Age | Commit message (Collapse) | Author |
|
This commit reverts 3749d9af3745c0eaff7239e379578e4e2af89e9d
which removes the dependency on the external library cuckoo
Without using cuckoo the same file in tdf#130795 takes
real 0m4,892s
user 0m5,298s
sys 0m0,449s
With it, it takes
real 0m4,914s
user 0m5,276s
sys 0m0,444s
pretty much the same time
Change-Id: I4cc9000ac5bf26de22bb9835283ae8d5b3230196
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/138435
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
Change-Id: I2e6ec0a32b859cead16965284359115db0628f5f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137277
Tested-by: Jenkins
Reviewed-by: Tor Lillqvist <tml@collabora.com>
|
|
Loongarch is a new RISC ISA , which includes a 32-bit version and a 64-bit version, Here are some documents about Loongarch:
https://github.com/loongson/LoongArch-Documentation.
The loongarch64 bridges implementation refers to mips64 bridges, and the code related to abi and asm has been modified
Change-Id: I1d9cd3aadf63046c8cdecc9a64842567bfa1cecc
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137486
Tested-by: Stephan Bergmann <sbergman@redhat.com>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Disabled by default. It might improve sub-pixel glyph positioning a bit,
though it is only used inside VCL and the API boundary is still
sal_Int32, sound rounding to full pixels still happens at some point.
Change-Id: I702f9a521e5b556d198756d6cdbb455ebaffbab4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137664
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
ad5cbcf6ba0afdc1d8d7405c2641cce8de4a360b "try to autodetect flags needed to
build Clang plugins" had added both, in that order, first a check for a
hand-curated list of individual -lclangTooling -lclangDriver -lclangFrontend ...
libs, and only then a fallback check for the combined -lclang-cpp. It did not
give any rationale for that particular order. But checking for the hand-curated
list first turned out to have the following drawback: My build against recent
Clang 16 trunk succeeded the check for the hand-curated list but then failed
during the build because the hand-curated list missed to include a new
-lclangSupport dependency. And my naive fix
5078591de9a0e65ca560a4f1913e90dfe95f66bf "CLANGTOOLLIBS needs to include
-lclangSupport now" caused further issues:
For one, I had naively assumed that libclangSupport.a would be available on our
compiler plugins Clang 12.0.1 baseline, not noticing that the corresponding
Jenkins build
<https://ci.libreoffice.org/job/gerrit_linux_clang_dbgutil/119581/> only
succeeded with
> checking for clang libraries to use... -lclang-cpp -L/home/tdf/lode/opt_private/clang-llvmorg-12.0.1/lib -lLLVMWindowsManifest -lLLVMXRay -lLLVMLibDriver -lLLVMDlltoolDriver -lLLVMCoverage -lLLVMLineEditor -lLLVMX86Disassembler -lLLVMX86AsmParser -lLLVMX86CodeGen -lLLVMX86Desc -lLLVMX86Info -lLLVMOrcJIT -lLLVMMCJIT -lLLVMJITLink -lLLVMOrcTargetProcess -lLLVMOrcShared -lLLVMInterpreter -lLLVMExecutionEngine -lLLVMRuntimeDyld -lLLVMSymbolize -lLLVMDebugInfoPDB -lLLVMDebugInfoGSYM -lLLVMOption -lLLVMObjectYAML -lLLVMMCA -lLLVMMCDisassembler -lLLVMLTO -lLLVMPasses -lLLVMCFGuard -lLLVMCoroutines -lLLVMObjCARCOpts -lLLVMHelloNew -lLLVMipo -lLLVMVectorize -lLLVMLinker -lLLVMInstrumentation -lLLVMFrontendOpenMP -lLLVMFrontendOpenACC -lLLVMExtensions -lLLVMDWARFLinker -lLLVMGlobalISel -lLLVMMIRParser -lLLVMAsmPrinter -lLLVMDebugInfoDWARF -lLLVMSelectionDAG -lLLVMCodeGen -lLLVMIRReader -lLLVMAsmParser -lLLVMInterfaceStub -lLLVMFileCheck -lLLVMFuzzMutate -lLLVMTarget -lLLVMScalarOpts -lLLVMInstCombine -lLLVMAggressiveInstCombine -lLLVMTransformUtils -lLLVMBitWriter -lLLVMAnalysis -lLLVMProfileData -lLLVMObject -lLLVMTextAPI -lLLVMMCParser -lLLVMMC -lLLVMDebugInfoCodeView -lLLVMDebugInfoMSF -lLLVMBitReader -lLLVMCore -lLLVMRemarks -lLLVMBitstreamReader -lLLVMBinaryFormat -lLLVMTableGen -lLLVMSupport -lLLVMDemangle -lrt -ldl -lpthread -lm -lz -ltinfo -lxml2
by silently falling back to -lclang-cpp after the check for the hand-curated
list had apparently failed due to the missing libclangSupport.a. It turns out
libclangSupport.a only got added with
<https://github.com/llvm/llvm-project/commit/f26c41e8dd28d86030cd0f5a6e9c11036acea5d2>
"[RISCV] Moving RVV intrinsic type related util to clang/Support" towards
Clang 15.
