focusing new window, wmaker handles all
pending events. If it receives another change-workspace combination,
you may end up with a workspace without any focused window.
Ctrf-F1 and Ctrl-F2 used to switch to 1st and 2nd workspaces.
Hold Ctrl and press F1 and F2 almost at the same time.
This patch ignores all change-workspace commands while workspace
change is in progress.
See also https://bugzilla.altlinux.org/show_bug.cgi?id=7230
[crmafra: Thanks to Alexey I. Froloff for the above explanation]
The man page of XChangeProperty() says:
"If the specified format is 32, the property data must be a
long array."
And as we call it with format 32, the type of 'data' must
be 'long'. It happens to work nowadays in 32-bit architectures
because sizeof(CARD32) = sizeof(long), but that is no longer
true in 64-bit mode.
This patch was downloaded from
www.openbsd.org/cgi-bin/cvsweb/ports/x11/windowmaker/patches/patch-WINGs_wwindow_c
and I thank Alexey I. Frolov and Vladimir Nadvornik for helping me
to understand it on a wmaker-dev thread.
For those not familiar with the way Macs cycle windows, the Command-Tab
sequence (Alt-Tab elsewhere) switches between DIFFERENT application windows
and Command-Grave (key above tab) switches between windows owned by the
SAME application as is currently focused. So if you had three Safari and
two Finder windows open, and Safari had focus, Command-Tab would switch to
Finder; Command-Tab would switch back to Safari; Command-Grave would switch
to a different Safari window etc.
This patch implements "something like" the above by only populating the
switchpanel with windows matching the currently-focused WWindow's wm_class
when the new cycling mode is activated. In practice this means you can
switch to The Next XTerm or The Next Firefox Window using this method.
The configuration names for these new shortcuts are GroupNext and
GroupPrev. The patch tells WPrefs.app about them. Of course switching to
The Next Window is still possible with the (unchanged) FocusNext and
FocusPrev keys.
Allow wmsetbg to revert to pre-0.90.0 behaviour when choosing a
background image that is large enough to span several heads in a
Xinerama setup. If the -X flag is passed to wmsetbg, the background
image will be stretched to fill the logical screen (as it would be by
default in older versions of wmsetbg) instead of being tiled across
screens.
Shift the workspace name, shown when switching workspaces, by 32
pixels. This is purely for aesthetic reasons. It just looks better
(IMO) than having the name flush against the screen edge.
When switching workspaces, force the name to be shown entirely within
one head of a Xinerama display. Previously the name would span heads if
set to TOP, CENTER or BOTTOM alignment, and was hence hard to read when
the display comprised an even number of heads.
This patch combines two small patches which fix issues with x86_64 and mmx in
the configure script.
Submitted by: Gilbert Ashley
Author: unknown?
Origin: ALT/Sisyphus Linux, Alexey Voinov <voins@altlinux.ru>
If the old_geometry coordinates were saved while the window
was in another xinerama head, use current coordinates as
the old ones to avoid making the window jump heads when
un-maximizing it.
This patch removes bits and pieces of xmu that for some reason
have been bundled with wmaker. I can very well imagine that at the time
they were bundled, xmu was not, was not available everywhere, or
whatever.
I couldn't come across a system that doesn't have these things in its
system-supplied x libs, though i've been carrying this since january,
and my memory is quite flakey at times.
certainly nothing xorg is at danger, nor xfree86 dating back to at
least 3.3. i believe this stuff should be everywhere that calls itself
to be on par with x11r6.3 at the very least.
it would be incredibly useful if people having access to commercial
unixes could check this on things released in the past, say, 15 years.
i believe back then i had solaris 8+and sco openserver 5+ covered.
When a maximized window is resized (either using wmaker or because the
application changed the window size), and you want to maximize it again,
you first have to unmaximize it (because wmaker thinks the window is still
maximized), and only then you can maximize it. The following patch
corrects this behaviour.
The patch "Clean up maximization and un-maximization logic" introduced
a regression wrt to the left/half maximization feature, due to a C
operator order precedence issue.
With this patch, minimized windows do not reshuffle anymore.
See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=361241
where Martin Hinsch reported:
"The behaviour of icons (minimized applications) is severely broken when
automatic placement is switched on.
* When switching the workspace icons often stick thereby covering (and
hiding) icons on the workspace one is switching to.
* After minimizing an application, clicking (singly or doubly) on one of
the icons causes all of them to reshuffle, usually in a way that the
icon in question changes its place (which is extremely annoying).
* Icon placement ignores the dock so that icons disappear behind
it/cover it.
All of these suddenly appeared about a year ago (I think with 91.0). The
problems are not architecture-specific since they occur in exactly the
same way on my pentium machine. Wiping the complete configuration (rm -r
~/GNUstep) did not make a change either."
This fixes a compilation failure with gcc-4.3.4, where it complains
nxpm.c: In function RGetImageFromXPMData:
nxpm.c:99: error: expected expression before ] token
...
There is no point in having a buffer capable of 512 simultaneous events,
as it will typically be only 1. But increase the value for the length
of the path + filename to 64 (which is a reasonable size for a
typical /home/something/GNUstep/Defaults/WMRootmenu filename).
This patch adds history to some dialog boxes. To use it replace %a with
%A in menu file (like in "Run..." menu item). You can specify third
parameter to %A to use different histories for differen dialogs.
