Bug#907152: vulkan: Porting to non-linux systems
Source: vulkan-loader Followup-For: Bug #907152 X-Debbugs-Cc: tjaal...@debian.org Control: tag -1 = upstream fixed-upstream Hi! The vulkan source changed a while since this bug was originally reported: it was split into components, with this bug reassigned to to the loader, and in it was ported to more architectures. I recently ported this to Hurd again, with good results: the upstream tests pass, and it was reviewed and merged upstream: https://github.com/KhronosGroup/Vulkan-Loader/pull/1244 Because of this, I submitted the enablement of src:vulkan-loader to non-Linux architectures, as the actual Hurd porting will come in soon in new upstream releases (I guess in versions greater than 1.3.250): https://salsa.debian.org/xorg-team/vulkan/vulkan-loader/-/merge_requests/14 I'm *not* submitting the upstream Hurd patch to downstream backport: the code upstream was refactored, so a good part of it would need to be rewritten, which is not that worth of effort for something that has never been available yet. -- Pino
Bug#951754: libxkbcommon: allow to compile again on !linux archs
Source: libxkbcommon Version: 0.9.1-1 Severity: important Tags: patch Control: found -1 0.10.0-1 Hi, starting from 0.9.1-1, libxkbcommon cannot be built on non-Linux architectures due to the unconditional Wayland-related build dependencies. Considering that Wayland is Linux-specific, and the usage of it is only in a demo/test, it is possible to restrict its usage only on Linux architectures. Patch attached for this. Thanks, -- Pino --- a/debian/control +++ b/debian/control @@ -13,10 +13,10 @@ Build-Depends: meson, pkg-config, quilt, - libwayland-dev, + libwayland-dev [linux-any], libx11-dev, libxcb-xkb-dev (>= 1.10), - wayland-protocols, + wayland-protocols [linux-any], x11-xkb-utils, x11proto-core-dev, x11proto-kb-dev (>= 1.0.5), --- a/debian/rules +++ b/debian/rules @@ -1,9 +1,15 @@ #!/usr/bin/make -f +include /usr/share/dpkg/architecture.mk + +ifneq ($(DEB_HOST_ARCH_OS),linux) + extra_meson_args+="-Denable-wayland=false" +endif + # We need to point to xkb-data's files. The default should be OK but # let's be cautious: override_dh_auto_configure: - dh_auto_configure -- -Dxkb-config-root=/usr/share/X11/xkb + dh_auto_configure -- -Dxkb-config-root=/usr/share/X11/xkb $(extra_meson_args) # Kill *.la files, and forget no-one: override_dh_install:
Bug#760621: libxkbcommon: please enable parallel building
Source: libxkbcommon Version: 0.4.3-1 Severity: wishlist Tags: patch Hi, libxkbcommon seems to build fine with multiple build jobs when building. Thus, my suggestion is to enable the parallel build (with the --parallel option of dh) to speed up the compilation when requested (see also Policy §4.9.1). Thanks, -- Pino --- a/debian/rules +++ b/debian/rules @@ -14,7 +14,7 @@ override_dh_makeshlibs: dh_makeshlibs -- -c4 %: - dh $@ --with autoreconf + dh $@ --parallel --with autoreconf # For maintainer use only, generate a tarball
Bug#760624: libxkbcommon: FTBFS on !linux archs
Source: libxkbcommon Version: 0.4.3-1 Severity: serious Tags: patch Justification: fails to build from source Control: forwarded -1 https://bugs.freedesktop.org/show_bug.cgi?id=83551 Hi, libxkbcommon 0.4.3-1 fails to build on non-Linux architectures, because of an unconditional linux/input.h include. I just reported (with a patch) the issue upstream, see [1]. [1] https://bugs.freedesktop.org/show_bug.cgi?id=83551 Thanks, -- Pino --- a/test/x11comp.c +++ b/test/x11comp.c @@ -28,7 +28,6 @@ #include signal.h #include sys/types.h #include sys/wait.h -#include linux/input.h #include test.h #include xkbcommon/xkbcommon-x11.h
Bug#749286: xserver-xorg-input-synaptics: unbuildable on !linux archs
