Re: [Development] macOS CI node segfault
Just got a segmentation fault on OS X 10.9 machine, while compiling, which doesn't seem related to moc. http://testresults.qt.io/logs/qt/qtbase/9859c09bddaf87f49e73c1e8734decab497730ec/OSXOSX_10_08x86_64OSXOSX_10_08x86_64Clangqtci-osx-10.8-x86_64Release_NoFramework/85d6b000f945a84bc84a4f01f53ac65bc05cbe86/testrun_1476370856/testlog.txt.gz http://testresults.qt.io/coin/integration/qt/qtbase/tasks/1476370856 Regards, Alex. On 11 Oct 2016, at 09:23, Jedrzej Nowacki mailto:jedrzej.nowa...@qt.io>> wrote: Yes, it is possible, but that would be a manual work. From that perspective I would prefer to focus on automatic uploading core dumps. Fredrik volunteered to look into that (thanks!) Cheers, Jędrek From: Development mailto:development-bounces+jedrzej.nowacki=qt...@qt-project.org>> on behalf of Sérgio Martins mailto:iamser...@gmail.com>> Sent: Monday, October 10, 2016 4:25 PM To: Thiago Macieira Cc: development@qt-project.org<mailto:development@qt-project.org> Subject: Re: [Development] macOS CI node segfault On Sat, Oct 8, 2016 at 9:52 PM, Thiago Macieira mailto:thiago.macie...@intel.com>> wrote: That said, we've known moc has been crashing on Mac for a couple of months and we have no idea what's causing it. I recently hit a moc crash when compiling KHTML, then replaced moc with a script that called "valgrind moc $@", gave the report to olivier and he fixed it immediately. Is it possible to trigger a manual build on macOS node and try until it crashes ? Regards, Sérgio Martins ___ Development mailing list Development@qt-project.org<mailto:Development@qt-project.org> http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
Yes, it is possible, but that would be a manual work. From that perspective I would prefer to focus on automatic uploading core dumps. Fredrik volunteered to look into that (thanks!) Cheers, Jędrek From: Development on behalf of Sérgio Martins Sent: Monday, October 10, 2016 4:25 PM To: Thiago Macieira Cc: development@qt-project.org Subject: Re: [Development] macOS CI node segfault On Sat, Oct 8, 2016 at 9:52 PM, Thiago Macieira wrote: > That said, we've known moc has been crashing on Mac for a couple of months and > we have no idea what's causing it. I recently hit a moc crash when compiling KHTML, then replaced moc with a script that called "valgrind moc $@", gave the report to olivier and he fixed it immediately. Is it possible to trigger a manual build on macOS node and try until it crashes ? Regards, Sérgio Martins ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
Bug report for it doesn't exist it is something hanging in back of our heads for months, but feel free to fill the form. Cheers, Jędrek From: Sean Harmer Sent: Monday, October 10, 2016 9:11 PM To: Jedrzej Nowacki; development@qt-project.org Subject: Re: [Development] macOS CI node segfault On 10/10/2016 09:14, Jedrzej Nowacki wrote: > As Thiago said. Moc is crashing on OSX for an unknown reason. It is > unlikely to be RAM problem, because we would see much wider > distribution of crashes (related bug report: > https://bugreports.qt.io/browse/QTQAINFRA-990). Argh, sorry I indeed misread the segfault error. It is moc, not the compilation of moc's output as I thought. > Regarding overwriting logs, it is Coin bug. For some reason we naively > believed that builds are deterministic and there is only one build log > 100 % reproducible... Well something to fix. Is there a JIRA for this or should I file one? Cheers, Sean > > > Cheers, > > Jędrek > > > *From:* Development > on behalf of > Sean Harmer > *Sent:* Sunday, October 9, 2016 9:51:47 AM > *To:* development@qt-project.org > *Subject:* Re: [Development] macOS CI node segfault > > Hi Thiago, > > On 08/10/2016 21:52, Thiago Macieira wrote: >> Em sábado, 8 de outubro de 2016, às 14:44:59 CEST, Sean Harmer escreveu: >>> Hi, >>> >>> Just seen another segfault on one of the mac mini CI nodes: >>> >>> http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498 >>> b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRel >>> ease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz >> >> Where? That log file does not contain the word "segfault", "error" or the >> phrase "segmentation fault". > > Something very wrong is going on then. Look at