All history files is kept in ~/GNUstep/.AppInfo/WindowMaker/ and the number of
history lines controlled by DialogHistoryLines parameter (one for
all dialogs) defaults to 500 lines.
It also adds Tab completion.
Origin: ALT/Sisyphus Linux, Alexey Voinov <voins@altlinux.ru>
This patch is from the contrib directory. It adds an option to allow starting
DockApps with a single click. It's a handy option that adds only a few lines to
the code.
It is a good feature patch example because it includes modifications
to the WPrefs application so that the feature can be easily enabled or
disabled. The one-click ability allows wmaker to integrate more seemlessly with
programs like ROX-Filer which can be configured to use single or double click
actions.
The README in the contrib/ directory says,
author:
Sebastien Bauer <seb_bauer@bigfoot.com>
John Morrissey <jwm@horde.net>
updated for Window Maker 0.65.0 by:
Daniel Richard G. <skunk@graphics.lcs.mit.edu>
updated for Window Maker 0.80.2 by:
Martial Daumas <martial@nasgaia.org>
update for 0.80.2+ by:
steve lion <steve.lion@verizon.net> and vlaad
We should not try to un-maximize the windows from inside
the function wMaximizeWindow(), as that makes no sense.
If the window is already maximized then we don't call
wMaximizeWindow() anymore and call wUnmaximizeWindow()
instead, which will use the old geometry regardless
of which maximization is active (horizontal, vertical,
maximus, etc). And the old geometry now is also saved
when we enter wMaximizeWindow().
So when we call wMaximizeWindow() or wUnmaximizeWindow()
we really mean "maximize to the specified state" or
"go back to whatever old geometry coordinates we had before".
This patch introduces the "tiled maximization" feature, a.k.a. Maximus.
By pressing the keyboard shortcut associated with Maximus, the focused
window will be maximized to the greatest area satisfying the constraint
of not overlapping existing windows.
Bug overview: right after start up, omnipresent AppIcons (living in the Clip)
are displayed but non-functionnal.
How to reproduce it: place two AppIcons in the Clip, make the first one (A)
omnipresent and leave the second one (B) as is. Make the Clip "autocollapse".
Switch to the second workspace and restart wmaker. Wmaker starts in the first
workspace; the Clip is closed (its text color corresponds to the closed state,
AppIcon B is not shown). However, AppIcon A is displayed. Moreover, A does
not react when the Clip expands or collapses. Finally, a click on A makes it
disappear. Fortunately, changing to another workspace fixes the problem
definitively.
Explanation and correction: internally, wmaker maintains as many clips as
workspaces. When the user switches to another workspace, the omnipresent
AppIcons are moved to the new "current" clip. In the situation above (trying
to reproduce the bug), when wmaker restarts, the omnipresent AppIcons are
restored in the second workspace, whereas the first workspace is active. In
the previous code, a "hack" (calling XMapWindow()) unconditionally displayed
the omnipresent AppIcons. The proposed patch makes sure the omnipresent
AppIcons are moved to the first workspace on wmaker startup.
This is a bug fix. Bug overview: if an AppIcon is moved rapidly over a Clip
set to auto-expand, the latter may erroneously auto-expand afterwards.
How to reproduce it: set a Clip to auto-collapse, and make sure it contains a
(random) AppIcon, so as to easily visualise its open/close state. Now move
rapidly an AppIcon over the Clip. Try to move it so fast that the cursor
sometimes is out of the AppIcon's tile. Then, replace the AppIcon out of the
Clip.
Explanation and correction: if, while the AppIcon was being moved, the mouse
cursor entered at least once in the Clip's tile, the latter is going to
receive an EnterNotify event (after the AppIcon is replaced) and thus expand
automatically after the relevant delay. The solution is to simply "gobble"
(i.e., ignore) all EnterNotify events when moving an AppIcon.
Moved the half_scr_width calculation to after the
usableArea = wGetUsableAreaForHead(scr, head, &totalArea, True);
call. This fixes the problem where windows that get "half screen
maximized" cover the dock or clip.
Instead of open coding the animation routines in the function bodies,
put them in a helper function called shade_animate(wwin, what), whose
second argument can be SHADE or UNSHADE.
This adds Left Half / Right Half Maximize capability to WindowMaker.
It allows you to maximize a window to only the left or right half
of your screen.
It is useful on widescreen displays where one might to bring up
two different windows side-by-side.
This patch cleans up an obvious code duplication case in the
different "placement" algorithms.
Avoid this needless repetition by using a helper function,
which in turn lets those functions a bit easier to read and
also makes wmaker 0.11% smaller :-)
[mafra@Pilar:wmaker.git]$ size src/.libs/wmaker.*
text data bss dec hex filename
619824 19160 8544 647528 9e168 src/.libs/wmaker.new
620552 19160 8544 648256 9e440 src/.libs/wmaker.old
The "Kill Application" dialog window doesn't always show the
application name when killing DockApps. It shows its wm_class
instead, which is not always the same as the app name.
This patch allows the Kill Application dialog window to show
the application name in the window -- as expressed by using the
basename of the path to the application.