Source: xserver-xorg-input-synaptics Version: 1.8.0-1~exp1 Severity: serious Tags: patch Justification: fails to build from source (but built successfully in the past) Hi, xserver-xorg-input-synaptics 1.8.0-1~exp1 cannot be built on non-Linux archs because of the libevdev-dev B-D (which is Linux-specific). The attached patch limits the libevdev-dev B-D as linux-any, as it is used only for the eventcomm Linux backend. Thanks, -- Pino --- a/debian/control +++ b/debian/control @@ -9,7 +9,7 @@ Build-Depends: libx11-dev, libxext-dev, libxi-dev (= 2:1.2.0), - libevdev-dev, + libevdev-dev [linux-any], x11proto-core-dev, xserver-xorg-dev (= 2:1.12), pkg-config,
Bug#742335: mesa: extra libegl1-mesa-drivers dependency recommend on Hurd
Source: mesa
Version: 10.1.0-4
Severity: important
Tags: patch
User: debian-h...@lists.debian.org
Usertags: hurd
Hi,
thanks for the various egl/gles fixes (#729260, #741572, and the other
packaging commits) on Hurd!
The only left issue is that libegl1-mesa-dev depends on
libegl1-mesa-drivers, which is not built on Hurd (and libegl1-mesa
recommends it).
Attached there is a simple patch to remove such dependency and recommend
on hurd-any, which should make libegl1-mesa-dev finally installable.
Thanks,
--
Pino
--- a/debian/control
+++ b/debian/control
@@ -252,7 +252,7 @@ Architecture: any
Depends:
${shlibs:Depends},
${misc:Depends},
-Recommends: libegl1-mesa-drivers
+Recommends: libegl1-mesa-drivers [!hurd-any]
Provides: libegl1-x11
Conflicts: libegl1-x11
Replaces: libegl1-x11
@@ -287,7 +287,7 @@ Section: libdevel
Architecture: any
Depends:
libegl1-mesa (= ${binary:Version}),
- libegl1-mesa-drivers (= ${binary:Version}),
+ libegl1-mesa-drivers (= ${binary:Version}) [!hurd-any],
libdrm-dev (= 2.4.52) [!hurd-any],
x11proto-dri2-dev (= 2.6),
x11proto-gl-dev (= 1.4.14),
Bug#724675: xserver-xorg-video-savage: 1:2.3.7-1 is really version 2.3.6
Source: xserver-xorg-video-savage Version: 1:2.3.7-1 Severity: important Hi, while investigating the FTBFS of 1:2.3.7-1 on hurd-i386, I discovered that basically this version is the orig tarball of 2.3.7... but all the 2.3.6..2.3.7 changes are reverted in the debian.diff.gz file, making it effectively as 2.3.6. Please fix by removing the extra differences that downgrade the driver, and please be more careful when doing uploads. Thanks, -- Pino -- To UNSUBSCRIBE, email to debian-x-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130926133836.8943.92531.reportbug@drak
Bug#720736: xorg-server: FTBFS on hurd-i386: implicit declarations as errors with GCC 4.8
forwarded 720736 http://lists.x.org/archives/xorg-devel/2013-August/037628.html thanks In data mercoledì 28 agosto 2013 12:10:29, Julien Cristau ha scritto: On Sun, Aug 25, 2013 at 00:40:05 +0200, Pino Toscano wrote: Source: xorg-server Version: 2:1.12.4-6 Severity: important Tags: patch User: debian-h...@lists.debian.org Usertags: hurd Hi, xorg-server fails to compile on Hurd with GCC 4.8. The problem is basically similar to #701372, just happening on parts compiled only on Hurd. Attached there is a patch fixing all the issues: - os/access.c: moves the arpa/inet.h include from the defined(SIOCGIFCONF) block up in the source, so it applies for every !win32 system - hw/xfree86/os-support/hurd/hurd_init.c: - hw/xfree86/os-support/hurd/hurd_mmap.c: - hw/xfree86/os-support/hurd/hurd_video.c: include hurd.h, needed for Hurd's get_privileged_ports. The patch applies fine and work for both the version currently in sid (2:1.12.4-6) and in experimental (2:1.14.2.901-2). Thanks! Can you also send it to xorg-de...@lists.x.org? Done, http://lists.x.org/archives/xorg-devel/2013-August/037628.html -- Pino Toscano signature.asc Description: This is a digitally signed message part.
