Note: This is a public test instance of Red Hat Bugzilla. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback at bugzilla.redhat.com.
Bug 1803711
Summary: | perl-Tk-804.034-9.fc33 FTBFS: t/entry.t test fails | ||||||
---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Petr Pisar <ppisar> | ||||
Component: | perl-Tk | Assignee: | Xavier Bachelot <xavier> | ||||
Status: | CLOSED ERRATA | QA Contact: | Fedora Extras Quality Assurance <extras-qa> | ||||
Severity: | unspecified | Docs Contact: | |||||
Priority: | unspecified | ||||||
Version: | rawhide | CC: | andreas.bierfert, jplesnik, mchehab, perl-devel, tremble, xavier | ||||
Target Milestone: | --- | Keywords: | Reopened | ||||
Target Release: | --- | ||||||
Hardware: | Unspecified | ||||||
OS: | Unspecified | ||||||
URL: | https://koji.fedoraproject.org/koji/buildinfo?buildID=1463067 | ||||||
Whiteboard: | |||||||
Fixed In Version: | perl-Tk-804.035-1.fc33 perl-Tk-804.035-1.fc32 perl-Tk-804.036-3.fc34 perl-Tk-804.036-3.fc33 | Doc Type: | If docs needed, set a value | ||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2021-07-08 00:57:04 UTC | Type: | Bug | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Bug Depends On: | |||||||
Bug Blocks: | 1818509 | ||||||
Attachments: |
|
Description
Petr Pisar
2020-02-17 09:57:12 UTC
Those are the rebased packages between last successful build and now: Removed packages: annobin-9.01 binutils-2.33.1 binutils-gold-2.33.1 cpio-2.12 fedora-gpg-keys-32 fedora-release-32 fedora-release-common-32 fedora-repos-32 fedora-repos-rawhide-32 freetype-2.10.0 freetype-devel-2.10.0 gdb-minimal-9.0.50.20191119 glib2-2.63.4 glibc-2.30.9000 glibc-common-2.30.9000 glibc-devel-2.30.9000 glibc-headers-2.30.9000 glibc-minimal-langpack-2.30.9000 hwdata-0.331 kernel-headers-5.5.0 krb5-libs-1.17.1 libarchive-3.4.0 libblkid-2.35 liberation-fonts-common-2.00.5 liberation-sans-fonts-2.00.5 libfdisk-2.35 libmount-2.35 libsmartcols-2.35 libtasn1-4.15.0 libuuid-2.35 libxcrypt-4.4.11 libxcrypt-devel-4.4.11 libXt-1.1.5 mesa-libGL-19.3.3 mesa-libglapi-19.3.3 pcre-8.43 perl-Devel-PPPort-3.56 perl-parent-0.237 perl-Scalar-List-Utils-1.53 python3-3.8.1 python3-libs-3.8.1 redhat-rpm-config-149 sed-4.5 sqlite-libs-3.30.1 systemd-libs-244.1 util-linux-2.35 xkeyboard-config-2.28 Added packages: annobin-9.07 binutils-2.34 binutils-gold-2.34 cpio-2.13 dejavu-fonts-common-2.37 dejavu-sans-fonts-2.37 fedora-gpg-keys-33 fedora-release-33 fedora-release-common-33 fedora-repos-33 fedora-repos-rawhide-33 freetype-2.10.1 freetype-devel-2.10.1 gdb-minimal-9.1 glib2-2.63.5 glibc-2.31 glibc-common-2.31 glibc-devel-2.31 glibc-headers-2.31 glibc-minimal-langpack-2.31 guile-2.0.14 hwdata-0.332 kernel-headers-5.6.0 krb5-libs-1.18 langpacks-core-font-en-3.0 libarchive-3.4.2 libblkid-2.35.1 liberation-fonts-common-2.1.0 liberation-sans-fonts-2.1.0 libfdisk-2.35.1 libmount-2.35.1 libsmartcols-2.35.1 libtasn1-4.16.0 libuuid-2.35.1 libxcrypt-4.4.13 libxcrypt-devel-4.4.13 libXt-1.2.0 mesa-libGL-20.0.0~rc2 mesa-libglapi-20.0.0~rc2 pcre-8.44 perl-Devel-PPPort-3.57 perl-parent-0.238 perl-Scalar-List-Utils-1.54 python3-3.8.2~rc1 python3-libs-3.8.2~rc1 redhat-rpm-config-151 sed-4.8 sqlite-libs-3.31.1 systemd-libs-245~rc1 util-linux-2.35.1 xkeyboard-config-2.29 xxhash-libs-0.7.2 804.035 doesn't help, same test failures, both on F32 and F33. F31 is okay. The problem is probably related to the fonts update. Notes from Tk-804.035/README: * There are a lot more tests in Tk804. Some notably t/entry.t and t/listbox.t very dependant on the available fonts and to a lesser extent the window manager used. * By default Perl/Tk makes use of client side fonts via freetype2, fontconfig and Xft on Unix systems. This gives anti-aliased fonts for regular text (if you have TrueType or