[crmafra: Added wtokensplit() to get only the app name]
These are some of the fixes sent to the wmaker-dev list by
Vladimir Nadvornik, with minor modifications to address Dan
Pascu's concerns.
Original-post: http://lists.windowmaker.info/dev/msg00293.html
In commit d6c134f420 ("Do not switch
workspace to follow new windows in others") the default behavior
was changed, and workspace switching to follow focus requests was
strictly forbidden.
Although that seems to be a sane thing to do by default, that raises
concerns about whether Window Maker could be more flexible in that
respect -- allowing the user to choose which applications are or
are not allowed to do that.
This patch adds such configuration, located in the "Advanced Options"
submenu of the top-level "Attributes" menu.
New windows should only be focused if they are in the current workspace.
Not performing this check can lead to WM switching workspaces if a
window is opened which:
(a) is configured to appear in a particular workspace
and
(b) sends a _NET_ACTIVE_WINDOW message to give focus to
something, e.g. a sub window.
This behaviour was observed with firefox if a session with more
than one tab open was restored at startup because:
"If a Client wants to activate another window, it MUST send a
_NET_ACTIVE_WINDOW client message to the root window:"
Original-post: http://lists.windowmaker.info/dev/msg00442.html
[crmafra: Added comment]
The switch panel was not being destroyed with the realease of
FOCUSPREV, only with FOCUSNEXT. Fix this.
One can reproduce this bug as follows.
In the "Keyboard Shortcut Preferences" of WPrefs set "Focus next window"
to e.g. "Alt+Tab" and "Focus previous window" to "Ctrl+Tab".
The switchpanel is not destroyed if we release "Ctrl+Tab" but it
is upon releasing of "Alt+Tab".
Retrieved-from: http://git.altlinux.org/people/voins/packages/?p=WindowMaker.git;a=commit;h=51c95a55c9310f499b1fdeca138106ca7bf74423
[crmafra: Commit log]
Why?
1. The reason for its existence is to "Disable some stuff that are
duplicated in kde", and I don't think I will ever need that.
Furthermore, even the description in the configure script reads
"disable some stuff (dont use it)".
2. It makes the code uglier at some places, e.g.,
#ifdef LITE
{
#if 0
}
#endif
#else
if (!wRootMenuPerformShortcut(event)) {
#endif
which by the way is the ugliness which motivated this patch.
3. Does not even compile anymore. It fails with
CC dockedapp.o
CC event.o
event.c: In function 'executeButtonAction:
event.c:711: error: WScreen has no member named root_menu
event.c:712: error: WScreen has no member named root_menu
event.c:713: error: WScreen has no member named root_menu
event.c:715: error: WScreen has no member named root_menu
event.c:720: error: WScreen has no member named switch_menu
event.c:721: error: WScreen has no member named switch_menu
event.c:722: error: WScreen has no member named switch_menu
event.c:724: error: WScreen has no member named switch_menu
make[2]: *** [event.o] Error 1
make[1]: *** [all] Error 2
make: *** [all-recursive] Error 1
But instead of fixing this (it would be trivial), let's get
rid of the whole ugliness altogether.
It is safer if the functions are declared with proper
prototype specifications.
This patch addresses some of the warnings with -Wstrict-prototype,
like this:
event.c:105: warning: function declaration isn't a prototype
After the fix to avoid a memory leak in wtokenfree(), WPrefs crashes
when opening the 'Applications Menu Definition' dialogue.
The problem is that WPrefs code relied on the fact that the first token
in the array would not be free'd by wtokenfree(), a misbehaviour which
is correctly addressed with this patch.
Retrieved-from: http://www.mail-archive.com/ports@openbsd.org/msg12731.html
[crmafra: small changes in the commit log from the webpage]
wtokenfree() does not free the first entry of an array.
If count is 1 then
while (--count)
will be
while (0)
and the inner body of that while-loop will not be entered.
Therefore a memory leak happens every time wtokenfree() is called.
Retrieved-from: http://paldium.homeunix.org/tobias/wmaker/patches/
[crmafra: This patch, altough correct, breaks WPrefs.app, which will be
fixed by the next patch. ]
Daniel Déchelotte reported one problem with the escape handling in
the switchpanel:
"Start with two windows: a fullscreen one, and a smaller window
that appears above the fullscreen one. With the mouse, focus the
bigger window. Start alt-tabbing, then press Escape. The bigger
window will then be focused (good) *and raised* (small problem)."
Fix this by adding a test for the escape key before calling
raiseWindow().
Problems:
1.
During expansion of path, the resulting path can overflow the supplied
area of PATH_MAX+2 (buffer as well as buffer2). A tampered environment
variable can be used to modify program flow.
Proof:
[note: wmaker has been compiled with propolice]
$ export A="[tested with 4096x A]"
$ GNUSTEP_USER_ROOT="\$A\$A/\$A/\$A/" wmaker --for-real
*** stack smashing detected ***: wmaker terminated
Aborted
2.
Way too many functions handle a return value of NULL for wexpandpath
improperly, resulting in segfaults (and maybe other problems). To
prove the existance of these issues:
Proof:
$ GNUSTEP_USER_ROOT=~nouser wmaker --for-real
wmaker error: could not get password entry for user nouser: Success
Segmentation fault
Solution:
hard exit with error message about what is going on.
3.
The improper parsing of environment variables can lead to expansion
of path names that were not intended to be expanded.