Bug#720736: xorg-server: FTBFS on hurd-i386: implicit declarations as errors with GCC 4.8
Source: xorg-server
Version: 2:1.12.4-6
Severity: important
Tags: patch
User: debian-h...@lists.debian.org
Usertags: hurd
Hi,
xorg-server fails to compile on Hurd with GCC 4.8.
The problem is basically similar to #701372, just happening on parts
compiled only on Hurd.
Attached there is a patch fixing all the issues:
- os/access.c:
moves the arpa/inet.h include from the defined(SIOCGIFCONF) block up
in the source, so it applies for every !win32 system
- hw/xfree86/os-support/hurd/hurd_init.c:
- hw/xfree86/os-support/hurd/hurd_mmap.c:
- hw/xfree86/os-support/hurd/hurd_video.c:
include hurd.h, needed for Hurd's get_privileged_ports.
The patch applies fine and work for both the version currently in
sid (2:1.12.4-6) and in experimental (2:1.14.2.901-2).
Thanks,
--
Pino
--- a/os/access.c
+++ b/os/access.c
@@ -163,6 +163,10 @@
/* #endif */
#endif
+#if defined(IPv6) defined(AF_INET6)
+#include arpa/inet.h
+#endif
+
#endif /* WIN32 */
#define X_INCLUDE_NETDB_H
@@ -461,10 +465,6 @@
#endif
#if defined(IPv6) defined(AF_INET6)
-#include arpa/inet.h
-#endif
-
-#if defined(IPv6) defined(AF_INET6)
static void
in6_fillscopeid(struct sockaddr_in6 *sin6)
{
--- a/hw/xfree86/os-support/hurd/hurd_init.c
+++ b/hw/xfree86/os-support/hurd/hurd_init.c
@@ -42,6 +42,7 @@
#include sys/file.h
#include assert.h
#include mach.h
+#include hurd.h
int
xf86ProcessArgument(int argc, char **argv, int i)
--- a/hw/xfree86/os-support/hurd/hurd_mmap.c
+++ b/hw/xfree86/os-support/hurd/hurd_mmap.c
@@ -27,6 +27,7 @@
#includemach.h
#includedevice/device.h
#includemach/machine/mach_i386.h
+#include hurd.h
#include X11/X.h
--- a/hw/xfree86/os-support/hurd/hurd_video.c
+++ b/hw/xfree86/os-support/hurd/hurd_video.c
@@ -28,6 +28,7 @@
#include mach.h
#include device/device.h
#include mach/machine/mach_i386.h
+#include hurd.h
#include X11/X.h
#include input.h
Bug#719493: libqt5gui5:i386 uninstallable
Alle martedì 13 agosto 2013, Koos Vriezen ha scritto: On Mon, Aug 12, 2013 at 03:40:56PM +0200, Pino Toscano wrote: reassign 719493 xkb-data forcemerge 677884 719493 thanks Are you sure that an importand (or is it grave) bug like uninstallable deb file should be merged with a normal bug report? Surely this is not grave at all not any other kind of release critical. Considering this is another case of #677884, merging is just the right thing to do. If you want to argue for a severity=important for #677884 (or ask for a fix), please followup on that bug report. Thanks, -- Pino Toscano signature.asc Description: This is a digitally signed message part.
