Re: Unit Tests failing when built with clang 12

I just did a git pull and UITest_calc_tests9 is passing on master. I'm bisecting the fix now. -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org

Re: Unit Tests failing when built with clang 12

Yes, I just verified that you'll need an --enable-optimized build to reproduce this failure. When I built with clang 12 + --enable-debug, all unit tests passed. -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html

Re: Missing _sysconfigdata_m_linux_i686-linux-gnu.py

Does the master branch build under the same configuration? I do not see this or any issue building i686 master under Arch Linux 32 or Cent OS 7. -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html ___

[Libreoffice-commits] core.git: sw/qa

sw/qa/extras/layout/layout.cxx | 20 ++-- 1 file changed, 14 insertions(+), 6 deletions(-) New commits: commit 85b29e0be7bbade080c2978aef598839a89cfd3e Author: slacka AuthorDate: Thu Dec 26 21:14:36 2019 -0500 Commit: Noel Grandin CommitDate: Fri Dec 27 13:26:50 2019

Re: [ABANDONED] Re: Build fail on Libgpg-error on Windows with gawk 5.0

We are still using libgpg-error v1.27 but the latest version on their download page is 1.36. Can we update this tarball so that new LODE installs can work again? https://gnupg.org/download/index.html#libgpg-error -- Sent from:

Re: [TDF infra] Announcing Gitiles VCS browser (gitweb replacement) and https:// anon git URIs

Thanks Guilhem! I like the look of gitiles, but they both need better integration. Could you add hyperlinks to referenced commit messages and to bugzilla? cgit: https://cgit.freedesktop.org/libreoffice/core/commit/?id=048b8e45813f6a19a4ff56e1d676fe9450325cd2 gitiles:

Re: Cppcheck: Reduction of False Positives: Manual Approach

The goal of my manual approach was to configure Cppcheck to minimize false positives. In doing so, I was forced to only scan the Linux code base, as only Linux has Linux system headers and does not have Window's or BSD's... So I am not surprised that some valid issues were not reported. There

Re: can't build libreoffice !

After I move the source folder around, I run a `./autogen.sh`to clean everything up so commands like 'make clean' work again. -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html ___ LibreOffice mailing list

Re: make check always fail

Rahul, Please take a few minutes to include the relevant details before emailing several hundred recipients a list like this. Telling us your 'make check' fails without a single detail on the error message or your build environment just wastes everyone's time including yours. Here is what a

Re: Windows Debug builds failure: 'DbgGUIInitSolarMutexCheck': identifier not found

Sorry I was in the wrong folder: this is the correct one Microsoft (R) C/C++ Optimizing Compiler Version 19.14.26433 for x64 Copyright (C) Microsoft Corporation. All rights reserved. -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html

Re: Windows Debug builds failure: 'DbgGUIInitSolarMutexCheck': identifier not found

Here is my build environment: $ cat autogen.input --enable-debug $ cl /? Microsoft (R) C/C++ Optimizing Compiler Version 17.00.60610.1 for x86 Do you need anything else? I'll try reverting Noel's patches to check if that is the cause or not. -- Sent from:

Re: Windows Debug builds failure: 'DbgGUIInitSolarMutexCheck': identifier not found

Here is my build environment: $ cat autogen.input --enable-debug $ cl /? Microsoft (R) C/C++ Optimizing Compiler Version 17.00.60610.1 for x86 Do you need anything else? -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html

Re: Segmentation fault when compiling

This is not an issue specific to his environment. I'm also seeing the same seg fault on a clean Windows 32-bit debug build. $ cat autogen.input --enable-debug After a $ make clean && git pull && ./autogen.sh && make check I'm seeing: [build CUT] vcl_lifecycle [build SLC] drawinglayer [build

Re: Compiler baselines

devtoolset-7 for i386 is over here: https://copr-be.cloud.fedoraproject.org/results/mlampe/devtoolset-7/epel-6-i386/00756569-devtoolset-7/ Is that what you're looking for? -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html