(a) If a string like "$(var" is found, Window Maker tries to expand "var"
(environment variable) although the syntax is wrong.
Proof:
$ export PROOF=foo
$ GNUSTEP_USER_ROOT=/\$\(PROOF wmaker --for-real
wmaker warning: could not find user GNUstep directory (/foo/Defaults/WindowMaker).
(b) If the variable out of a) cannot be resolved, a closing bracket will be
added.
Proof:
$ unset PROOF
$ GNUSTEP_USER_ROOT=/\$\(PROOF wmaker --for-real
./wmaker warning: could not find user GNUstep directory ($(PROOF)/Defaults/WindowMaker).
Author: Tobias Stoeckmann
Retrieved-from: http://paldium.homeunix.org/tobias/wmaker/
Submitted-by: Gilbert Ashley <amigo@ibiblio.org>
There is no point in checking
if (dock_state)
if the function would have already returned with
the check if(!dock_state) right above.
Retrived from http://yo.dan.free.fr/wmaker.phtml.en
[crmafra: wrote the commit message]
Gcc-4.3.2 warns:
texture.c: In function 'wTextureMakeFunction':
texture.c:393: warning: 'fallbackColor.pixel' is used uninitialized in this function
Based on a patch by Vladimir Nadvornik.
Bug overview:
New windows are sometimes mis-placed when the "smart" algorithm is used.
How to reproduce it:
be sure (with WPrefs.app, for instance) to use the so-called "smart"
window placement algorithm. In an empty workspace, open a first window
(for instance, an xterm). The window should appear at the upper left corner.
Shade it (hide its contents below its title bar by double-clicking on it)
and select another empty workspace. Open another window. Instead of placing
it at the exact same place as the first window, wmaker places the new
window a little bit lower, or frankly to the right.
Explanation and correction:
when placing a new window, wmaker avoided all shaded windows, instead of
only avoiding the shaded windows on the active workspace.
1. Setup two windows in a workspace, one at the center and the
other at a corner. Move the mouse to the center of the screen, so
that the focus goes to the center window. Now, with the help of the
keyboard (with Alt-tab, typically), try and switch the focus to the
other window. In doing so, the switch panel shows up, gives the
focus to the other window and then disappears. However, its
disappearance make it seem to wmaker that the mouse has just
entered the center window, so wmaker gives the focus to that
window again.
2. It is a lit bit more involved. "Raise window when switching
focus with keyboard" needs to be set. In a given workspace, maximize
a first window A, then setup "above" window A two windows B and C
(one in the upper left corner and the other one in the lower right
corner, for example). Move the mouse so as to give the focus to
window B. Press the Alt key, hit the key tab once (window A moves
up to the "top"), then another time (window C is then selected).
Eventually relase the Alt key: window B is given the focus again.
Correction: it is a matter of ignoring some (EnterNotify) events
when the switch panel is active or has just been used.
Pressing the escape key (ESC) while the switchpanel is active
cancels it and gives the focus back to the window which had it
before the switchpanel was invoked.
If all windows are minimized before the switchpanel was called,
they will continue to be if ESC is pressed.
In other words, pressing ESC is like going to the parallel universe
where you never entered the switchpanel in the first place.
Based on a patch by Nicolas Bonifas from 17.08.2009.
This way we avoid code duplication in 6 places.
This changes the previous behavior of the first
instance because the helper function has an extra
CommitStacking(scr);
compared to the original code. But it should not
hurt to have it.
We can get rid of one overall tab by moving the
if(swpanel)
test -- which was done everytime in each individual case --
to the beginning, therefore encompassing all cases.
for arq in `git ls-files *.c`; do
echo $arq;
indent -linux -l115 $arq;
done
The different line break at 115 columns is because
I use a widescreen monitor :-)
GCC has an extension to deal with ranges within case statements.
Let's use it to make the code more readable and ~5% smaller,
[mafra@Pilar:wmaker.git]$ size src/event.o.*
text data bss dec hex filename
13087 0 1056 14143 373f src/event.o.new
13711 0 1056 14767 39af src/event.o.old
gcc-4.3.2 warns:
defaults.c: In function 'getModMask':
defaults.c:2586: warning: comparison of unsigned expression < 0 is always false
The line in question is
if (mask < 0)
and 'mask' is the return of wXModifierFromKey() which is an 'int' and can
indeed return a negative value (see src/xmodifier.c), so let 'mask' be
an 'int' too.
There is no point in having parameter in wDefaultsCheckDomains() and
not using it, so let's simply remove the parameter altogether and
avoid silly-looking things like
wDefaultsCheckDomains("bla");
gcc was complaining in a few places things like:
defaults.c:890: warning: comparison between signed and unsigned
so let's just make the loop counting variables be unsigned, as
the never get negative anyway.
I also spotted other two variables which can be unsigned too.
gcc-4.3.2 warns:
defaults.c:3136: warning: declaration of 'index' shadows a global declaration
/usr/include/string.h:309: warning: shadowed declaration is here
and in a few other places too. Fix this by renaming 'index' to 'widx'.