Bug#719493: libqt5gui5:i386 uninstallable
Alle martedì 13 agosto 2013, Koos Vriezen ha scritto: On Tue, Aug 13, 2013 at 02:54:32PM +0200, Pino Toscano wrote: Alle martedì 13 agosto 2013, Koos Vriezen ha scritto: On Mon, Aug 12, 2013 at 03:40:56PM +0200, Pino Toscano wrote: reassign 719493 xkb-data forcemerge 677884 719493 thanks Are you sure that an importand (or is it grave) bug like uninstallable deb file should be merged with a normal bug report? Surely this is not grave at all not any other kind of release critical. Considering this is another case of #677884, merging is just the right thing to do. If you want to argue for a severity=important for #677884 (or ask for a fix), please followup on that bug report. I already had rebuilt xkd-data myself because I needed to test this setup. It's more that I noticed that merged bug reports don't show up in the appriopriate mailing lists, and thus unnoticed. I'm not sure which mailing list you are referring to. #677884 is more or less in a 'who needs it' state. Then followup there and point out the dependency chain different than install a foreign xorg server? Anyhow, maybe I can use this correspondence This is getting off-topic now though; I will reply to you, but for more user-like questions please email us at debian-qt-...@lists.debian.org. to mention another problem we have with qtbase5-dev in Jessie, it's missing qplatformmenu.h, qplatformnativeinterface.h and qplatformintegration.h /usr/include/qt5/QtGui/5.1.0/QtGui/qpa/ for our product. Do you have any suggestion how this is handled, e.g. should I file a br or is this deliberate left out? You want qtbase5-private-dev, since you are asking for such private headers. -- Pino Toscano signature.asc Description: This is a digitally signed message part.
Bug#717723: libxkbcommon: FTBFS on hurd-i386: unconditional usage of PATH_MAX
Source: libxkbcommon
Version: 0.3.1-1
Severity: important
Tags: patch fixed-upstream
User: debian-h...@lists.debian.org
Usertags: hurd
Control: forwarded -1 https://bugs.freedesktop.org/show_bug.cgi?id=67229
Hi,
libxkbcommon 0.3.1 does not compile on GNU/Hurd [1].
The problem is the unconditional usage of PATH_MAX, optional in POSIX
and not provided on the Hurd.
I reported the problem upstream [2] with a patch, which got accepted.
Could you please backport it? It applies fine to the released 0.3.1.
[1]
https://buildd.debian.org/status/fetch.php?pkg=libxkbcommonarch=hurd-i386ver=0.3.1-1stamp=1370737530
[2] https://bugs.freedesktop.org/show_bug.cgi?id=67229
Thanks,
--
Pino
From b0ad0cbf99c835ff346f0cd0f3aa0d2e45686923 Mon Sep 17 00:00:00 2001
From: Pino Toscano toscano.p...@tiscali.it
Date: Wed, 24 Jul 2013 10:05:02 +0200
Subject: [PATCH] Get rid of the usage of PATH_MAX
PATH_MAX is optional in POSIX, so avoid its unconditional usage
allocating and freeing buffers as needed.
To avoid too many malloc/free in the for loop in FindFileInXkbPath,
a buffer is grown according to the size needed at each iteration.
---
src/xkbcomp/include.c | 34 +++---
test/common.c | 41 ++---
test/test.h | 2 +-
3 files changed, 62 insertions(+), 15 deletions(-)
diff --git a/src/xkbcomp/include.c b/src/xkbcomp/include.c
index b4a4014..280bbbd 100644
--- a/src/xkbcomp/include.c
+++ b/src/xkbcomp/include.c
@@ -199,17 +199,34 @@ FindFileInXkbPath(struct xkb_context *ctx, const char *name,
{
unsigned int i;
FILE *file = NULL;
-char buf[PATH_MAX];
+char *buf = NULL;
const char *typeDir;
+size_t buf_size = 0, typeDirLen, name_len;
typeDir = DirectoryForInclude(type);
+typeDirLen = strlen(typeDir);
+name_len = strlen(name);
for (i = 0; i xkb_context_num_include_paths(ctx); i++) {
-int ret = snprintf(buf, sizeof(buf), %s/%s/%s,
- xkb_context_include_path_get(ctx, i),
- typeDir, name);
-if (ret = (ssize_t) sizeof(buf)) {
-log_err(ctx, File name (%s/%s/%s) too long\n,
+size_t new_buf_size = strlen(xkb_context_include_path_get(ctx, i)) +
+ typeDirLen + name_len + 3;
+int ret;
+if (new_buf_size buf_size) {
