Created attachment 895927 [details]
build.log
Created attachment 895928 [details]
environment // profile default/linux/amd64/17.1/desktop/plasma gcc-13.2.1_p20240210 O3 pipe march=native(verbose) lto graphite ggdb3
iteration 1: IT HAPPENS when just -O0 -pipe -ggdb3 -frecord-gcc-switches Interesting, I guess, but you're using old Qt 5.15.13 (not in tree anymore) and old libktorrent. Update your system and try to reproduce it with libktorrent-24.05.1. Created attachment 897435 [details]
with updated system, in `zip -9` : build.log , testlog , environment, emerge --info, emerge -pqv libktorrent
I can reproduce with updated system on version 23.08.5
Let me check with new version.
(In reply to Andreas Sturmlechner from comment #4) > Interesting, I guess, but you're using old Qt 5.15.13 (not in tree anymore) > and old libktorrent. Update your system and try to reproduce it with > libktorrent-24.05.1. Sorry, do you mean to switch to plasma 6 ? (In reply to Arniii from comment #6) > (In reply to Andreas Sturmlechner from comment #4) > > Interesting, I guess, but you're using old Qt 5.15.13 (not in tree anymore) > > and old libktorrent. Update your system and try to reproduce it with > > libktorrent-24.05.1. > > Sorry, do you mean to switch to plasma 6 ? kde-apps/libktorrent does not depend on Plasma, it can be tested independently. (In reply to Arniii from comment #3) > iteration 1: IT HAPPENS when just -O0 -pipe -ggdb3 -frecord-gcc-switches Not reproducible here even with 23.08.5. Are you sure about your options? Latest logs still contain -flto. Please understand that options like LTO may affect libraries hence testing in a clean chroot and doing some work to figure out precisely what causes something is welcome. I am happy to look into exotic issues *once they've been reduced into something reproducible*, but I can't spend time on chasing dragons. In this particular case, it might well depend on the filesystem PORTAGE_TMPFS resides on. But my advice above applies in general. *TMPDIR. Created attachment 897447 [details] with updated system and only -O0 -pipe -ggdb3 -frecord-gcc-switches, in `zip -9` : build.log , testlog , environment, emerge --info, emerge -pqv libktorrent (In reply to Andreas Sturmlechner from comment #8) > (In reply to Arniii from comment #3) > > iteration 1: IT HAPPENS when just -O0 -pipe -ggdb3 -frecord-gcc-switches > Not reproducible here even with 23.08.5. Are you sure about your options? > Latest logs still contain -flto. Here's logs with such (In reply to Sam James from comment #9) > Please understand that options like LTO may affect libraries hence testing > in a clean chroot and doing some work to figure out precisely what causes > something is welcome. > > I am happy to look into exotic issues *once they've been reduced into > something reproducible*, but I can't spend time on chasing dragons. I'll try reproduce in clean chroot. (In reply to Sam James from comment #10) > In this particular case, it might well depend on the filesystem > PORTAGE_TMPFS resides on. But my advice above applies in general. After some researching, I've found you were right. It depends on filesystem of /tmp . ext4 - success `PARTUUID=c09c8e8a-0fca-4ee9-b72a-3bebbd3625eb /tmp ext4 defaults,noatime 0 1` tmpfs - success `tmpfs /tmp tmpfs rw,nosuid,nodev,size=4G,mode=1777 0 0` btrfs - success `PARTUUID=c774ac8a-3ccf-441f-b950-238696051d23 /tmp btrfs defaults,noatime 0 1` xfs - fail . Also, I've found next logs that can be produced by /var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5_build/bin/preallocationtest -v2 -o log.txt,txt : from ./preallocation.log : ``` Thu Jul 11 20:32:15 2024: Tot Size : 35858964 Thu Jul 11 20:32:15 2024: Num Chunks : 137 Thu Jul 11 20:32:15 2024: Chunk Size : 262144 Thu Jul 11 20:32:15 2024: Last Size : 207380 Thu Jul 11 20:32:15 2024: Created /tmp/preallocationtest-QKlEmptorrent Thu Jul 11 20:32:16 2024: Tot Size : 10825048 Thu Jul 11 20:32:16 2024: Num Chunks : 42 Thu Jul 11 20:32:16 2024: Chunk Size : 262144 Thu Jul 11 20:32:16 2024: Last Size : 77144 Thu Jul 11 20:32:16 2024: Created /tmp/preallocationtest-Nuvsdztorrent Thu Jul 11 20:32:16 2024: Preallocating file /tmp/preallocationtest-QKlEmpdata/movies/aaa.avi (11438415 bytes) Thu Jul 11 20:32:16 2024: file_size = 0 Thu Jul 11 20:32:16 2024: Preallocating file /tmp/preallocationtest-QKlEmpdata/movies/bbb.avi (9124816 bytes) Thu Jul 11 20:32:16 2024: file_size = 0 Thu Jul 11 20:32:16 2024: Preallocating file /tmp/preallocationtest-QKlEmpdata/movies/ccc.avi (15295733 bytes) Thu Jul 11 20:32:16 2024: file_size = 0 Thu Jul 11 20:32:16 2024: PreallocationThread has finished Thu Jul 11 20:32:16 2024: bw: 0, ts: 35858964 Thu Jul 11 20:32:16 2024: Preallocating file /tmp/preallocationtest-Nuvsdzdata/bla.avi (10825048 bytes) Thu Jul 11 20:32:16 2024: file_size = 0 Thu Jul 11 20:32:16 2024: PreallocationThread has finished Thu Jul 11 20:32:16 2024: bw: 0, ts: 10825048 ``` The code related that produces such is at /var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5/src/diskio/tests/preallocationtest.cpp Since you've isolated this to XFS, if it is reproducible 100% also with latest 24.05.2, you can file this bug upstream (on bugs.kde.org) instead. (In reply to Andreas Sturmlechner from comment #14) > Since you've isolated this to XFS, if it is reproducible 100% also with > latest 24.05.2, you can file this bug upstream (on bugs.kde.org) instead. Reported here : https://bugs.kde.org/show_bug.cgi?id=490231 reproducible with net-libs/libktorrent-24.05.2 reproducible with net-libs/libktorrent-24.08.1 |
Created attachment 895926 [details] /var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5_build/Testing/Temporary/LastTest.log Relevant part of /var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5_build/Testing/Temporary/LastTest.log : ``` ---------------------------------------------------------- ********* Start testing of PreallocationTest ********* Config: Using QtTest library 5.15.13, Qt 5.15.13 (x86_64-little_endian-lp64 shared (dynamic) release build; by GCC 13.2.1 20240210), gentoo 2.15 PASS : PreallocationTest::initTestCase() FAIL! : PreallocationTest::testPreallocationMultiFileCache() 'prealloc.bytesWritten() == multi_tor.getTotalSize()' returned FALSE. () Loc: [/var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5/src/diskio/tests/preallocationtest.cpp(87)] FAIL! : PreallocationTest::testPreallocationSingleFileCache() 'prealloc.bytesWritten() == single_tor.getTotalSize()' returned FALSE. () Loc: [/var/tmp/portage/net-libs/libktorrent-23.08.5/work/libktorrent-23.08.5/src/diskio/tests/preallocationtest.cpp(111)] PASS : PreallocationTest::cleanupTestCase() Totals: 2 passed, 2 failed, 0 skipped, 0 blacklisted, 1360ms ********* Finished testing of PreallocationTest ********* <end of output> Test time = 1.39 sec ---------------------------------------------------------- Test Failed. "preallocationtest" end time: Jun 16 16:38 UTC "preallocationtest" time elapsed: 00:00:01 ---------------------------------------------------------- ```