GCC warns:
event.c: In function 'handleDestroyNotify':
event.c:676: warning: declaration of 'index' shadows a global declaration
/usr/include/string.h:309: warning: shadowed declaration is here
event.c: In function 'handleKeyPress':
event.c:1435: warning: declaration of 'index' shadows a global declaration
/usr/include/string.h:309: warning: shadowed declaration is here
So let's rename "index" to "widx".
Note that there is a more serious shadowing in this same function,
event.c:1686: warning: declaration of 'wwin' shadows a previous local
but this is more complicated to solve.
inotifyHandleEvents() was allocating a buffer of size
(sizeof(struct inotify_event) + FILENAME_MAX)*1024
where FILENAME_MAX is #defined to be 4096 in stdio_lim.h, therefore
it was more than 4 MB!
Reduce it by using 16 instead of FILENAME_MAX and 512 instead of 1024.
Now valgrind does not complain about things like
Invalid write of size 8
at 0x42002F: inotifyHandleEvents (event.c:323)
by 0x7FF00020F: ???
Address 0x7febfc148 is on thread 1's stack
I also made some small coding style changes.
When investigating possible memory leaks with 'valgrind' I noticed
this leak report:
602 bytes in 13 blocks are definitely lost in loss record 77 of 128
at 0x4C2362E: malloc (vg_replace_malloc.c:207)
by 0x4576ED: wmalloc (memory.c:88)
by 0x45B4F7: wstrconcat (string.c:214)
by 0x426FC2: main (main.c:608)
which happens here,
str = wstrconcat("WMAKER_BIN_NAME=", argv[0]);
putenv(str);
There is a comment further below (in another context)
/* return of wstrconcat should not be free-ed! read putenv man page */
so this leak report is probably a false positive anyway. However,
https://www.securecoding.cert.org/confluence/display/seccode/POS34-C.+Do+not+call+putenv%28%29+with+a+pointer+to+an+automatic+variable+as+the+argument
has some nice discussion about putenv() and after reading it I decided to
use setenv() instead, making 'valgrind' happy along the way. It also
makes the code a bit simpler, avoiding a call to wstrconcat().
I also changed the name of another variable called 'str' to 'pos', just
in case.
This patch addresses this warning of gcc:
wfontpanel.c: In function 'listFamilies':
wfontpanel.c:588: warning: missing sentinel in function call
Fix it by using the 'NULL' pointer instead of 0 (zero) as the
sentinel.
For more information, see
http://www.linuxonly.nl/docs/2/2_GCC_4_warnings_about_sentinels.html
in particular,
"On most systems, there is no difference between 0 and (char *)0.
On 64 bit systems, however, the integer 0 is 32 bits and the
pointer 0 is 64 bits. The compiler does not know whether it is
an integer or a pointer, and defaults for the integer. This will
not clear the upper 32 bits and the function will not stop
scanning its parameters."
Note that here in my 64-bit Mandriva I don't need to cast (char *)NULL.
Based upon a patch by Tobias Stoeckmann
Small code size reduction,
text data bss dec hex filename
620412 19144 8544 648100 9e3a4 src/.libs/wmaker
621347 19152 8544 649043 9e753 src/.libs/wmaker.old
The WINGs-function wtokensplit does not set argv to NULL if no string has been
split - instead argc is set to 0.
You can only observe this issue if you compile Window Maker without any
optimization:
compile Window Maker with CFLAGS=""
run Window Maker and save a session
change session-file and replace a 'command xyz' line with 'command " "'
restart Window Maker
watch "Fatal error"-message
Retrieved-from: http://paldium.homeunix.org/tobias/wmaker/
Submitted-by: Gilbert Ashley <amigo@ibiblio.org>
This is a small patch to fix an issue with the switchpanel and a large number
of windows which happens only on a Xinerama setup.
When the number of open windows is so large that displaying all of them would
cause the switchpanel to be too wide for the screen, the panel is supposed to
shrink and scroll to accomodate them all.
In Window Maker 0.92.0 this works for single head displays but not for
Xinerama. The panel extends to the next head and gets garbled. This patch fixes
the issue by correctly constraining the panel to the head with the cursor.
Submitted-by: Gilbert Ashley <amigo@ibiblio.org>
A minor bug has been bothering me for a long time. When you maximize a
borderless window in Window Maker, the window ends up too narrow and too short
by two pixels.
Submitted-by: Gilbert Ashley <amigo@ibiblio.org>
wmaker keeps the names of all workspaces together in
the string 'buf' with fixed length of 1024, therefore
allowing buffer overflows if the number of workspaces
is big enough.
For the default names "Workspace X" (from 1 to 9)
and "Workspace XX" (from 10 to 99) etc, the approximate
number of workspaces necessary to make the buffer
overflow occur is 80, because
(11*9) + (71*12) + 80 = 1031
The fix is to set the size of 'buf' as
the maximum number of workspaces times their maximum
name length.
The problem was reported by John H. Robinson in the wmaker-dev
list ( http://lists.windowmaker.info/dev/msg00214.html ):
"http://www.youtube.com/watch?v=fkNJZvKwmhE
Michael reported a problem with Window Maker where it crashes with a
SIGSGV when trying to create an 82nd workspace.
/usr/local/WindowMaker-0.92.1pre/bin/wmaker warning: Window Maker exited
due to a crash (signal 11) and will be restarted.
I was able to reproduce it by making 81 workspaces, then creating an 82nd."