+void *buf_new = realloc(buf, new_buf_size);
+if (buf_new) {
+buf_size = new_buf_size;
+buf = buf_new;
+} else {
+log_err(ctx, Cannot realloc for name (%s/%s/%s)\n,
+xkb_context_include_path_get(ctx, i), typeDir, name);
+continue;
+}
+}
+ret = snprintf(buf, buf_size, %s/%s/%s,
+ xkb_context_include_path_get(ctx, i),
+ typeDir, name);
+if (ret 0) {
+log_err(ctx, snprintf error (%s/%s/%s)\n,
xkb_context_include_path_get(ctx, i), typeDir, name);
continue;
}
@@ -242,11 +259,14 @@ FindFileInXkbPath(struct xkb_context *ctx, const char *name,
xkb_context_failed_include_path_get(ctx, i));
}
+free(buf);
return NULL;
}
if (pathRtrn)
-*pathRtrn = strdup(buf);
+*pathRtrn = buf;
+else
+free(buf);
return file;
}
diff --git a/test/common.c b/test/common.c
index 7b4ee00..796904e 100644
--- a/test/common.c
+++ b/test/common.c
@@ -138,13 +138,22 @@ test_key_seq(struct xkb_keymap *keymap, ...)
return ret;
}
-const char *
+char *
test_get_path(const char *path_rel)
{
-static char path[PATH_MAX];
+char *path;
+size_t path_len;
const char *srcdir = getenv(srcdir);
-snprintf(path, PATH_MAX - 1,
+path_len = strlen(srcdir ? srcdir : .) +
+ strlen(path_rel ? path_rel : ) + 12;
+path = malloc(path_len);
+if (!path) {
+fprintf(stderr, Failed to allocate path (%d chars) for %s\n,
+(int) path_len, path);
+return NULL;
+}
+snprintf(path, path_len,
%s/test/data/%s, srcdir ? srcdir : .,
path_rel ? path_rel : );
@@ -155,10 +164,15 @@ char *
test_read_file(const char *path_rel)
{
struct stat info;
-char *ret, *tmp;
+char *ret, *tmp, *path;
int fd, count, remaining;
-fd = open(test_get_path(path_rel), O_RDONLY);
+path = test_get_path(path_rel);
+if (!path)
+return NULL;
+
+fd = open(path, O_RDONLY);
+free(path);
if (fd 0)
return NULL;
@@ -195,6 +209,7 @@ test_get_context(enum test_context_flags test_flags)
{
enum xkb_context_flags ctx_flags;
struct xkb_context *ctx;
+char *path;
ctx_flags = XKB_CONTEXT_NO_DEFAULT_INCLUDES;
if (test_flags CONTEXT_ALLOW_ENVIRONMENT_NAMES) {
@@ -212,7 +227,12 @@ test_get_context(enum
Bug#715560: libxkbcommon: please re-enable the tests on !linux archs
Source: libxkbcommon Version: 0.3.1-1 Severity: wishlist Tags: patch Hi, the upload of libxkbcommon 0.3.1-1, among the other things, completely disables the tests (not even having them non-fatal) on non-Linux architectures. This has been done in commit b751b42 [1] in packaging's Git repository by Timo Aaltonen, with no whatsoever explanation nor comment on the reasons about this. Timo? In the meanwhile, I tried a build with them enabled on kfreebsd-i386 and hurd-i386 (the latter after having been fixed), and they passed on both. Attached there is the patch reverting that commit and thus re-enabling tests on non-Linux archs. [1] b751b429470a7b7cd4515167ec120fa4e5d728c9 Thanks, -- Pino --- a/debian/rules +++ b/debian/rules @@ -1,7 +1,5 @@ #!/usr/bin/make -f -DEB_HOST_ARCH_OS ?= $(shell dpkg-architecture -qDEB_HOST_ARCH_OS) - # We need to point to xkb-data's files. The default should be OK but # let's be cautious: override_dh_auto_configure: @@ -15,10 +13,6 @@ override_dh_install: override_dh_makeshlibs: dh_makeshlibs -- -c4 -ifneq ($(DEB_HOST_ARCH_OS), linux) -override_dh_auto_test: -endif - %: dh $@ --with autoreconf
Re: Bug#660940: usbfs error causes kdm to not load on startup or reboot
Hi debian-x people, Alle giovedì 23 febbraio 2012, david mckisick ha scritto: Update to this. The issue is related to - xf86OpenConsole: VT_WAITACTIVE failed: Interrupted system call. https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/441653 For now I have inserted a sleep 1 just before the kdm start line. This (bug #660940) would seem a xserver-xorg issue, fixed with upstream commit 88c4622b594a1725d0cee86bc82ad640d241c520 (part of 1.10.99.901 and onward). Would you be able to confirm that could be indeed the case, or anyway provide hints about the issue? Thanks, -- Pino Toscano signature.asc Description: This is a digitally signed message part.