Re: Compiler baselines

While it's true there is no 32-bit RHEL 7, CentOS 7 has had a 32-bit release since 2015. You can download it from here: http://mirror.centos.org/altarch/7/isos/i386/ More details can be found here: https://wiki.centos.org/SpecialInterestGroup/AltArch/i386 -- Sent from:

Re: Question, please

As of a few weeks ago, LODE was working. You just have to be careful to follow every step carefully. You might find some helpful developers on IRC or you can try to fix it yourself. Assuming you are in the lode/dev/core folder, if that make command is failing, you likely missed one of the

Re: CppunitTest_sc_tabviewobj failing on Windows MSVC 2017

Tested on another totally different system with MSVC 2015. Same failure as reported in my original post with ScTabViewObj::testGetReferredCells. I would be surprised if this works on any windows system. Should this be disabled for windows or does someone want to take a look at it? -- Sent

Re: CppunitTest_sc_tabviewobj failing on Windows MSVC 2017

Jens, Could you be a bit more specific? What debug info would you like me to gather? -Luke -- Sent from: http://document-foundation-mail-archive.969070.n3.nabble.com/Dev-f1639786.html ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org

Re: CppunitTest_sc_tabviewobj failing on Windows MSVC 2017

Sorry for the noise. We can talk about the OpenCL issue caused by the same commit in another thread. Re: the original issue, when I only run: $ make CppunitTest_sc_tabviewobj On my Broadwell MSVC 2015 system, it fails with the same error as I reported originally seen MSVC 2017 Sandybridge

Re: Subsequent Test PythonTest_dbaccess_python stuck in Infinite Loop

I can confirm that the build failure was caused by https://cgit.freedesktop.org/libreoffice/core/commit/?id=1c5c9b84 Refactor: migration warn dlg without modifying API and that Mike Kaganski's commit does resolve the issue. -- Sent from:

Re: PythonTest_pyuno_pytests_insertremovecells Unit Test Failing on Windows

With https://cgit.freedesktop.org/libreoffice/core/commit/?id=e21b2a61d64b697413c0fc9013746b9d358da4c6 and https://cgit.freedesktop.org/libreoffice/core/commit/?id=04f4702eb896a330bc3b10443b7303a1d07d7a11 Windows 'make check' builds now succeed again. Thanks Stephan! -- Sent from:

Re: vclplug_gtk - vclplug_gtk3

Assuming you are a developer and are trying to fix this issue yourself, I suggest you start by filing a bug report. Be sure to include your bug doc and system configuration so that others can reproduce this issue. --

Re: UITest_writerperfect_epubexport hang (only sometimes)

I experienced this hang today in UITest_writerperfect_epubexport on a clean build. A subsequent clean build on the same revision succeeded. This appears to be a intermittent issue. -- Sent from: http://nabble.documentfoundation.org/Dev-f1639786.html

Re: Unit test PythonTest_solenv_python failing on Linux on est_gbuildtojson

If you are getting this failure, you have to manually $ rm xmlsecurity/qa/unit/signing/data/S.gpg-agent Neither a "make clean" nor a "git clean -fdx" will remove this file. After that plus Thorsten's fix, "make check" now successfulness runs on Linux. Thanks Thorsten -- Sent from:

Re: Unit test PythonTest_solenv_python failing on Linux on est_gbuildtojson

I'm still getting the same failure and same "[Errno 6] No such device or address: '/core/xmlsecurity/qa/unit/signing/data/S.gpg-agent'")] Was this a blind fix? If so, you need anything else from me? -- Sent from: http://nabble.documentfoundation.org/Dev-f1639786.html

Re: CppunitTest_dbaccess_hsqldb_test / CppunitTest_dbaccess_RowSetClones Failing after System Update

Thank you Jan-Marek Glogowski. The workaround: kernel parameter stack_guard_gap=1 resolved the issue and the build successfully completed. This issue was also reported in our bug tracker here: https://bugs.documentfoundation.org/show_bug.cgi?id=108619 -- View this message in context:

Re: Link error: undefined reference to `__mulodi4' with Clang 32-bit

I build with x86 because my home laptop is a Dell Precision M90 Core Duo T2700 which is a 32-bit CPU. This machine has a gorgeous 16:10 display which is hard to find these days. The other machine is just a 32-bit image on a VM that I spool up to verify bugs. It does have the added benefit of less

Re: Link error: undefined reference to `__mulodi4' with Clang 32-bit

> Whatever the trouble with your toolchain If you have access to a Fedora x86 box with Clang, I'm sure you'd see this too as I'm seeing this on both my 32-bit Fedora box with Clang 5.0 master and 32-bit Ubuntu box with clang 3.9. > tweak the #elif so that you use the #else fallback

Re: Clang 5.0 32-bit builds Failing on CppunitTest_vcl_pdfexport

> SIGABRT failure of CppunitTest_vcl_filters_test only happen on one of those "2 different machines(Ubuntu/Fedora)" or on both? To Answer your other 2 Questions. It's failing on my Fedora box with Clang 5.0 master. It's failing on my Ubuntu box with clang 3.9 release. Both are x86. > * Which

Re: Build error: 'GetMap' is a protected member of 'SwAccessibleContext'

That commit fixed the error. I just finished a successful build with Clang. Thanks Markus! -- View this message in context: http://nabble.documentfoundation.org/Build-error-GetMap-is-a-protected-member-of-SwAccessibleContext-tp4213098p4213122.html Sent from the Dev mailing list archive at

Re: Clang 5.0 32-bit builds Failing on CppunitTest_vcl_pdfexport

I think my initial bisect was wrong because I didn't do a "make clean" between all the builds. Reverting https://cgit.freedesktop.org/libreoffice/core/commit/?id=6889fa826eef6bd1074d77507818e71dfe8ba152 Caused a failing build to pass, so it looks like Michael Stahl's hunch was correct. $ cat

Re: Clang 5.0 32-bit builds Failing on CppunitTest_vcl_pdfexport

Here is the backtrace: https://pastebin.com/vhYRQDSG This was with clang version 4.0.0 (trunk 287326) On Ubuntu 16.04.2 LTS 3 4.4.0-64-generic #85-Ubuntu SMP Mon Feb 20 11:49:39 UTC 2017 i686 i686 i686 I'm not sure if it was timing, ccache. or what but whatever is causing this crash started

Re: Possible bug in libreoffice online

Aditya Dewan, This list is more for coding-level discussion, not bug reports. To file a bug report, please follow the instructions on our wiki: https://wiki.documentfoundation.org/QA/BugReport -- View this message in context:

Re: Post-Build Limited Functionality

'libreoffice-common is needed for your distro's version of libreoffice. And your symptoms sound like a corrupt user profile. Are you running your distro's, soffice? or the one you just built, ./instdir/program/soffice ? If you are running the version you just built, i'd try making a clean

Re: MSVC 2015 Error C2664

sberg wrote > fixed with > https://cgit.freedesktop.org/libreoffice/core/commit/?id=a4d091485c1cc59a35fb3df17b23a9e10cf9c311; > > "Blind fix for MSVC 2015 error C2664" Yes that did the trick, and with Markus's commit:

Re: MSVC 2015 builds Failing After 2017 Support added

Thanks for the work-around, but that doesn't really answer my question. To clarify, before support for MSVC 2017 was added, I did not need the Win 10 SDK. Even after, I do not need the Win 10 SDK to build with MSVC 2013. Only MSVC 2015 has this new requirement. Was this change intentional? If

Re: minutes of ESC call ...