the failed run at Oct 8 > 2:38 PM from https://codereview.qt-project.org/#/c/168122/ which is > where I copied the link from. The report on gerrit says segmentation > fault, as did the report when I checked it at the time, yet now indeed > the report has no mention of a segfault at all. > > Sounds like something overwrote the compile report which is a little > concerning. > >> >> That said, we've known moc has been crashing on Mac for a couple of months >> and >> we have no idea what's causing it. > > This was a segfault in the compiler when building > .moc/release/moc_qrenderaspect.cpp rather than when running moc but I am > suspicious of some bad RAM or something on the hardware. > > Cheers, > > Sean > ___ > Development mailing list > Development@qt-project.org > http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
On 10/10/2016 09:14, Jedrzej Nowacki wrote: As Thiago said. Moc is crashing on OSX for an unknown reason. It is unlikely to be RAM problem, because we would see much wider distribution of crashes (related bug report: https://bugreports.qt.io/browse/QTQAINFRA-990). Argh, sorry I indeed misread the segfault error. It is moc, not the compilation of moc's output as I thought. Regarding overwriting logs, it is Coin bug. For some reason we naively believed that builds are deterministic and there is only one build log 100 % reproducible... Well something to fix. Is there a JIRA for this or should I file one? Cheers, Sean Cheers, Jędrek *From:* Development on behalf of Sean Harmer *Sent:* Sunday, October 9, 2016 9:51:47 AM *To:* development@qt-project.org *Subject:* Re: [Development] macOS CI node segfault Hi Thiago, On 08/10/2016 21:52, Thiago Macieira wrote: Em sábado, 8 de outubro de 2016, às 14:44:59 CEST, Sean Harmer escreveu: Hi, Just seen another segfault on one of the mac mini CI nodes: http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498 b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRel ease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz Where? That log file does not contain the word "segfault", "error" or the phrase "segmentation fault". Something very wrong is going on then. Look at the failed run at Oct 8 2:38 PM from https://codereview.qt-project.org/#/c/168122/ which is where I copied the link from. The report on gerrit says segmentation fault, as did the report when I checked it at the time, yet now indeed the report has no mention of a segfault at all. Sounds like something overwrote the compile report which is a little concerning. That said, we've known moc has been crashing on Mac for a couple of months and we have no idea what's causing it. This was a segfault in the compiler when building .moc/release/moc_qrenderaspect.cpp rather than when running moc but I am suspicious of some bad RAM or something on the hardware. Cheers, Sean ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
On 10/10/2016 15:25, Sérgio Martins wrote: On Sat, Oct 8, 2016 at 9:52 PM, Thiago Macieira wrote: That said, we've known moc has been crashing on Mac for a couple of months and we have no idea what's causing it. I recently hit a moc crash when compiling KHTML, then replaced moc with a script that called "valgrind moc $@", gave the report to olivier and he fixed it immediately. Is it possible to trigger a manual build on macOS node and try until it crashes ? I suspect hw here as it was the compiler, not moc, which crashed in this case. But that's a good idea to try in general in case there is a moc specific issue too. Cheers, Sean Regards, Sérgio Martins ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development -- Dr Sean Harmer | sean.har...@kdab.com | Managing Director UK KDAB (UK) Ltd, a KDAB Group company Tel. +44 (0)1625 809908; Sweden (HQ) +46-563-540090 Mobile: +44 (0)7545 140604 KDAB - Qt Experts ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
On Sat, Oct 8, 2016 at 9:52 PM, Thiago Macieira wrote: > That said, we've known moc has been crashing on Mac for a couple of months and > we have no idea what's causing it. I recently hit a moc crash when compiling KHTML, then replaced moc with a script that called "valgrind moc $@", gave the report to olivier and he fixed it immediately. Is it possible to trigger a manual build on macOS node and try until it crashes ? Regards, Sérgio Martins ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
