Age | Commit message (Collapse) | Author |
|
Regression from commit d3c6ac6d0f23 ("tdf#125922 rename kde5 to
kf5 + plasma5").
Change-Id: I47f2a3977094acc0c7b4ae0af28c3774eba2daca
Reviewed-on: https://gerrit.libreoffice.org/76078
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Just as the gtk3 plugin isn't named GNOME, rename kde5 to kf5, as
it is based on the KDE frameworks 5 libraries.
This also includes:
* a convenience alias to load the kf5 VCL plugin in case someone
requests the kde5 plugin.
* keep convenience kde5 configure switch, but warn about it
* rename detected desktop from kde5 to plasma5
Change-Id: I6764a05b81a5edbf284484c234fee2649aacf735
Reviewed-on: https://gerrit.libreoffice.org/75313
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: Iad48c663708dc9cda00d2a8534981f34c1c6f9d0
Reviewed-on: https://gerrit.libreoffice.org/75577
Tested-by: Jenkins
Reviewed-by: Andras Timar <andras.timar@collabora.com>
|
|
... it's already impossible with 6.2; and was only needed prior to
6.0, where Windows XP support was needed.
Change-Id: Ia462f0b6566ae35bd68545d2d34d2987ee7907b9
Reviewed-on: https://gerrit.libreoffice.org/75334
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: Ic6d27877d757ff1a379e278053630dd5f3d14807
Reviewed-on: https://gerrit.libreoffice.org/75307
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
With older releases the C++ 17 will not work and the configure would
error out with weird message.
Change-Id: Ife83669f0b55d3b013ca33f0b2a2709884309d5d
Reviewed-on: https://gerrit.libreoffice.org/75217
Tested-by: Jenkins
Reviewed-by: Tomáš Chvátal <tchvatal@suse.cz>
Tested-by: Tomáš Chvátal <tchvatal@suse.cz>
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
Thanks to the extensive debugging information by the reporter of
tdf#126272, I found the embedded Qt5 version tag:
$ objdump -T libvclplug_qt5lo.so | grep qt_version_tag
000000.. DO *UND* 000000.. Qt_5.9 qt_version_tag
So even if we try to stay Qt 5.6 compatible, this adds a dependeny
to the build Qt version. But fortunatly this symbol dependency can
be prevented by using -DQT_NO_VERSION_TAGGING, which this patch
does. We still can't be sure we don't use newer symbols, but this
way we won't prevent running with older Qt versions at all.
Change-Id: Ia85c14f888d9743645909f59867b7ad568dc6f42
Reviewed-on: https://gerrit.libreoffice.org/75280
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
x86intrin.h is only available for gcc/clang, specific headers are
available in all compilers (if they support that specific intrin.
at all)
Change-Id: Ic6dec052c150032f67ae76e0ca362c4bc77b55b0
|
|
Adds CPU intrinsics detection in configure pass for compile time
detection and "cpuid" runtime detection of which CPU instruction
sets are available on the user device.
Change-Id: I0ee4d0b22a7c51f72796d43e7383a31d03b437ad
Reviewed-on: https://gerrit.libreoffice.org/75175
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
My 5-year-old desktop can build 8 jobs on its own, so 10 is a rather
pointless default nowadays. Icecream will spawn only as many jobs
as the cluster can handle, so this should work fine even for small
machines/clusters.
Change-Id: Iccdf3adf51f55428e89e042d83979a1e1327659e
Reviewed-on: https://gerrit.libreoffice.org/75029
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
* Add CC_BASE and CXX_BASE to shorten output of compiler check
messages
* Sets the g++ check to be 7.0 as documented
* Apply win_short_path_for_make for more directories and files
* Get rid of the last few tabs
* Untangle some mixed AC_MSG_CHECKING output
* Don't test the unset $CLANGVER for Apple clang
Change-Id: I83baf31c10630d2a8d356b10544632c35717a84f
Reviewed-on: https://gerrit.libreoffice.org/74725
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
The bundle helps to generate QR code in LO.
Change-Id: Iaa9225a72d15806c929d30951cefd3f3fee8960e
Reviewed-on: https://gerrit.libreoffice.org/73302
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>
|
|
Change-Id: Id62e648bf0faa037ddd5f77f30f7008132f54e19
Reviewed-on: https://gerrit.libreoffice.org/74985
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
This allows to use 64-bit tools on x64 hosts for all modules,
both 32- and 64-bit, and so avoid memory problems, like in
https://ci.libreoffice.org/job/gerrit_windows/39174/console
LINK : the 32-bit linker (C:\PROGRA~2\MIB055~1\2017\COMMUN~1\VC\Tools\MSVC\1414~1.264\bin\HostX86\x86\link.exe) failed to do memory mapped file I/O on `out\zlib.lib' and is going to restart linking with a 64-bit linker for better throughput
LINK : restarting link with 64-bit linker `C:/cygwin/bin\link.exe'
where it then picks wrong linker, failing the build.