[ crmafra: Wrote the changelog ]
Actually, the whole thing is so chock full of signed vs unsigned
mixups and the assumption of all the world is 32-bit that i'm not sure
anymore that correcting them one by one is a good idea.
what would probably work best is that if someone deeply familiar with
the code (*looks in dan's general direction*) cleaned the data types
up, then legions of grunts like myself could start cleaning the rest
with relative confidence...
Here I tried to prevent the menu of windows to be displayed in the
other side of the screen if the window is half present in the screen
horizontally (ie x < 0)
a worse problem with menu is that it disappears when
the window is half present in the window vertically
ie y < 0, that makes it not usable.
This patch fixes the following warnings (with gcc 4.2.3)
defaults.c:980: warning: unused variable 'replace'
main.c:504: warning: 'watchPath' may be used uninitialized in this function
main.c:504: note: 'watchPath' was declared here
startup.c:616: warning: implicit declaration of function 'XrmUniqueQuark'
The handling of user defined shortcuts was not checking the length
of the shortcut before copying it to a fixed-length temporary buffer,
char buf[128];
strcpy(buf, shortcutDefinition);
and strcpy() is well known for not checking if overflows will occur.
In particular, wmaker was crashing here if a big 'shortcut' was defined
either through WPrefs or by directly editing the configuration files.
This is now avoided by using strncpy() instead.
And this patch also fixes a similar buffer overflow for big workspace
names too.
Furthermore, use MAX_SHORTCUT_LENGTH instead of raw number and define
it to be 32 instead of 128.
There may be issues with running applications in 64-bit mode when
they were written with tacit assumptions about 32-bit platforms.
For example,
* Assuming that a pointer can be cast back and forth to an integer
The reason is that the size of the integer and pointer may be different.
See the description of "[PATCH] Warn when casting a pointer (constant)
to an integer of different size." in the gcc mailing list
http://gcc.gnu.org/ml/gcc-patches/2005-12/msg01881.html
where it was also suggested the use of casts to uintptr_t. This is
what this patch does.
As a result the following warnings are fixed, leaving us with an
almost warning-free compilation in 64-bit platforms:
defaults.c:1446: warning: cast to pointer from integer of different size
defaults.c:1457: warning: cast to pointer from integer of different size
defaults.c:1471: warning: cast to pointer from integer of different size
defaults.c:1486: warning: cast to pointer from integer of different size
icon.c:67: warning: cast from pointer to integer of different size
menu.c:112: warning: cast from pointer to integer of different size
switchmenu.c:452: warning: cast from pointer to integer of different size
window.c:140: warning: cast from pointer to integer of different size
window.c:2217: warning: cast to pointer from integer of different size
workspace.c:135: warning: cast to pointer from integer of different size
workspace.c:214: warning: cast to pointer from integer of different size
workspace.c:634: warning: cast to pointer from integer of different size
workspace.c:1330: warning: cast to pointer from integer of different size
workspace.c:1514: warning: cast to pointer from integer of different size
wfilepanel.c:135: warning: cast from pointer to integer of different size
wfilepanel.c:171: warning: cast from pointer to integer of different size
wfontpanel.c:499: warning: cast to pointer from integer of different size
wfontpanel.c:500: warning: cast to pointer from integer of different size
wfontpanel.c:505: warning: cast to pointer from integer of different size
wfontpanel.c:506: warning: cast to pointer from integer of different size
wfontpanel.c:776: warning: cast from pointer to integer of different size
wfontpanel.c:777: warning: cast from pointer to integer of different size
wfontpanel.c:877: warning: cast from pointer to integer of different size
wfontpanel.c:878: warning: cast from pointer to integer of different size
wpanel.c:363: warning: cast from pointer to integer of different size
fontl.c:42: warning: cast from pointer to integer of different size
fontl.c:42: warning: cast from pointer to integer of different size
fontl.c:42: warning: cast from pointer to integer of different size
fontl.c:90: warning: cast to pointer from integer of different size
puzzle.c:138: warning: cast from pointer to integer of different size
puzzle.c:225: warning: cast to pointer from integer of different size
wtableview.c:1031: warning: cast to pointer from integer of different size
wtableview.c:1067: warning: cast to pointer from integer of different size
wtableview.c:1069: warning: cast to pointer from integer of different size
wtableview.c:1074: warning: cast to pointer from integer of different size
wtabledelegates.c:234: warning: cast from pointer to integer of different size
wtabledelegates.c:250: warning: cast from pointer to integer of different size
wtabledelegates.c:265: warning: cast from pointer to integer of different size
wtabledelegates.c:287: warning: cast to pointer from integer of different size
wtabledelegates.c:351: warning: cast from pointer to integer of different size
wtabledelegates.c:372: warning: cast from pointer to integer of different size
wtabledelegates.c:393: warning: cast from pointer to integer of different size
wtabledelegates.c:410: warning: cast to pointer from integer of different size
test.c:44: warning: cast from pointer to integer of different size
test.c:47: warning: cast to pointer from integer of different size
test.c:55: warning: cast from pointer to integer of different size
test.c:58: warning: cast from pointer to integer of different size
This patch adds the --disable-verbose-compile switch to the
configure script.
When this option is used it reduces the verbosity of compilation
messages to the essential. However compiler warnings are not affected
and thus gain visibility by not standing in the middle of a storm
of other messages.