And for another, as discussed in the mailing list thread starting at
<https://lists.freedesktop.org/archives/libreoffice/2022-August/089212.html>
"Re: Libreoffice error during autogen", there are apparently environments that
lack both libclangSupport.a and libclang-cpp.so, which then started to fail the
configure check completely.
Checking for the combined -lclang-cpp first avoids all of that (so this commit
effectively reverts 5078591de9a0e65ca560a4f1913e90dfe95f66bf "CLANGTOOLLIBS
needs to include -lclangSupport now" again). The clang-cpp library should
generally be available since Clang 9 (see
<https://github.com/llvm/llvm-project/commit/a80a3a2b239591f62422686dae9071df89a3b339>
"Document clang-cpp in the release notes for clang"), but it can presumably be
configured not to be built, and there are apparently environments lacking it
(see above). So keep checking for the (original, lacking the problematic
-lclangSupport) hand-curated list as a fallback, even though that list is known
to become outdated and dysfunctional.
Change-Id: I0edf23f5ca30b125e6438db0a42c849f1b4677d3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137789
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...presumably since
<https://github.com/llvm/llvm-project/commit/7a5cb15ea6facd82756adafae76d60f36a0b60fd>
"[RISCV] Lazily add RVV C intrinsics", otherwise causing
> [GEN] compilerplugins/clang/sharedvisitor/analyzer
> /usr/bin/ld: ~/llvm/inst/lib/libclangSema.a(SemaRISCVVectorLookup.cpp.o): in function `clang::CreateRISCVIntrinsicManager(clang::Sema&)':
> ~/github.com/llvm/llvm-project/clang/lib/Sema/SemaRISCVVectorLookup.cpp:195: undefined reference to `clang::RISCV::RVVIntrinsic::computeBuiltinTypes(llvm::ArrayRef<clang::RISCV::PrototypeDescriptor>, bool, bool, bool, unsigned int, bool, clang::RISCV::PolicyScheme, clang::RISCV::Policy)'
[...]
with recent Clang 16 trunk.
It would have been nice to extend the configure check to actually catch that,
but I didn't find any small code sample that would diagnose that:
* Reverting 79ca031fb2e8802aa4383362f83306fad6895729 "make the clang plugins
configure check faster" would cause the configure check to erroneously fail
due to "no member named 'tooling' in namespace 'clang'".
* And then replacing clang/AST/RecursiveASTVisitor.h with
clang/Tooling/Tooling.h would make the configure check succeed, and not
diagnose the above issue.
Change-Id: I6b2eb38daaa1c514fb61652e5bd522f91aa67059
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137755
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
as already handled as such in postprocess/CustomTarget_registry.mk
Change-Id: Idc7ead94db8f14e656c36db0438c916b0f2565e6
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137498
Tested-by: René Engelhard <rene@debian.org>
Reviewed-by: René Engelhard <rene@debian.org>
|
|
as that what it actually is and the compiler defines are already
-DMIPS/-DMIPS64 anyway (as are the PLATFORMID and the bridges'
name mips/mips64)
Change-Id: I274e7a3b0e140375a8e697c1790cbdb972251d29
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137489
Tested-by: René Engelhard <rene@debian.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: René Engelhard <rene@debian.org>
|
|
Change-Id: Id357ee8e27eb3e8b6f5f076350c43bb75e7d0683
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137348
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: Ie33eb8e256a38c5120fdd713b3fc34160a8728b7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137347
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
- Enabling some Cygwin checks also for WSL.
- Handling of Windows paths as needed for WSL.
- Reading of registry using wsl-lo-helper as WSL doesn't provide registry
in /proc the way Cygwin does.