Also, it theoretically allows using 32-bit tools when building on
32-bit host, and build both 32- and 64-bit modules (no idea if it
is actually possible though).
Removed checks for tools locations from older VS versions.
Change-Id: I4798a1c66df580027243ca8c7b4d375148214984
Co-authored-by: Jan-Marek Glogowski <glogow@fbihome.de>
Reviewed-on: https://gerrit.libreoffice.org/74924
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
E.g. on Windows, I see three duplicated paths in the resulting PATH,
so this aims to make PATH tidy.
Change-Id: If553eefe278f442dbe44c73ca955e481b5cabd63
Reviewed-on: https://gerrit.libreoffice.org/74947
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I99497f44baf1b51082c7c34e5fcfa8eeacff94ed
Reviewed-on: https://gerrit.libreoffice.org/74933
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
... hence require $(file) support.
https://lists.freedesktop.org/archives/libreoffice/2019-June/083004.html
Change-Id: Ic7bdb3d47a78e4f620c2efd9ec58042378dbf2f3
Reviewed-on: https://gerrit.libreoffice.org/74730
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>
|
|
Change-Id: I21fbcdc947c9ce5d99b14c1aba8fb4e1fb2d9c00
Reviewed-on: https://gerrit.libreoffice.org/74680
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: I0fd132965af051b6a717349943fa6827567c65dd
Reviewed-on: https://gerrit.libreoffice.org/74557
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
No need to reinvent the wheel and hardcode it in configure.ac.
Change-Id: Idb08ea0e5ce228bb0758dbdf023f3aee44da76eb
Reviewed-on: https://gerrit.libreoffice.org/74247
Reviewed-by: Andras Timar <andras.timar@collabora.com>
Tested-by: Andras Timar <andras.timar@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/74442
Tested-by: Jenkins
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
|
|
Uses the same naming then the gtk3 backend, i.e. "libreoffice"
for the instance name and different class names for the module
windows, like "libreoffice-writer".
These names are referenced in the desktop files a StartupWMClass
and for example used to pin an app to the task bar.
Change-Id: Ic9b8890536f6413ab59d2e0da866e2280ab3181a
Reviewed-on: https://gerrit.libreoffice.org/74014
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
There will presumably be more X11 specific code later to implement some
"fix" for tdf#119202. Moving the screensafer inhibitor is rather
uncontroversial in comparion, so start with it and use it to carry the
matching configure.ac changes.
A little "nightmare" are all the clashing X11 / Qt type undefs, but I
couldn't find a better solution while the inhibitor continues to
include the X11 headers in it's header.
Change-Id: I55c89c76726d30a890178488484e954207267e89
Reviewed-on: https://gerrit.libreoffice.org/74015
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Build spsupp for both x64 and x86, regardless of target platform.
This allows to install the ActiveX component to be used by both
64-bit and 32-bit applications on 64-bit systems (especially IE,
which runs both 64-bit and 32-bit processes simultaneously at
least on Win10), no matter which LO (32/64) was installed.
Move the DLLs from activex feature to ooo, to copy unconditionally.
Registration of LO-specific component will be also unconditional;
registration of replacement of MSO component will need own feature.
This doesn't yet register the component in system: TODO later.
Change-Id: Iccf5e73dfae306cb777f844d40611e23c4520a13
Reviewed-on: https://gerrit.libreoffice.org/71925
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
While the VCL plugins are dynamically loaded and therefore just
load their depending toolkit libraries, the GStreamer avmedia
backend now links against Qt and GTK+. The GStreamer API itself
is toolkit agnostic and the toolkit setup just uses a single
GStreamer symbol to create the specific video sink.
So the toolkit binding can simply be moved into the VCL plugin.
At the point of the GStreamer toolkit setup call the GStreamer
library is loaded by avmediagst, so the dlsym lookup should
never fail.
I also dropped the special GtkWidget handling. Using g_object_get
will increase the refcount of the widget. A g_object_unref after
adding it to the container seems to destroy it correctly.