Type1 fonts and they are in your fontconfig config file). See also http://fontconfig.org . To disable this feature run the Makefile.PL with "perl Makefile.PL XFT=0". I tried to build Tk-804.035 with XFT=0 and the package was build successfully. I am preparing rebuild for Perl 5.32. May I update the package to Tk-804.035 and change XFT from 1 to 0? Sure, please go ahead. FEDORA-2020-a726a87414 has been submitted as an update to Fedora 32. https://bodhi.fedoraproject.org/updates/FEDORA-2020-a726a87414 FEDORA-2020-a726a87414 has been pushed to the Fedora 32 testing repository. In short time you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2020-a726a87414` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2020-a726a87414 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-2020-a726a87414 has been pushed to the Fedora 32 stable repository. If problem still persists, please make note of it in this bug report. (In reply to Jitka Plesnikova from comment #3) > The problem is probably related to the fonts update. I don't think so. Building it by hand, with XFT=1 and either doing "make test" or running the unit test alone: $ perl -It -MTkTest -e "checked_test_harness('./xt', 0, 'blib/lib', 'blib/arch')" t/entry.t t/entry.t .. ok All tests successful. Files=1, Tests=351, 2 wallclock secs ( 0.04 usr 0.01 sys + 2.10 cusr 0.03 csys = 2.18 CPU) Result: PASS Works. The only difference I see is that, when building with mock, the output is not displayed. I'm starting to suspect that the unit test is failing due to some limitation at mock to emulate a X11/Wayland console, e. g. probably some incompatibility with xorg-x11-server-Xvfb. (In reply to Mauro Carvalho Chehab from comment #9) > (In reply to Jitka Plesnikova from comment #3) > > The problem is probably related to the fonts update. > > I don't think so. Building it by hand, with XFT=1 and either doing "make > test" or running the unit test alone: > > > $ perl -It -MTkTest -e "checked_test_harness('./xt', 0, 'blib/lib', > 'blib/arch')" t/entry.t > t/entry.t .. ok > All tests successful. > Files=1, Tests=351, 2 wallclock secs ( 0.04 usr 0.01 sys + 2.10 cusr > 0.03 csys = 2.18 CPU) > Result: PASS > > Works. > > The only difference I see is that, when building with mock, the output is > not displayed. I'm starting to suspect that the unit test is failing due to > some limitation at mock to emulate a X11/Wayland console, e. g. probably > some incompatibility with xorg-x11-server-Xvfb. Funny enough, running: xvfb-run -a perl -It -MTkTest -e "checked_test_harness('./xt', 0, 'blib/lib', 'blib/arch')" t/entry.t or xvfb-run -a make test works fine too. Perhaps it requires some extra package for the tests to pass on mock? Ok, with this small patch: diff --git a/t/entry.t b/t/entry.t index c9e99b87ab81..7192776685ad 100755 --- a/t/entry.t +++ b/t/entry.t @@ -677,6 +677,7 @@ eval { $e->destroy }; $e = $mw->Entry(-font => $fixed, -bd => 2, -relief => "raised")->pack; $e->insert(end => "0123"); $e->update; +diag sprintf "Font: $fixed, Geometry data: %s", $e->geometry; is($e->index('@10'), 0, "index with raised relief"); with_fixed_font { is($e->index('@11'), 0) }; is($e->index('@12'), 1); It can be seen the differences between a build on a machine with X11 properly installed: $ perl -It -MTkTest -e 'checked_test_harness('\''./xt'\'', 0, '\''blib/lib'\'', '\''blib/arch'\'')' t/entry.t t/entry.t .. 