Configure now passes for me (with Skia and Java disabled).
Change-Id: I325c4e6f9f825b3b6d0aa6cb350bafabc4011ce5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137346
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
It's not set in my WSL, but since it's going to be c:\program files (x86)
in the vast majority of cases, just hardcode a fallback.
Change-Id: I3bf41d6bae0e5bb36c53f0a4cf913a980a058fcf
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137345
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
It is needed to even find MSVC, so configure cannot easily build it.
Change-Id: Ie4e950cf01b9d8778cdc3e9a53718954b0c3c166
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137344
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
There's no wslsys in my WSL setup.
I also don't see why WSL should be at least version 2, they both
should(?) work and it is recommended to use version 1 with NTFS
(and version 2 also doesn't work e.g. inside VirtualBox).
Change-Id: I5b9440f65624f49e331d84235046e4dd2e31af4e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137343
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
of course it should only be added when the bluetooth part of the
sdremote is enabled, not the other way round...
fixes c78b49ec132eb5126445ebee7d259d3df7fcaa68
Change-Id: I0085c61a4bafece80be92c56928095c2726e11b4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137025
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
|
|
On 2022-07-04 I submitted 268f0f3f0d70902acaeb3f99e0d206ad66ffd67c
"run CppunitTest_chart2_xshape without ENABLE_CHART_TESTS" to
see if CppunitTest_chart2_xshape would fail for anyone.
Since then, no one has complained about it, so I suppose
the tests are stable nowadays, thus removing
--enable-chart-tests which does nothing since the mentioned
commit
Change-Id: I0407c0379ea45eee93bd6689c90d1759b48e2844
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137022
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
having one is one of the prerequisites of using TestFlight
Change-Id: I9e20eb99905071fade4179dfbe2da5b7e5dd1c24
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/136618
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
|
|
the network.server entitlement is only needed when the sdremote is
enabled, and the bluetooth one only when the bluetooth part is built.
Also warn about bluetooth not being built because of obsoleted/removed
API. It was "silently" dropped before (lost in the flood of configure
output).
Change-Id: I38d83aa1132307b5b4b5d6dceba8f9020ffb2373
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/136617
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
|
|
also reorder the logic to avoid duplicating the matches for default
value and manually provided one.
Change-Id: I4466cbeaf5abd7168f21e22cd910f63568c14e24
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/136224
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
|
|
Change-Id: I0a1e5289a36e39eeea00e0fd67fbcf1fe7c902f8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135772
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: René Engelhard <rene@debian.org>
Reviewed-by: René Engelhard <rene@debian.org>
|
|
Regression from be2252c8d3e5d892ec1e6b12a2b55c2026597cb9 where spaces
in --with-vendor end up being poorly quoted.
Opted to pass down --with-vendor to sub-configure unconditionally even
if empty, due to the need of word splitting generally being needed for
the sub_conf_opts variable.
Change-Id: Id3740255bc8fbf98adf7464de7e2836fcde9f2fd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135826
Reviewed-by: Balazs Varga <balazs.varga.extern@allotropia.de>
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
|
|
This gives us full support for variant, optional, any and visit libraries
Change-Id: I9f1bff5d7c0e2d5acc8c8b92c9a269b00e317574
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135804
Tested-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>
Reviewed-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>
|
|
1. error - if disable JAVA is configured, set '--without-java',
otherwise it breaks python for the build part.
2. error - forward '--with-vendor' setting to sub config,
to avoid stupid windows system variable error.
Change-Id: Icebdfc5d6d2816fa53538d83d3ea206a6e99ddd3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135724
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
|
|
Change-Id: I3073330ed25676d5fbf9f2af008e42828466fc3e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135565
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
* Client code must replace uses of idlc and regmerge with uses of unoidl-write,
see the changes to odk/examples/ and ure/source/uretext/ in
40f2aee6584eafcf4cd1d95fcf1f775e5435440d "Provide unoidl-write also for the
SDK" for examples.
* The new types.rdb format is not compatible with LibreOffice < 4.1. Clients
generating extensions containing such files are advised to use appropriate
LibreOffice-minimal-version elements.
* For compatibility with old extensions, reading the legacy types.rdb format is
still supported.
* The SDK no longer ships an idl/ sub-directory containing the udkap and offapi
.idl files (as, unlike idlc, unoidl-write does not need them).
odk/config/cfgWin.js had to be adapted to look (somewhat arbitrarily) for an
examples/ sub-directory instead of idl/ when checking for "an sdk folder".
gb_UnoApi_package_idlfiles became unused and has been removed.