VS 2015 has much better support for Support For C++11/14 Features https://msdn.microsoft.com/en-us/library/hh567368.aspx When VS 2013 is depreciated, we can probably remove some ugly IFDEFs. I really hope that hour search was a bad joke. Google "Download VS 2013" and the first link is the

Re: Does anyone use windows

That is not what the wiki says. It says "Visual Studio 2013 is the preferred compiler for new developers." It also goes on to say "Visual Studio 2015 is known to work on current master." If you are a new developer, I suggest you follow the wiki's advice. The wiki contains everything you need to

Re: system-wide location for dictionary files and dictionary file name format

To answer your first and third question, we use hunspell not myspell. For the spellcheck function to work, LibreOffice requires both the main hunspell package along with a language dictionary (hunspell-en for English, hunspell-de for German, etc). For a package maintainer, your main concern should

Re: helpful stuff

How are spammers getting malicious messages like this on the list. It's showing up in both our Archive and Nabble. There seems to be a moderations system that blocks non-subscribers. Do we need more mods or just stricter controls to make it harder to circumvent. If you need more mods, who should

Re: build breaks in embeddedobj/source/msole/olemisc.cxx

I saw that same exact failure earlier today. I did a $ make clean && git pull && ./autogen.sh && make And now had a successful build of commit e15fd1371567fd4b78dab37d843562ff676c52b8 I guess that build error was fixed by:

Re: export-validation in tests by default causing 32 bit Ubuntu to crash

This appears to be an OpenJDK 1.8 bug that was fixed in 1.8.0_112. At least with 32-bit Arch, 1.8 can now pass all the Unit Tests without the need for "--without-export-validation". Java version "1.7.0_111" : GOOD OpenJDK Runtime Environment (build 1.8.0_91-8u91-b14-3ubuntu1~16.04.1-b14):

Re: Need help for building

clean everything up with $ git checkout master $ git clean -fdx $ git reset --hard origin/master $ git pull then make everything is sync'd with master $ git status If that doesn't do it, Try paring down your ./autogen.input. I'm having no issues with a simpler version on VS2013 -- View

Re: Warning and error during settingup lode in windows

Nikhil, What GPU and version of Windows do you have? Markus, I posted about the same testFinancialXirrFormula error here. I was only seeing this on my Win 10

Re: MSVC 2015's broken static initializer_list causing build failure

It's working now for me too. Thanks! -- View this message in context: http://nabble.documentfoundation.org/MSVC-2015-s-broken-static-initializer-list-causing-build-failure-tp4202928p4203154.html Sent from the Dev mailing list archive at Nabble.com.

Re: MSVC 2015's broken static initializer_list causing build failure

Per your recommendation, I upgrade to Microsoft Visual Studio Community 2015 Version 14.0.25431.01 Update 3 After the update, I'm still seeing the same error. configure: error: working support for static initializer_list needed -Luke -- View this message in context:

Re: MSVC 2015's broken static initializer_list causing build failure

Microsoft Visual Studio Community 2015 Version 14.0.25123.00 Update 2 Is it working for you? What version are you running? -- View this message in context: http://nabble.documentfoundation.org/MSVC-2015-s-broken-static-initializer-list-causing-build-failure-tp4202928p4202950.html Sent from

Re: OpenCL enabled by default after OS reinstall causing Unit Test Failure

If someone wants to dig deeper, I’d be glad to run special debug builds like I did for Ashod’s failing Unit tests. Or, should we just blacklist my driver? What info do you need for that? -Luke -- View this message in context:

Re: OpenCL enabled by default after OS reinstall causing Unit Test Failure

Niklas Johansson wrote > Is it the only OpenCL test that fails for you? Yes, Build finishes without error after I comment the testFinancialXirrFormula out.Afterwards when I open:core\sc\qa\unit\data\ods\opencl\financial\XIRR.odsThe actual results seem to match the expected results. -Luke --

Re: OpenCL enabled by default after OS reinstall causing Unit Test Failure

I attached the old HDD to an external USB and reconfigured the old Cygwin.bat to run from the new drive. With the old make 4.0 and all the old LODE tools, the weird " [0m [37;40m[ [0m [36;1;40mCUT [0m " on every line went away, but the build still failed on Test name:

Re: Budgeting & investing in features ...

