tag 815292 + upstream thanks Hi,
On Sat, Feb 20, 2016 at 05:41:46PM +0100, Andreas Beckmann wrote: > libreoffice FTBFS on i386: Yes, I know. I *do* look at build logs. Filing a bug report for that is not exactly helpful. > https://buildd.debian.org/status/fetch.php?pkg=libreoffice&arch=i386&ver=1%3A5.1.1~rc1-1&stamp=1455588828 > > [...] > FDO74774.docx,526 > File tested,Execution Time (ms) > FDO74774.docx,1483 > xmltesttools.cxx:99:Assertion > Test name: testFlipAndRotateCustomShape::Import_Export_Import > equality assertion failed > - Expected: 2351 > - Actual : 2350 > - In <file:///tmp/lu24199oryx4t.tmp>, attribute 'x' of > '//a:custGeom/a:pathLst/a:path/a:lnTo[1]/a:pt' incorrect value. > > Failures !!! > Run: 104 Failure total: 1 Failures: 1 Errors: 0 >From IRC from Tuesday, 16: 09:15 < _rene_> is a sw_ooxmlexport7 test failure (only on i386, not on amd64(!)) known (in 5.1.1 rc1)? 09:16 < _rene_> https://buildd.debian.org/status/fetch.php?pkg=libreoffice&arch=i386&ver=1%3A5.1.1~rc1-1&stamp=1455588828 09:17 -!- sheikha_ [~sheikha@82.178.73.79] has quit [Ping timeout: 240 seconds] 09:17 < chris_wot> oh great, another off by one error 09:18 -!- aerozeppelin [~aeroze...@pool-108-23-243-158.lsanca.fios.verizon.net] has quit [Quit: Leaving] 09:18 < beluga_> yes it was me 09:18 < chris_wot> ah - :-) sorry about that 09:19 < chris_wot> anyway, I'm a very happy chappy :-) 09:19 -!- loircbot [~ir...@vm148.documentfoundation.org] has joined #libreoffice-dev 09:19 < loircbot> LibreOffice (core) vmiklos * (11 files in 7 dirs): Related: tdf#59699 DOCX import: fix linked graphics with relative URLs 09:19 < chris_wot> hopefully the users will be too 09:19 < chris_wot> I've got a gerrit patch queued to be backported to 5-1 09:20 * chris_wot idly wonders how to configure only i386... 09:20 < _rene_> hmm? you mean me? 09:20 < _rene_> of course by building on a i386 box? :) 09:20 < chris_wot> _rene_ sorry, yeah 09:20 -!- marcosps [~Thunderbi@177.96.45.3] has joined #libreoffice-dev 09:21 < chris_wot> isn't there a way of cross compiling and forcing it back to i386 though? 09:21 < _rene_> (64bit kernel might be fine, though, possible you need linux32) 09:21 < chris_wot> _rene_ yeah, I suspect so because I'm not getting that error on my 64 system 09:21 -!- Rohan_Kumar [~socomo22@118.102.179.106] has quit [Quit: Rohan_Kumar] 09:22 < chris_wot> the unit test that's failing is literally finding that the pixel count is off by 1 pixel 09:22 < chris_wot> this happens on my Mac also, same test, same error 09:23 < chris_wot> only on mine it's out by 2 pixels 09:23 < chris_wot> some sort of rounding error? 09:23 < _rene_> chris_wot: as I said, 64bit works 09:24 < _rene_> https://buildd.debian.org/status/package.php?p=libreoffice :) 09:24 < chris_wot> _rene_ yeah, I know :-) 09:24 -!- flook [flook@nat/iiit/x-ihhjpxegzzvrthyz] has joined #libreoffice-dev 09:24 < chris_wot> I'm sure it's happening on my system for a similar reason, even though it's a Mac with a 64bit processor 09:24 < _rene_> well, some 32bit also works (powerpc), but.. ;) 09:25 < chris_wot> yeah, got to be a rounding error :( 09:25 < chris_wot> not sure the best way of fixing unfortunately, I got rather s tuck on that one 09:25 < loircbot> LibreOffice (core) [libreoffice-5-0] oliver.specht * sw/source/core/tox/ToxWhitespaceStripper.cxx: tdf#89505: replace tab character with space in table of contents 09:28 < logerritbot> A change was merged to help: tdf#95282 Help page for the COLOR function https://gerrit.libreoffice.org/22300 09:28 < IZBot> gerrit: »tdf#95282 Help page for the COLOR function« by Olivier Hallot for master [MERGED] 09:28 < loircbot> LibreOffice (help) ohallot * (3 files in 2 dirs): tdf#95282 Help page for the COLOR function 09:30 < beluga_> guys, do we need retrograde sorting? #97755 it means sorting by what is in the end of words 09:30 < IZBot> LibreOffice-Calc minor/medium UNCONFIRMED Retrograde sorting of columns https://bugs.documentfoundation.org/show_bug.cgi?id=97755 09:30 -!- quikee [~qui...@cpe-92-37-63-42.dynamic.amis.net] has quit [Remote host closed the connection] 09:31 < loircbot> LibreOffice (core) ohallot * helpcontent2: Updated core 09:31 -!- quikee [~qui...@cpe-92-37-63-42.dynamic.amis.net] has joined #libreoffice-dev 09:33 < _rene_> chris_wot: the problem is that test failures are fatal and failing the build ;) (at least on i386/amd64), so... if you get it on your own machine, why was this sone in the first place, may I ask? Regards, Rene