In summary, the compilation messages are reduced to a stream of
CC array.o
CC bagtree.o
CC configuration.o
CC connection.o
CC data.o
[...]
instead of a mind-boggling flux of
gcc -DHAVE_CONFIG_H -I. -I../src -I../WINGs/WINGs -I../wrlib -I../src
-I/usr/include/freetype2 -I/usr/local/include
-DLOCALEDIR=\"/usr/local/lib/loca\"/usr/local/share/WINGs\"
-DDEBUG -fno-strict-aliasing -g -O2 -c array.c
gcc -DHAVE_CONFIG_H -I. -I../src -I../WINGs/WINGs
-I../wrlib -I../src -I/usr/include/freetype2 -I/usr/local/include
-DLOCALEDIR=\"/usr/local/lib/loca\"/usr/local/share/WINGs\"
-DDEBUG -fno-strict-aliasing -g -O2 -c bagtree.c
[...]
event handler timer.
After upgrading my kernel recently I noticed that dnotify has been
depreciated, so I decided to try and implement the new inotify code in
Window Maker instead.
During testing, I also found that one of the timers which was removed
(the one causing the most wake-ups), calling delayedAction, was
responsible for handling signals. Basically with this timer removed,
signals were only handled after an X event occurs.
After looking at the delayedAction function, I couldn't see the purpose of it.
It certainly wouldn't cause any delay as it was called by the timer every
500ms, so there is no time correlation with when a signal was received.
Also, it appeared to count the signals and call DispatchEvent for each
one, but it appears DispatchEvent would just handle the most recent signal
and take action on that. The signals handled by delayedAction are the
various exit and reset signals, so only one need to be handled. I
therefore have commented out delayedAction (it wasn't called by any other
procedure) and added a call to DispatchEvent imediately after the signal
is registered, in handleExitSig.
I'm not sure what problems this may cause with dead children - these are
only cleaned up after an Xevent now that the timer is removed- but I
haven't observed any problems since a few months ago.
Pedro Gimeno explains:
"As you most likely already know, every X Window event comes with a
timestamp. You can see it using e.g. xev.
I don't know if this is Debian-specific and I'm no X Window expert, but
the fact is that in my machine this timestamp, according to my
/usr/include/X11/X.h, is a 32-bit unsigned integer and it holds the
event time in milliseconds according to my tests.
This timestamp appears to be in milliseconds from the start of the Unix
epoch, modulo 2^32. The problem is that 32 bits are too little for this
purpose: 2^32 milliseconds are exactly 49 days, 17 hours, 2 minutes and
47.296 seconds.
Now, the WindowMaker code rejects any focus event whose timestamp is
less than the last one seen, using code similar to this:
if (timestamp < LastTimeStamp) return;
LastTimeStamp = timestamp;
This is a disaster when timestamp wraps around because of the 32-bit
limit. Say LastTimeStamp equals 2^32-1 and a focus event comes two
milliseconds after. Its timestamp will equal 1 because of the
wraparound. Obviously, 1 < 2^32-1 so the event will be rejected even if
it comes from the future (relative to LastTimeStamp) and not from the
past. Focus events are no longer accepted because nothing can be greater
than 2^32-1 (unless you click on that exact millisecond, 49 days after).
If you look in http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=102314
you'll see two people confirming the periodicity and predictability of
the bug. Arnaud Giersch was who noted that the period was 2^32
milliseconds (and not 2^31 as I thought it could be because of a signed
int involved) and who noted the coincidence with the Unix epoch.
My fix works by subtracting one timestamp from the other (an operation
which is done modulo 2^32) and rejecting the focus event only if the new
timestamp is within 1 minute (60000 milliseconds) in the past. Given the
periodicity of the timestamp, actually after 49.7 days there will be
another minute in the future during which focus events will not be
accepted; again after 99.4 days, etc., but thanks to the subtraction,
they will now be relative to the *last* focus change and not to the Unix
epoch.
The patch could be simplified to get rid of the compareTimes function.
It's written like that because it comes from the previous patch (which
you can see in the same bug report).
So the date comparison line could end up looking like this:
if (scr->flags.ignore_focus_events || LastFocusChange - timestamp <
60000)
so you can get rid of the compareTimes function. However please respect
the other change of 'int' to 'Time'.
As for why you aren't experiencing this problem, perhaps the event
timestamp does not come from the Unix epoch in your X Window system, but
rather starts counting when you start it or your computer, I don't know.
Or maybe your X Window server's timestamps are 64 bits wide instead of
32 bits, which I don't think because I doubt the X11 protocol specifies
a 64-bit record for the transmission of the timestamp. In either case,
the output of xev will probably help understanding the cause.
Or maybe even it's a click-to-focus only bug and you're not using that
mode. I haven't checked who calls wSetFocusTo.
Note that I can reproduce it just by manipulating the system clock,
which seems to be tied to the event timestamps."