Here are some ideas * Regular scan-build reports like we do with CppCheck[1] * Improve CppCheck (uncheck low priority/quality warnings, show new detected problem since last run, and [2]). * More variety in Jenkins bots like switching 1 of Linux to 32-bit or adding 64-bit MSVC 2015 * Improve

OpenCL enabled by default after OS reinstall causing Unit Test Failure

I just reinstalled windows 10, LODE, and VS2013. Now my new builds are failing OpenCL Unit tests. This is not a regression as I built the same version as I made on my old OS. After disabled the failing Unit Tests, I discovered that the new binaries enable OpenCL by default, where as in my old

Re: I'm new and I'm having build issues

> Do I need to delete the failed build before building again? Not usually. If you do get stuck trying a 'make clean' before asking for help never hurts. > Does it need to be connected to the Internet while building? (so that I > could disconnect the net while disabling my firewall to build) This

Re: Needs help

Use a text editor or IDE to edit the code. But first you need to pick an easy hack that look doable and interesting to you. https://wiki.documentfoundation.org/Development/EasyHacks -- View this message in context: http://nabble.documentfoundation.org/Needs-help-tp4201846p4201865.html Sent

RE: Build Problem with Lode

Jason Marshall wrote > From your experience, is it best to disable anti-virus entirely during the > build, or should it be configured to ignore certain programs associated > with the build? As you can see from the build notes

Re: Merging feature/commonsallayout branch

I just came back from a trip to China and found many machines there still running XP with old versions of MS Office. I used the opportunity to introduce several people there, including one IT manager, to LibreOffice. Had we already dropped support for XP, it would have been an embarrassing

Re: Fresh install of Win 10 + LODE + vs2013-ide-integration = Error 127

Hi Bjoern, Should this be working with a vanilla LODE install? With the master I'm getting: $ ../../opt/bin/make vs2013-ide-integration cd C:/cygwin/home/luke/lode/dev/core && (LC_MESSAGES=C C:/cygwin/home/luke/lode/opt/bin/make cmd="C:/cygwin/home/luke/lode/opt/bin/make -npf Makefile.gbuild

Re: Fresh install of Win 10 + LODE + vs2013-ide-integration = Error 127

Yes, I had a successful build with LODE's default autogen.input="--enable-debug". After it finished, I ran './instdir/program/soffice.exe' to make sure everything was actually working. It seemed fine. So I tried several runs of 'C:/cygwin/home/luke/lode/opt/bin/make vs2013-ide-integration'. I also

Re: export-validation in tests by default causing 32 bit Ubuntu to crash

The error "- failed to execute: sh /core/bin/officeotron.sh /tmp/lu16851plw1cl.tmp > /tmp/lu16851plw1cn.tmp" along with an apport crash report of openjdk 8, led me to believe this was a java issue. Here are a summary of my results: Ubuntu 16.04 64-bit - builds successfully Ubuntu 16.04 32-bit -

Re: Test File: sc/qa/unit/data/functions/fods/chiinv.fods: fails with Assertion

Eike Rathke-2 wrote > If you want to get a working build without disabling all checks then you > can use > https://gerrit.libreoffice.org/gitweb?p=core.git;a=commitdiff;h=b9a27d5856f60688456762bfcc29c38670009254 > > that disables the function tests for linux_x86 builds. > > However, if you'd

Re: Recommended build instructions ...

Is the "make vs2013-ide-integration" script working for anyone? With a clean Win 10 install + VS 2013 + LODE, I always get: make[1]: *** [cmd] Error 127 From, http://nabble.documentfoundation.org/Error-building-Visual-Studio-2013-project-file-for-LibreOffice-on-Windows-tp4140307p4140941.html

Re: Recommended build instructions ...