* The idlc and regmerge executables have been removed. Module idlc has been
removed except for idlc/test/parser/, which is also used by
CustomTarget_unoidl/unoidl-write_test, and which may eventually be moved into
module unoidl. Module external/ucpp and the corresponding configure options
have also been removed.
Change-Id: I42a0231699b863b5ebe2bee63bc32c8f79278cc1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/122363
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: I4c696b0942edc5afe7a3d2babe62e7b70463f543
|
|
with an eye to win_arm64 cross build failure
Change-Id: I6872859f288168352e3e70c1d559f4d318439958
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135479
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
This way at least e.g. std::vector::operator[] will still check
bounds even though things like iterator debugging will not work.
Change-Id: If3c550cfec68eee9a19050fc8e60fca07148b4a4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134946
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
commit 88d57cf241209ffec9eaed3e523942ab51af6db6
Date: Wed Sep 29 11:09:51 2021 +0200
qt6: Add a qt6 VCL plugin
mentioned:
> 1) At least in my self-compiled Qt versions,
> 'moc' (the meta-object compiler) is located
> in the 'libexec' subdirectory in 'QT6DIR', while the
> Qt 5 equivalent is located in the "bin" subdirectory
> of 'QT5DIR', so the configure.ac check uses the former.
Search for moc in the "libexec" subdirectory of
all paths in `lib_dir` before trying `$PATH`,
so e.g. the one provided as `/usr/lib/qt6/libexec/moc`
by package "qt6-base-dev-tools" in Debian testing is found
before a Qt 5 version in `$PATH`.
With this change in place, building on Debian testing
using the system Qt 6 works when building with autogen option
`--enable-qt6` but no `QT6DIR` set and `/usr/bin/moc` points
to the Qt 5 version of moc.
Change-Id: I0471c73315224ffab018b350aff8d797a234a905
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134814
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
|
|
Debian testing/unstable provides a `/usr/bin/qmake6`.
Change-Id: I3dba55fc1a2d5667d9fc16aecacf8db6119ced02
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134813
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
|
|
If available, use GNU parallel to run N make_installer.pl scripts in
parallel, to scale packaging LibreOffice up with the rest of gbuild.
* fallback if no GNU parallel found - run make_installer sequentially
as before
* push most of the make_installer.pl input param tweaks from gbuild
down into a shared call_installer.sh script
* call gnu parallel with generated number of
"templ:lang:prodname:ext:pkgfmt:strip-flag" tuples, one for each
package to build (empty templ for non-windows, to save on cmd line
length)
* such that we can run all those in parallel (taking into account
the build's PARALLELISM parameter)
* there's still the main package build running epm sequentially for
umpteen sub-packages from within _one_ make_installer.pl instance, but
that's much harder to parallelize from inside Perl (so we punt on
that here)
Change-Id: Ie7d3084ed60d003d587c5e64dc9fb1809b23e409
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133957
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
|
|
Change-Id: I9cd29a48a0b46bae9d265b502b250709ec8a3247
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134710
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Id9b6e1355147c3f68b9922db14f1b4904a05c686
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134650
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I8e4d439f0e03ca9e9e3088b451efd8d09bf37203
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134462
Tested-by: Jenkins
Reviewed-by: Tor Lillqvist <tml@collabora.com>
|
|
Change-Id: I19f54042613769993b92cfaf04de292cfbb72b5f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134459
Tested-by: Jenkins
Reviewed-by: Tor Lillqvist <tml@collabora.com>
|
|
No major changes.
See https://icu.unicode.org/download/71
Change-Id: I7929d175962ff13e4369005633a4135f17f97e8c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134404
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
|
|
Now includes a configure check for the std::to_string() crash.
Change-Id: I45c3b804b7a1ff5f6520f7bdf772497f72ddee47
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134338
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
File writerperfect/qa/unit/data/writer/libmwaw/pass/ScriptWriter_1
causes CppunitTest_writerperfect_writer to fail here with 0.3.20.
Change-Id: Id5e66b7c4531287357502b7b5ff7880af7186016
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134306
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: Ifde1b0dba9dbbd9c4bba614098afff530cc0fdd9
|
|
It triggers an assert in mold's --gdb-index handling because of missing
debug info. It sort of makes sense that --gdb-index should be tested
with debug info actually used.