Change-Id: I693947e441bceb4b09bc38920e308e39142d0a35
Reviewed-on: https://gerrit.libreoffice.org/73849
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
From its description, --disable-gui appears to be experimental, and it causes
failures like
> /usr/bin/ld: /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/workdir/CxxObject/vcl/headless/svpgdi.o: in function `dl_cairo_surface_set_device_scale(_cairo_surface*, double, double)':
> /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/vcl/headless/svpgdi.cxx:2181: undefined reference to `dlsym'
> /usr/bin/ld: /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/workdir/CxxObject/vcl/headless/svpgdi.o: in function `dl_cairo_surface_get_device_scale(_cairo_surface*, double*, double*)':
> /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/vcl/headless/svpgdi.cxx:2189: undefined reference to `dlsym'
> /usr/bin/ld: /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/workdir/CxxObject/vcl/unx/generic/glyphs/freetype_glyphcache.o: in function `(anonymous namespace)::dlFT_Done_MM_Var(FT_LibraryRec_*, FT_MM_Var_*)':
> /lo/home/tdf/lode/jenkins/workspace/lo_tb_random_config_linux/vcl/unx/generic/glyphs/freetype_glyphcache.cxx:191: undefined reference to `dlsym'
> collect2: error: ld returned 1 exit status
(<https://ci.libreoffice.org//job/lo_tb_random_config_linux/1912/>) that are
probably not worth fixing.
Change-Id: Ia36a8665967601e1c6ef5af3b67d17d25822e339
Reviewed-on: https://gerrit.libreoffice.org/73960
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: Iaaeb479698c79f2cea3fd2e3914c17d3a2692981
Reviewed-on: https://gerrit.libreoffice.org/73837
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Drop the extra '/' resulting in the include in path, like e.g.
'-isystem /usr/include//KF5/KCoreAddons'.
Change-Id: I6b1814b91c17926309b5a0567d4d1338a691e92e
Reviewed-on: https://gerrit.libreoffice.org/73894
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
|
|
Change-Id: I7048171501e32d29f9d5a58f0d5fc0b96b318a4e
Reviewed-on: https://gerrit.libreoffice.org/73777
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
...along with PYTHON, PYTHON_CFLAGS, and PYTHON_LIBS. (PYTHON_VERSION is
apparently set by AM_PATH_PYTHON. When explicitly passing in PYTHON_CFLAGS and
PYTHON_LIBS under --enable-python=system, AM_PATH_PYTHON does not get called,
but PYTHON_VERSION needs to be set, so needs to be passed in explicitly.)
Change-Id: Ia8c3877fffd8dcc179a56be04627c61a667d8cb0
Reviewed-on: https://gerrit.libreoffice.org/73662
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Drop all GStreamer 0.10 support according to
ESC decision of 2019-06-06.
GStreamer 0.10 is obsolete and no longer needed,
superseded by GStreamer 1.0 which is available in
baseline (RHEL 7 or CentOS 7) and all relevant distros.
Change-Id: Ic317eba04d2c17e141acc983f37fbfa4301c9f3f
Reviewed-on: https://gerrit.libreoffice.org/73619
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
|
|
Change-Id: Ibbc731f998b00a0ea1235a961290746a849f81cf
|
|
Change-Id: I8fffa4cef9628e6872c881cd0cbdfe85495fa324
|
|
Change-Id: I2369db6b76af67d71146903c081ce6427d45c211
|
|
Add the 'ENABLE_GTK3' and 'ENABLE_QT5' there
to allow using sth like
#include <config_vclplug.h>
#if ENABLE_GTK3
// your code here
#endif
for these as well.
Remove 'config_gtk3_kde5.h{,.in}', 'config_kde5.h{,.in}'
and 'config_qt5.h{,.in}'; they aren't included anywhere.
('QT5_HAVE_GLIB' is only used in Makefiles, so there's no
need to have it in a config header and 'KDE5_HAVE_GLIB' is
not used anywhere at all.)
Change-Id: Iea7906880f57dde782c50f2520e8d9b358f6bc82
Reviewed-on: https://gerrit.libreoffice.org/73204
Tested-by: Jenkins
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
|
|
By now, any outdated builders where this would fail should have been identified
and taken care of.
Change-Id: Ie94b76838cc3c8a6d3ee6e5a9e16ba32ed7c600a
Reviewed-on: https://gerrit.libreoffice.org/66998
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: I5378a7460c94118973e3ff7e4a03ba1769205890
|
|
Change-Id: I765ff8a80ac37b09e16bb16652f28534d1e575eb
Reviewed-on: https://gerrit.libreoffice.org/73161
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
This reverts commit efe28895498b03f1468a9dc9f510452f36affc2f.
It is that commit that is broken. The only thing that does not work is
that not all compilerplugin sources depend on config_clang.h, the rest is fine.
So instead of reverting something that in principle works (and even complaining
in the commit message about the original problem), just fix it
(will do in next commit).
Change-Id: Ic7766a97220d5b7ef1cd195320899564140fdf1c
Reviewed-on: https://gerrit.libreoffice.org/72975
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Apparently test does not short-circuit evaluation.