Speaking of wiki build instructions, the IDE instructions for Windows have not worked for almost a year now. Both the IDE instructions and Windows Instructions

Re: Test File: sc/qa/unit/data/functions/fods/chiinv.fods: fails with Assertion

This may be a regression, but I'm not seeing it with the 32-bit 5.1.4.2 official LibreOffice for Ubuntu 16.04. With minverse.fods, C10 changes to FALSE after pressing Shift+Ctrl+F9. With vdb.fods, C11 changes to FALSE after a force recalculate. Interestingly the minverse.fods test also fails with

Re: export-validation in tests by default causing 32 bit Ubuntu to crash

with the following autogen.input: --enable-ext-validator --with-export-validation --enable-symbols I get a crash in bfbd66f151002af6f65b9c062cc511797690e625 Test name: SdImportTest::testDocumentLayout equality assertion failed - Expected: 0 - Actual : 34304 - failed to execute: sh

Re: Recommended build instructions ...

Michael Meeks-5 wrote > * I was excited about Chocolatey ... eventually I gave up, and moved on to > manual installation. As did I. If you look at the history all of the Chocolatey info was recently added. It seems to cause more problems than it solves. Should we should move it off into it's own

Re: MSVC 2015 Error C2672: no matching overloaded function found

Thanks Stephan! I verified that both the 32 and 64-bit builds are working again under MSVC 2015 after your commit. -- View this message in context: http://nabble.documentfoundation.org/MSVC-2015-Error-C2672-no-matching-overloaded-function-found-tp4192415p4192564.html Sent from the Dev

Re: LibreOffice's New Area Dialog Box crashes when compiled with Clang

I sent the backtrace to Stephan directly, which he used to resolve the clang crash with https://cgit.freedesktop.org/libreoffice/core/commit/?id=e11931da72eabd98f9ea49ae93bd76039b283268 Why was this working with gcc and MSVC? Undefined behavior or a compiler bug? Thanks Stephan! -- View

Re: LibreOffice's New Area Dialog Box crashes when compiled with Clang

Here is the backtrace made with $ instdir/program/soffice --backtrace, and the steps listed in my first post. Is there anything else I can do? -- View this message in context:

Re: Another C2220: warning treated as error with MSVC 2015

Can we trust MSVC 2015 unreachable code analysis like we with this similar issue? The patch below fixes the build: diff --git a/comphelper/source/misc/threadpool.cxx

Re: New MSVC 2015 build failure

Yes, that fixed it. Thanks Michael! Now the only hold up with MSVC 2015 is the C4702: unreachable code Error. http://nabble.documentfoundation.org/C2220-warning-treated-as-error-64-bit-MSVC-2015-td4184053.html#a4189120 -- View this message in context:

Re: New C2220: warning treated as error (MSVC 2015 build failure)

After Michael Stahl's commit, https://cgit.freedesktop.org/libreoffice/core/commit/?id=59e4bd07211d8509dcb3e89f6645c0ba3085df23 MSVC 2015 can successfully build master again. Thanks everyone! -- View this message in context:

Re: Error building Visual Studio 2013 project file for LibreOffice on Windows

Ashod Nakashian-2 wrote > I haven't tried a different version of make as Juergen suggests, but I > confirm that the issue has to do with the fact that make doesn't generate > the invariant data that the python script excepts. Specifically, all the # > lines that contain the paths and filenames are

Re: vs2013 + Intel GPU = build failure on LO 4.4+

After https://cgit.freedesktop.org/libreoffice/core/commit/?id=40b0b9ab7703a165295b008f47df14d2ec076fb1 LibreOffice can now be successfully built with VS 2013 again. Thanks Mark! -- View this message in context:

Re: C2220: warning treated as error (64 bit MSVC 2015)

With sberg's and Noel's patches, I can now build the 64-bit version again. Nice work team! To answer my own question, the C4334 warning is new in VS 2015 Update 2. -- View this message in context:

Re: C2220: warning treated as error (64 bit MSVC 2015)

Even after: https://cgit.freedesktop.org/libreoffice/core/commit/?id=4a8caa37fe83c389f4fd43219e8ea0f21f03c6fe With the 64-bit build, I'm still getting the following error: [build CXX] vcl/source/filter/ixbm/xbmread.cxx [build CXX] vcl/source/filter/ixpm/xpmread.cxx