Change-Id: I5adced29355deec9f0a6af390867f4ff2f2bd29c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134153
Reviewed-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
Tested-by: Jenkins
|
|
Change-Id: I62ebdd0c1339eb4a804e2ea277f043f68c35c601
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134082
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...which should always be available with our C++17 toolchains
Change-Id: If3b231f58c3c27a10d68ef4946b7ee2bb4f31a4e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134083
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
We require at least GCC 7.0.0 anyway, per README.md
Change-Id: Ib974690ef4e4a1861a43f33157fde9f4bb3ac8ac
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134080
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
It's currently too broken (https://reviews.llvm.org/D125184),
but they're working on fixing it, so hopefully somewhen later.
Change-Id: I3ca243a57d41bd9d8c4cdbdc4a6a3b5fdc49e4c6
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134023
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Change-Id: Id657d45a00c60e281891791dd7e2f178c7857a71
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133990
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
During the compilation with Visual Studio 2019 v16.5, the dragonbox
was failing the build, but it was OK when ugprading to the latest
version, 16.11.
It should be noted that according to the list of predfined macros in
Visual Studio, v16.10 and v16.11 use the same value for _MSC_VER,
which is 1929. Thus, the distinction between these 2 versions can not
be distinguished.
Predefined macros
https://docs.microsoft.com/en-us/cpp/preprocessor/predefined-macros
For not having the Visual Studio version > 16.10, a warning is
shown, and if the Visual Studio version is < 16.5, just like before,
an error is generated.
Change-Id: I6661ee5121b03ca43e1f7503b74191abcc8d6b40
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/132907
Tested-by: Jenkins
Reviewed-by: Hossein <hossein@libreoffice.org>
|
|
Nowadays the bin/find-unneeded-includes script is a better
solution for this problem
This essentially reverts the commits:
c716b3888e7e8150d1c1053ee6550afb56438b1f
096c7e889f3b9bd42a81fb0216e2a68fb27159fc
40e7eecb7efeeb9af59206d7a9c82ac55adf5279
Change-Id: I404a4b34176efaacf961605559af9de6d6cba10a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133540
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
|
|
...to harmonize it with the HAVE_LIBCPP configuration variable recently
introduced in 6014f363a08f2edd6086cf30b3bbb913f1c621c6 "improve support for
using libc++ on Linux and for libc++ debug mode"
Change-Id: I6b3fee2e4e564d75caf1e60095d0b554153bce9e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133745
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
It seems to run slightly faster in --enable-dbgutil mode and also
builds a bit faster (at least with Clang). But libc++ on Mac
isn't built with debug mode support.
Change-Id: Idf5dba9c4a56aba1f4163aa518a78d34b6837149
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133664
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
NSS is very ABI compatible and part of current LSB 5.0 and so it should
just work to use the one from the system, i'm not aware of a Linux that
doesn't ship it.
There used to be some feature patch in external/nss but this was removed
years ago, now it just contains build fixes.
Problems can occur if a mixture of system and bundled NSS libraries is
loaded, if first an old version .so is loaded and then newer version .so
depends on it but can't find some required symbol.
See for example:
https://www.mail-archive.com/libreoffice@lists.freedesktop.org/msg287043.html
Another issue is that e.g. Fedora ships libnsspem.so in the nss-pem
package, which may cause trouble in ASAN builds, where the internal NSS
is built to call __asan_poison_memory_region but the system
libnss-pem.so doesn't call __asan_unpoison_memory_region so we get:
==2568944==ERROR: AddressSanitizer: use-after-poison on address 0x61d0019afab0
WRITE of size 192 at 0x61d0019afab0 thread T28 (utl::Moderator)
0 in __interceptor_memset.part.0 (instdir/program/soffice.bin+0x466460)
1 (/lib64/libnsspem.so+0x15f3d)
2 (/lib64/libnsspem.so+0x16185)
3 (/lib64/libnsspem.so+0x8a9b)
4 (/lib64/libnsspem.so+0xe13b)
5 in secmod_ModuleInit workdir/UnpackedTarball/nss/nss/lib/pk11wrap/pk11load.c:244:11
..
21 in curl_easy_perform workdir/UnpackedTarball/curl/lib/easy.c:715:10
Also, with system NSS, things like p11-kit integration become
responsibility of the OS.
Change-Id: I50caedd46914dd5d6905c5d32d44a599cd78119e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/120388
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
|