As Thiago said. Moc is crashing on OSX for an unknown reason. It is unlikely to be RAM problem, because we would see much wider distribution of crashes (related bug report: https://bugreports.qt.io/browse/QTQAINFRA-990). Regarding overwriting logs, it is Coin bug. For some reason we naively believed that builds are deterministic and there is only one build log 100 % reproducible... Well something to fix. Cheers, J?drek From: Development on behalf of Sean Harmer Sent: Sunday, October 9, 2016 9:51:47 AM To: development@qt-project.org Subject: Re: [Development] macOS CI node segfault Hi Thiago, On 08/10/2016 21:52, Thiago Macieira wrote: > Em s?bado, 8 de outubro de 2016, ?s 14:44:59 CEST, Sean Harmer escreveu: >> Hi, >> >> Just seen another segfault on one of the mac mini CI nodes: >> >> http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498 >> b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRel >> ease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz > > Where? That log file does not contain the word "segfault", "error" or the > phrase "segmentation fault". Something very wrong is going on then. Look at the failed run at Oct 8 2:38 PM from https://codereview.qt-project.org/#/c/168122/ which is where I copied the link from. The report on gerrit says segmentation fault, as did the report when I checked it at the time, yet now indeed the report has no mention of a segfault at all. Sounds like something overwrote the compile report which is a little concerning. > > That said, we've known moc has been crashing on Mac for a couple of months and > we have no idea what's causing it. This was a segfault in the compiler when building .moc/release/moc_qrenderaspect.cpp rather than when running moc but I am suspicious of some bad RAM or something on the hardware. Cheers, Sean ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
Hi Thiago, On 08/10/2016 21:52, Thiago Macieira wrote: Em sábado, 8 de outubro de 2016, às 14:44:59 CEST, Sean Harmer escreveu: Hi, Just seen another segfault on one of the mac mini CI nodes: http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498 b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRel ease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz Where? That log file does not contain the word "segfault", "error" or the phrase "segmentation fault". Something very wrong is going on then. Look at the failed run at Oct 8 2:38 PM from https://codereview.qt-project.org/#/c/168122/ which is where I copied the link from. The report on gerrit says segmentation fault, as did the report when I checked it at the time, yet now indeed the report has no mention of a segfault at all. Sounds like something overwrote the compile report which is a little concerning. That said, we've known moc has been crashing on Mac for a couple of months and we have no idea what's causing it. This was a segfault in the compiler when building .moc/release/moc_qrenderaspect.cpp rather than when running moc but I am suspicious of some bad RAM or something on the hardware. Cheers, Sean ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
Re: [Development] macOS CI node segfault
Em sábado, 8 de outubro de 2016, às 14:44:59 CEST, Sean Harmer escreveu: > Hi, > > Just seen another segfault on one of the mac mini CI nodes: > > http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498 > b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRel > ease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz Where? That log file does not contain the word "segfault", "error" or the phrase "segmentation fault". That said, we've known moc has been crashing on Mac for a couple of months and we have no idea what's causing it. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development
[Development] macOS CI node segfault
Hi, Just seen another segfault on one of the mac mini CI nodes: http://testresults.qt.io/logs/qt/qt3d/8f273f12204fb46ea508393ea9ad3dcd41a498b4/OSXOSX_10_10x86_64OSXOSX_10_10x86_64Clangqtci-osx-10.10-x86_64DebugAndRelease_Release/3c7a8e84853b9a3cd649d77444b6f1a885819671/buildlog.txt.gz Cheers, Sean ___ Development mailing list Development@qt-project.org http://lists.qt-project.org/mailman/listinfo/development