For more information see:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=102314
Reported-by: Joey Hess <joeyh@debian.org>
Signed-off-by: Pedro Gimeno <parigalo@formauri.es>
WPrefs tries to use the fonts listed in the WMCreateFont() call (from WPrefs.c),
font = WMCreateFont(scr, "Lucida Sans,URW Gothic L,Times New Roman,serif"
":bold:pixelsize=26:antialias=true");
and 'font' is later used without accounting the possibility of it being NULL in
WMSetLabelFont(WPrefs.nameL, font);
WMReleaseFont(font);
In particular, WMReleaseFont(font) will kill WPrefs ungracefully with the
following message if font=NULL
WPrefs: wfont.c:193: WMReleaseFont: Assertion `font!=((void *)0)' failed.
Aborted
That happens because the return value of WMCreateFont() can be NULL, so this
patch makes WMCreateFont() never return NULL. If the font creation fails for
some reason (see below), we try to use the emergency DEFAULT_FONT and print
debugging information telling what's going on. If the use of DEFAULT_FONT
also fails, then we terminate WPrefs with exit(1) and let the user know
why exactly if failed.
This bug happened because the font "Lucida Sans" (the first possibility
in the initial call to WMCreateFont()) "exists" in my system as a stale
symbolic link to a location which no longer exists after an upgrade from
java from 1.5.0.14 to 1.5.0.15
/etc/alternatives/LucidaSansDemiBold.ttf ->
/usr/lib/jvm/java-1.5.0-sun-1.5.0.14/jre/lib/fonts/LucidaSansDemiBold.ttf
So the call to XftFontOpenName(display, scrPtr->screen, fname) with
"Lucida Sans" being the first possibility for Xft to try out ended up in 'font'
being NULL because, as far as the Xft library was concerned, "Lucida Sans"
produced a positive match (due to it existing as a symbolic link) but in the end
a NULL result was produced due to the missing symbolic link destination. This later
exposed the bug of WMCreateFont() returning font=NULL and WPrefs.c not checking
whether font=NULL before using it. Bang!
If "Lucida Sans" was the _second_ entry to try and the first one had suceeded,
this bug would not have surfaced.
This solves https://qa.mandriva.com/show_bug.cgi?id=39677
Signed-off-by: Carlos R. Mafra <crmafra@ift.unesp.br>
This patch removes wmaker from PowerTop's shame list, where
it appeared with ~3-4 wakeups/second.
It adds the linux kernel's dnotify mechanism (adapted from
the example in Documentation/dnotify.txt in the kernel source),
to detect when a configuration file in ~/GNUStep/Defaults has
changed to load it again on-the-fly. For me it usually means that
modifications to ~/GNUStep/Defaults/WMRootMenu via the 'genmenu'
script are automatically detected and loaded.
The use of dnotify makes the ancient behaviour of polling unecessary
and cuts down the wakeups count.
Other 'apparently' useless timers are also deleted and it's been almost
one year now that I use this patched exclusively without problems, so
I am pretty sure that it doesn't hurt to remove them.
The end result of all this is that wmaker generates 0 (zero) wakeups
when idle in a Linux system.
Signed-off-by: Carlos R. Mafra <crmafra@ift.unesp.br>
This also fixed an endless loop that could be entered by the previous code
in certain situations, after the patch to fix navigation/selection in a
textfiled with UTF8 chars was applied.
fix the 2 problems mentioend below:
- Fixed buggy handling of UTF8 characters in textfields in WINGs.
- Fixed segfault in WPrefs when some font description is missing from the
configuration file.
WINGs based applications should now work with both KDE and GNOME
applications (Sylvain Reynal <sreynal@nerim.net>)
- better check for the XDND protocol version when interoperating with other
applications. As it seems xdnd version 3 (which WINGs supports) and newer
are not backward compatible with xdnd version 1 and 2. This is why WINGs
applications cannot interoperate with GNUstep applications (which uses
xdnd version 2). Xdnd version 4 and 5 are backwards compatible with
version 3 though. (Sylvain Reynal <sreynal@nerim.net>)
small integer and causing unexpected color releases that can crash
Window Maker (Martin Frydl <martin.frydl@systinet.com>)
- fixed a small memory leak in WINGs/wview.c caused by not releasing the
background color of a view (Martin Frydl <martin.frydl@systinet.com>)
on 64bit platforms (Max Loparyev <max@city.veganet.ru>)
- fixed issue with icon colors on big endian platforms
(Max Loparyev <max@city.veganet.ru>)
patch based on a previous patch by Vladimir Nadvornik <nadvornik@suse.cz>
belong to, that works without auto-arranging the icons, while still avoiding
multiple icons in the same spot by moving the miniwindows to a new slot if
their old slot was occupied in the meantime.
- double clicking an appicon will also raise the miniwindows that belong to
that application to the front (along with the normal windows).
- Fixed WPrefs path in the installed WMState (changed after recent WPrefs
installation path changes)
- Fixed wmaker.inst (apparently some very noisy individual was unable to get
right his 10 line patch to rename a path)
the old .AppInfo/... to the new Library/WindowMaker/... on the fly when
Window Maker starts.
This should allow a transparent transition without any need for users to do
anything.
in the first place in his trivial 10 line patch which only attempted to
replace one path with another.
He must be a big time user of his own patch, to not have noticed where the
paths pointed after his change.
than 10 parameters in inline asm, else disable asm code.
This should enable it to work on any x86 platform (not just linux as before)
given that a recent enough gcc is available (no more need to guess by looking
at $host)
This should also fix the issues of compiling wrlib with gcc-2.95
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
