[08:16] (Photo, 806x335) https://i.imgur.com/HluyrWt.jpg [08:16] @mitya57 you aware of anything such? ^ [08:17] I have a feeling some autotests failed with similar? [08:18] In 5.14.2 I removed a patch that may be related. [08:19] http://autopkgtest.ubuntu.com/packages/n/notepadqq/groovy/amd64 [08:19] tests failing with similar error [08:19] https://salsa.debian.org/qt-kde-team/qt/qtwebengine/-/commit/e0bb0919ad7c7920dfdf59be36a2de6833cd72bf [08:20] Also we are building with bundled ICU because Debian has an old version, in Ubuntu there is no such problem so maybe building with system ICU again will help. [08:22] I have a groovy VM on another machine I can test any trial PPA build you might do [08:42] @mitya57 [Interesting, vorlon reinstantiated the previous qtwebengine upload, I thought su …], copy-package takes any version as an argument [08:42] And he's an AA so he has the powers [09:03] @RikMills [I have a groovy VM on another machine I can test any trial PPA build you might d …], Will it be OK if I upload to the same PPA (4057)? [09:05] Do you want to copy it to the archive from the PPA if it tests OK? if so, I think you will need a fresh ticket. If not, then should do no harm [09:05] If it is one package then I can copy it manually. [09:06] right. I have never tried that [09:06] anyway, using that ppa should not steop us finalizing the ticket [09:17] Uploaded. [11:10] I wonder if I should have created a separate qml-module-qtqml package. Neon guys did not add it. And it breaks a few packages, e.g. kitemmodels. [11:11] @RikMills what do you think? [11:15] Looks like a separate .so file for that module was added as part of https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=43573c8df170c566 [11:16] I guess it makes sense as you did it [11:17] I think... [11:17] The documentation lists it as a separate module (https://doc.qt.io/qt-5/qtqml-qmlmodule.html) and I think it is possible to use QtQml without importing it, so maybe that was correct, yes. [11:18] In that case kitemmodels tests need to depend on it (outside of autotests directory, it is not used). [11:49] shall I upload that then? [11:50] Yes, please [11:55] done [11:55] Thanks [11:56] missing symbols for qtwebengine :( [11:57] Oops [11:57] Will fix in 15 minutes [12:28] Done. [18:22] @mitya57 https://bugreports.qt.io/browse/QTBUG-82589 [18:24] https://bugs.mageia.org/show_bug.cgi?id=26200 [18:24] bugs.mageia.org bug 26200 in RPM Packages "Kontact does not start because it is looking for qtwebengine in the wrong place" [Critical,Resolved: fixed] [18:25] Interesting that qtdiag prints the right path. [18:25] But this is without usrmerge. Do you have it enabled? [18:26] I guess so. This is a default install [18:26] Let me try to build qtbase with `-no-feature-relocatable` in that PPA. [18:27] Ok. Thanks! [18:31] Uploaded. [20:08] aha https://packaging.neon.kde.org/qt/qtbase.git/commit/debian/rules?h=Neon/release&id=2cbf7f4d5d1ea1496d76f0a159187af96b3401c3 [20:19] So this was the right change :) [20:21] FWIW, I uploaded qcustomplot which should fix the autopkgtest failure. [20:21] FYI, that's what upstream Qt recommends for system-wide installations anyway [20:22] this change should fix notepadqq tests I hope [20:22] fvogt: Thanks for confirming! [20:24] @RikMills [this change should fix notepadqq tests I hope], Maybe you can already test how kontact behaves? Qtbase built on amd64 [20:25] @mitya57 [Maybe you can already test how kontact behaves? Qtbase built on amd64], I tried a few mins ago. Had not published the amd64 in the PPA yet [20:26] "Finished 50 minutes ago" — hopefully should be published soon [20:27] Looks like it just has! [20:29] @mitya57 fixed! [20:29] Great! I am copying it to archive then. [20:31] Or better to wait until it builds on all architectures and copy with binaries? What do you think? [20:34] I think either way, some arches will not be done until after midnight my time. and riscv64 took 8 hrs last time, so waiting to binary copy that means tomorrow [20:35] so I think I would source copy or upload [20:35] that way autotests will get queued overnight [20:35] as they do not wait for riscv64 to be done [20:36] Copied with existing binaries. Apparently it was possible :) [20:36] really? nice [20:37] I assumed that would fail. handy to know [20:37] For arm and riscv64 new builds are started. [20:39] ah. I think it errors when there are binaries, but they are not published in the source archive. that must be what failed for me once before [20:40] Maybe. [20:41] I have triggered kitemmodels retests [20:42] and I will probably do some more looking in the morning [20:43] Thanks! [20:43] sadly we are are queued behind about 1.5 billion perl tests 🙄 [20:44] 😱