Bug#672572: xserver-xorg-input-synaptics: unbuildable on !linux archs
Package: xserver-xorg-input-synaptics Version: 1.6.0-1 Severity: important Tags: patch Hi, xserver-xorg-input-synaptics 1.6.0 cannot be built on non-Linux architectures due to the unconditional libmtdev-dev build dependency. Such library is an optional dependency of the eventcomm backend only, which is Linux-specific. Thus, the libmtdev-dev build dependency can just be limited to linux archs (patch attached). Thanks, -- Pino --- a/debian/control +++ b/debian/control @@ -15,7 +15,7 @@ pkg-config, quilt, xutils-dev (= 1:7.5+4), - libmtdev-dev, + libmtdev-dev [linux-any], Standards-Version: 3.9.2 Vcs-Git: git://git.debian.org/git/pkg-xorg/driver/xserver-xorg-input-synaptics Vcs-Browser: http://git.debian.org/?p=pkg-xorg/driver/xserver-xorg-input-synaptics.git
Bug#672546: xserver-xorg-video-openchrome: FTBFS on hurd-i386: unconditional libdrm requirement
Package: xserver-xorg-video-openchrome Version: 1:0.2.904+svn1050-1 Severity: important Tags: patch User: debian-h...@lists.debian.org Usertags: hurd Hi, currently, xserver-xorg-video-openchrome does not build on GNU/Hurd. The problem is that libdrm is considered an hard requirement (by the buildsystem and the Debian packaging), while it is only if the DRI support is enabled (and indeed there seems to be only libdrm usage in the DRI-related code). Attached there are patches to fix this: - no-libdrm.diff: do not check for libdrm among the base Xorg packages; libdrm will be searched later as mandatory requirement if the DRI support is enabled - debian.diff: disable the libdrm-dev build dependency on Hurd architectures Thanks, -- Pino --- a/configure.ac +++ b/configure.ac @@ -70,7 +70,7 @@ XORG_DRIVER_CHECK_EXT(XF86DRI, xextproto XORG_DRIVER_CHECK_EXT(DPMSExtension, xextproto) # Checks for pkg-config packages -PKG_CHECK_MODULES(XORG, [xorg-server xproto fontsproto libdrm $REQUIRED_MODULES]) +PKG_CHECK_MODULES(XORG, [xorg-server xproto fontsproto $REQUIRED_MODULES]) PKG_CHECK_MODULES(XEXT, [xextproto = 7.0.99.1], HAVE_XEXTPROTO_71=yes; AC_DEFINE(HAVE_XEXTPROTO_71, 1, [xextproto 7.1 available]), HAVE_XEXTPROTO_71=no) --- a/debian/control +++ b/debian/control @@ -17,7 +17,7 @@ Build-Depends: x11proto-xf86dri-dev, x11proto-video-dev, x11proto-gl-dev, - libdrm-dev ( 2.0), + libdrm-dev ( 2.0) [!hurd-any], libx11-dev, libgl1-mesa-dev | libgl1-dev, libxvmc-dev, HAVE_XEXTPROTO_71=no)
Re: Help needed for mesa
Hi, Alle domenica 6 febbraio 2011, Cyril Brulebois ha scritto: xorg-server 1.9 build-depends on mesa with a version higher than the one available in sid, and both mesa 7.9 and 7.10 FTBFS on non-Linux for now. I'm not sure when I'll have time to work on fixing those FTBFS, so I'd appreciate if somebody could have a look. You can send patches to debian-x@ or to the BTS. Some months ago I managed to compile mesa 7.8.2 (from exp) on Hurd. I just rebased the patches I did, and started a build of mesa 7.10 on Hurd. (I was waiting for mesa to compile on kfreebsd.) -- Pino Toscano signature.asc Description: This is a digitally signed message part.