193/351 # Font: -adobe-courier-medium-r-normal--12-120-75-75-m-*-iso8859-1, Geometry data: 150x23+0+0 t/entry.t .. ok All tests successful. Files=1, Tests=351, 2 wallclock secs ( 0.04 usr 0.01 sys + 1.95 cusr 0.03 csys = 2.03 CPU) Result: PASS And a build with mock: + xvfb-run -a perl -It -MTkTest -e 'checked_test_harness('\''./xt'\'', 0, '\''blib/lib'\'', '\''blib/arch'\'')' t/entry.t # # # Font: -adobe-courier-medium-r-normal--12-120-75-75-m-*-iso8859-1, Geometry data: 170x25+0+0 # # # Failed test at t/entry.t line 684. # got: '0' # expected: '1' # Failed test at t/entry.t line 693. # got: '0' # expected: '1' # Failed test at t/entry.t line 746. # got: '5' # expected: '6' # Failed test at t/entry.t line 1219. # got: '4' # expected: '5' # Looks like you failed 4 tests of 351. t/entry.t .. Dubious, test returned 4 (wstat 1024, 0x400) Failed 4/351 subtests (less 39 skipped subtests: 308 okay) I suspect that some additional packages are required for the adobe-currier font, libXft and fontconfig to be properly installed/initialized inside mock chroot. Found it! events.t requires google-noto-sans-fonts in order to pass. I'll enclose a patch on this BZ. Created attachment 1729853 [details]
Re-enable TrueType support by adding a missing font
I doubt google-noto-sans-fonts is the cause. This package predates Noto font. Also you can see that the font is not listed in the installed package set <https://kojipkgs.fedoraproject.org//packages/perl-Tk/804.034/8.fc32/data/logs/x86_64/root.log> for the last successful build before the XFT was disabled. I'd rather bet on FontConfig cache not being updated or the metrics of the exhibited font being removed or moved to a different package. Also note that -adobe-courier-medium-r-normal--12-120-75-75-m-*-iso8859-1 is the old non-FT font specification. So if indeed that font was used, then no XFT was involved. (In reply to Petr Pisar from comment #14) > I doubt google-noto-sans-fonts is the cause. This package predates Noto > font. Also you can see that the font is not listed in the installed package > set > <https://kojipkgs.fedoraproject.org//packages/perl-Tk/804.034/8.fc32/data/ > logs/x86_64/root.log> for the last successful build before the XFT was > disabled. I'd rather bet on FontConfig cache not being updated or the > metrics of the exhibited font being removed or moved to a different package. > Also note that -adobe-courier-medium-r-normal--12-120-75-75-m-*-iso8859-1 is > the old non-FT font specification. So if indeed that font was used, then no > XFT was involved. I tested installing xorg-x11-fonts-ISO8859-1-75dpi, which should be carrying adobe-courier font, but installing it didn't make any change. I agree that this could be due to FontConfig cache, but it could also be due to some font replacement setup that would happen when this font is installed. In either case, before coming up with this solution, I tried to install other packages that would make more sense, but everything failed until this one got installed. After installing it, mock was able to successfully the package on both Fedora 33 and rawhide. This bug appears to have been reported against 'rawhide' during the Fedora 34 development cycle. Changing version to 34. FEDORA-2021-49127bd223 has been submitted as an update to Fedora 33. https://bodhi.fedoraproject.org/updates/FEDORA-2021-49127bd223 FEDORA-2021-49127bd223 has been pushed to the Fedora 33 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2021-49127bd223` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2021-49127bd223 See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-2021-639b8bfb6e has been pushed to the Fedora 34 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --advisory=FEDORA-2021-639b8bfb6e` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2021-639b8bfb6e See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates. FEDORA-2021-639b8bfb6e has been pushed to the Fedora 34 stable repository. If problem still persists, please make note of it in this bug report. FEDORA-2021-49127bd223 has been pushed to the Fedora 33 stable repository. If problem still persists, please make note of it in this bug report. |