| Age | Commit message (Collapse) | Author |
|---|
|
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
|
|
Includes adding many $(AM_V_GEN) calls for the AM_SILENT_RULES included
by XORG_DEFAULT_OPTIONS when automake-1.11 or newer is used
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
Allows parallel make and simpler build logs/error reporting
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
libXt uses IceConnectionNumber() and IceProcessMessages(), so make sure
we link against -lICE.
Signed-off-by: Julien Cristau <jcristau@debian.org>
|
|
At 36e9f0d351afbf7fd2595990b2d39e7c551f6420, a fix was added to use
the host gcc instead of the target gcc when cross-compiling
libXt. This fix works, but is not solve the whole problem: the CFLAGS
and LDFLAGS used with the host compilers are the one defined for the
target compiler (and the flags for both compilers might be very
different).
This new fix let libXt obey to CFLAGS_FOR_BUILD and LDFLAGS_FOR_BUILD
environment variables, and use them to compile the host tools in
util/.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
|
|
|
|
Commit c59dd27ecb1751f0b097046b2f892028e5a10a3e replaces __CYGWIN__ with WIN32
in a preprocessor check, as the code is appropriate for both Cygwin and Mingw.
But this isn't quite the right was to do this. Being a Win32 platform is
Cygwin's secret shame, and WIN32 isn't actually defined to avoid turning on
code specific to native-Win API ports
(See http://www.cygwin.com/faq/faq.programming.html#faq.programming.preprocessor)
Yes, this probably means WIN32 could bereplaced by __MINGW32__ everywhere in the
X sources...
|
|
Evaluation from original bug report against Solaris 2.6 in 1996:
According to my investigation with the problem using debuggable
libXt and libXm, keycode 0 keypress events, which are given to
Xt's Translation Manager when inputting any cyrillic/greek characters,
are not cached in Xt's Translation Manager. At the TRANSLATE macro
in the Xt's source lib/Xt/TMkey.c, keycode 0 keyevents are always
given to XtTranslateKeycode() and it consumes a long time to parse
Motif's Translation Manager tables as many times as keycode 0
keyevent are given.
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
Don't try to printf if returned string is NULL
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
|
|
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com>
|
|
Most "compiler" warnings were actually sparse warnings, due to
assigning a integer to a pointer, or an external symbol without
a previous declaration.
|
|
<http://bugs.freedesktop.org/show_bug.cgi?id=17942>
|
|
|
|
makestrs is used during the build, so use the native compiler (either
gcc or cc in $PATH, can be overridden as $CC_FOR_BUILD) instead of the
(possibly cross-) compiler.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
For some reason, when Solaris moved to X11R6.0, the intelabi version
of StringDefs.c/.h was used on SPARC, so now we're stuck with 10+ years
of binaries built to use it.
|
|
|
|
|
|
select() can modify the fd set, so be a bit pessimistic and recalculate
it more often.
|
|
|
|
|
|
|
|
|
|
<https://bugs.freedesktop.org/show_bug.cgi?id=9514>
|
|
|
|
build environment with settings for chown and make not work with
permissions of the builder. So use mkdir -p and cp instead. This
is for my bug #6599.
|
|
|
|
With autoconf 2.60, $datadir now expands to use $datarootdir rather than just
$prefix, so the .pc referenced the undefined $datarootdir. AC_DEFINE_DIR
expands it all out for us.
|
|
|
|
|
|
The function "GetResources" will return a copy of resources to the caller
and the caller has the responsibility to free the returned copy. However some
functions, such as "XtGetSubresources", "_XtGetResources", and
"XtGetApplicationResources" invoked "GetResources" without freeing the
returned resource copy.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
not created correctly when MANDIR & MANSUFFIX don't match.
|
|
|