Change-Id: I0d9003735a43d0b5715b9d5fd9b5323efd6fdac1
Reviewed-on: https://gerrit.libreoffice.org/72999
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: I3ea246c84bb81e4eca9fe569edb957bed3c788ee
Reviewed-on: https://gerrit.libreoffice.org/72674
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
It doesn't correctly make a dependency on the .gch or anything
the .gch is made from, so there would be stale cache hits even
if anything included from the .gch changed. Disable CCACHE_DEPEND.
https://github.com/ccache/ccache/pull/427
Change-Id: I6929ff6ec1f1e50498033cdf95b56c94522166eb
Reviewed-on: https://gerrit.libreoffice.org/72704
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: Ib43e7349bce04e42b247af3c87e0128350e6c1c1
|
|
... hence it will print a pointless warning about some problem that
doesn't affect JDK 8 anyway.
Change-Id: Ic7669950dbb16815c85e4b8eb9d07d75c68d1b82
Reviewed-on: https://gerrit.libreoffice.org/72172
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>
|
|
This avoids the preprocessing step normally done by ccache (see
ccache docs), which speeds up the build.
This way there may be incorrect ccache hits if a system header
(found using -isystem) changes, but that should be rare and such
headers generally should be backwards compatible, so the chance
of an actual breakage should be very low.
The patch also adds --enable-ccache=nodepend to allow enabling ccache
without the depend mode.
https://lists.freedesktop.org/archives/libreoffice/2019-May/082684.html
Change-Id: Id69ca43988a016a917beb75927b39d8bde719aa7
Reviewed-on: https://gerrit.libreoffice.org/71629
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
So that it works out of the box automagically.
Change-Id: I0a6c93824233eef6ce25c44762ec40770ec5110f
Reviewed-on: https://gerrit.libreoffice.org/71628
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
There are now 4 levels of PCH support, the previous 'full' level
adding to PCH whatever the update_pch script finds useful,
and new levels 'system', which adds only external headers,
'base', which is 'system' and LO basic headers (sal, osl, rtl, vcl)
and 'normal', which is 'full' without headers from the module
built itself.
With Clang/GCC even 'system' still saves some time (10-15%) and since external
headers should rarely if even change, it should be without most
of the disadvantages of PCH. And even 'base' should be pretty easy
to use, as those headers should be rarely changed while developing,
thus avoiding the need for massive rebuilds. Using 'normal' or 'full'
does not seem to be worth it with Clang or GCC, but with MSVC that still
makes a difference, so keep(?) 'full' the default there.
The update_pch script unfortunately does not include as many system
headers as it could, since it includes only what is directly included
by the .cxx, but not what's included indirectly by .hxx files.
https://lists.freedesktop.org/archives/libreoffice/2019-May/082685.html
Change-Id: If83a07a1fc9b77d0134502b0d89348944f82806b
Reviewed-on: https://gerrit.libreoffice.org/71580
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: I8b080ad44f10e52fa5a726bd7cdf7dd8ff7e2dd4
Reviewed-on: https://gerrit.libreoffice.org/71913
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
No idea why this has been disabled, Clang can handle both plugins
and PCH at the same time.
Change-Id: I957b57f5f86b041c0437a121990b40bfa0095a0d
Reviewed-on: https://gerrit.libreoffice.org/71564
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
*Before this patch, Windows SDK came with .NET 4.5
and .NET 4.6.1 or later was ignored.
* the author of this configure.ac did not seem to understand that
WOW6432Node should be given to reg_get_value_64
instead of reg_get_value_32.
sometimes the code redundantly fetched the same entry WOW6432
(for 64 bit, wrongly passed to 32bit)
and the 32 bit one. I removed the duplicate.
Change-Id: Iff972d727907c480ef927d53fa1f53557fdeb1fd
Reviewed-on: https://gerrit.libreoffice.org/71268
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>
Tested-by: Michael Stahl <Michael.Stahl@cib.de>
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
...with latest Clang trunk towards Clang 9. All the no-longer necessary hacks
are made conditional on new NEED_CLANG_LINUX_UBSAN_RTTI_VISIBILITY, which is
still set for UBSan builds with older Clang on Linux (but which should
eventually be purged).
Various classes needed additional SAL_DLLPUBLIC_RTTI annotations, as building
with UBSan instrumentation can generate references to RTTI symbols from
additional places like outside a dynamic library that used to hide those symbols
by default (but used to not hide them for old UBSan builds thanks to the
-fvisibility-ms-compat hack).
The odr-violation suppressions in solenv/sanitizers/asan-suppressions (which is
not referenced from anywhere in the code base, but meant to be included in an
ASan/UBSan build's ASAN_OPTIONS env var) are also no longer needed when
NEED_CLANG_LINUX_UBSAN_RTTI_VISIBILITY is false.
Change-Id: I24ec3e388b0cbab50dbe2bf008d9569bff7bf25a
Reviewed-on: https://gerrit.libreoffice.org/70829
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|