Re: Help needed for mesa
Alle mercoledì 9 febbraio 2011, Pino Toscano ha scritto:
Alle domenica 6 febbraio 2011, Cyril Brulebois ha scritto:
xorg-server 1.9 build-depends on mesa with a version higher than
the one available in sid, and both mesa 7.9 and 7.10 FTBFS on
non-Linux for now. I'm not sure when I'll have time to work on
fixing those FTBFS, so I'd appreciate if somebody could have a
look. You can send patches to debian-x@ or to the BTS.
Some months ago I managed to compile mesa 7.8.2 (from exp) on Hurd. I
just rebased the patches I did, and started a build of mesa 7.10 on
Hurd.
Got the same issue on the buildds, which then must be something new
after mesa 7.8.2: when building the dri flavour, it enables glx
(configure.ac:1075 and on) compiling the dri2 eglx driver, which fails.
I'm attaching the other patches I have, most probably something in the
Debian side is missing.
(Maybe the change of 10_hurd.diff could just become a check for
PIPE_OS_UNIX instead of most/all of the current PIPE_OS_*.)
--
Pino Toscano
--- a/src/gallium/auxiliary/os/os_time.c
+++ b/src/gallium/auxiliary/os/os_time.c
@@ -37,7 +37,7 @@
#if !defined(PIPE_OS_EMBEDDED)
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
# include sys/time.h /* timeval */
#elif defined(PIPE_SUBSYSTEM_WINDOWS_DISPLAY)
# include windows.h
--- a/src/gallium/auxiliary/os/os_thread.h
+++ b/src/gallium/auxiliary/os/os_thread.h
@@ -40,7 +40,7 @@
#include util/u_debug.h /* for assert */
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
#include pthread.h /* POSIX threads headers */
#include stdio.h /* for perror() */
@@ -306,7 +306,7 @@
* pipe_barrier
*/
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_HURD)
typedef pthread_barrier_t pipe_barrier;
@@ -434,7 +434,7 @@
*/
typedef struct {
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
pthread_key_t key;
#elif defined(PIPE_SUBSYSTEM_WINDOWS_USER)
DWORD key;
@@ -449,7 +449,7 @@
static INLINE void
pipe_tsd_init(pipe_tsd *tsd)
{
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
if (pthread_key_create(tsd-key, NULL/*free*/) != 0) {
perror(pthread_key_create(): failed to allocate key for thread specific data);
exit(-1);
@@ -466,7 +466,7 @@
if (tsd-initMagic != (int) PIPE_TSD_INIT_MAGIC) {
pipe_tsd_init(tsd);
}
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
return pthread_getspecific(tsd-key);
#elif defined(PIPE_SUBSYSTEM_WINDOWS_USER)
assert(0);
@@ -483,7 +483,7 @@
if (tsd-initMagic != (int) PIPE_TSD_INIT_MAGIC) {
pipe_tsd_init(tsd);
}
-#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN)
+#if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS) || defined(PIPE_OS_APPLE) || defined(PIPE_OS_HAIKU) || defined(PIPE_OS_EMBEDDED) || defined(PIPE_OS_CYGWIN) || defined(PIPE_OS_HURD)
if (pthread_setspecific(tsd-key, value) != 0) {
perror
reassign 480051 to compiz
# Automatically generated email from bts, devscripts version 2.10.28 reassign 480051 compiz -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
