This patch is just adding a single header, but because it also modifies
all the C files to add the #include, it was made as a patch on its own to
ease review.
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>
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>
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.
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>
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>
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 :-)
- Fixed bug in icon chooser dialog that could cause a segmentation fault
in some cases (Pascal Hofstee <caelian@gmail.com>)
- Fixed crash in asm code in wrlib, with new versions of gcc.
- Fixed bug in the x86_PseudoColor_32_to_8() function which incorrectly
used the r, g, b fields in the conversion.
- Fixed x86 ASM code in wrlib to work on 64 bit architectures.
- Fixed the focus flicker seen with some apps (notably gtk2)
(Alexey Spiridonov <snarkmaster@gmail.com>)
- Fixed all crashing bugs that were generated by wmaker starting with the
WMState file missing.
- Added NetWM support (a modified version of the patch originaly written
by Peter Zijlstra <a.p.zijlstra@chello.nl>)
- Applied patch to enhance the Virtual Desktop behaviour, and to integrate
it with the NetWM code (Peter Zijlstra <a.p.zijlstra@chello.nl>)
- Applied a few xinerama and placement fixes (Peter Zijlstra
<a.p.zijlstra@chello.nl>)
- Fixed memory leak in dock code.
- Fixed and enhanced the text wrapping in WINGs.
- Fixed the layout of some elements in WPrefs.app
- Added workaround for aplications that don't set the required hints on the
client leader window, but they set them on normal windows (observer with
KDE 3.3.0 mainly). This will allow these apps to get an appicon again.
(they should be fixed still)
- Added workaround for applications that do not set a command with
XSetCommand(), but instead they set the _NET_WM_PID property. This works
with operating systems that offer a /proc interface similar to what linux
has. (This also is to fix problems with KDE 3.3.0 apps, but not only them).
- Fixed bug with autostart and exit scripts not being executed if user
GNUstep path was different from ~/GNUstep (when setting GNUSTEP_USER_ROOT)
- Added utf8 support in WINGs (removed old X core font code)
- Added utility to convert old font names to new font names in style files
WMDrawString() and WMDrawImageString() now take WMColor instead of GC as
arguments. WMDrawImageString() receives 2 colors (text & background).
This is to allow easy extension for Xft/Xrender and hide X low level details
- Added alpha channel to WMColor. 2 new functions also:
WMCreateRGBAColor() and WMSetColorAlpha()
- Miscelaneous code cleanups in wtext.c
- Removed obsoleted acconfig.h and implemented its functionality using
AC_DEFINE and AC_DEFINE_UNQUOTED as autoconf 2.5x recommends.
This will definitely enforce the need to use autoconf 2.5x
- Separated the font caches for normal fonts and fontsets in WINGs (they can
have the same names and collide in the cache giving unwanted results)
- Updated the years in the copyright notices
- Fixed wrong display of images with alpha in StaticGray and GrayScale visuals
- Hermes lib is used now only to convert if the visual is TrueColor and no
dithering is necesarry. This is because currently hermeslib doesn't
support to convert to an indexed destination image (so it can't convert to
PseudoColor, StaticGray and GreyScale visuals). It can convert to
StaticColor since this visual uses masks as the TrueColor visual, but
without dithering. Also hermeslib only supports dithering for just 2
combinations of source/destination bits/masks, none of which are useful
for wrlib, so no conversion that needs dithering is currently done
through hermeslib.
The fix still doesn't look right (hermes seems to do weird things internally,
and there is no documentation for it)
People with big endian machines please test if it works for you
(install hermes lib first).
Then try to start wmaker in different screen depths (15, 16, 24, 32)
and check if there are depths that do not work (either crash, or
display other colors than you expect).
Little endian machines seem ok.
PowerPC architecture, because on LinuxPPC char is unsigned by default, not
signed like on the other platforms).
Bug fixed by Philip Derrin <philipd@student.unsw.edu.au>
- miscelaneous bug fixes
We would like people with cvs access experimenting the white 'speckles' on
images to test if they still have the problem.
code like
*ptr++ = *ptr++ = *ptr++ = color;
is wrong, because there is no guarantee that ptr will be incremented
_between_ the assignment operations. it can be incremented after all
assignment operations as well. Because of this both of these are valid
implementations for a compiler:
a. assign, increment, assign, increment, assign, increment
b. assign, assign, assign, increment by 3
In case b. only the first memory location of the 3 will be modified, being
assigned 3 times the same value, while the other 2 remain unchanged.
For example egcs-2.91.66 (and possibly gcc-2.95.x too) implement this in
the second way (like in case b.)
Also the order in which the assignement is made is undefined (left to right
or right to left).
this fixed the problem we had with greyscale jpegs showing up in red,
and possibly other problems related to pseudocolor and greyscale displays.
