summaryrefslogtreecommitdiffstats
path: root/t
Commit message (Collapse)AuthorAgeFilesLines
...
* cpan_testers: fix a warning when skipping testsThierry Vignaud2020-04-091-2/+10
| | | | should have been in commit ad953e4c4bca7e5bd9f5b5664298e04272537479
* cpan_testers: workaround 'nogroup' not existing on CentOSThierry Vignaud2020-04-091-1/+1
|
* cpan_testers: workaround disabled nobody accountThierry Vignaud2020-04-091-1/+1
| | | | | On CentOS7, the error message is: 'This account is currently not available.'
* cpan_testers: fix a test with rpm < 4.13.0Thierry Vignaud2020-04-091-3/+5
| | | | | | | | It was failing for older versions of rpm since commit 32c4bc874e3f7fc99188bf045902b1018f38855a This fix the counting & adjust to missing pkg name when erasing with older rpms
* cpan_testers: set PATHThierry Vignaud2020-04-091-0/+3
| | | | | rationale: CPAN testers installs genhdlist2 using cpan but do not have /usr/local/bin in their PATH
* cpan_testers: skip weak deps test on CentOS 7 (unsupported)Thierry Vignaud2020-04-095-3/+27
| | | | | | | | | | | | rationale: CentOS 7's rpm doesn't support eg Recommendstag, it only knows their definitions but is unable to parse a spec file using Recommends Ideally we should be able to detect a too old/unpatched rpm... Technically, we could check whether it's Mageia < 6 which did have an non upstream implementation (but was using Suggests tag instead of Recommends and so would need testsuite adaption -- not worth it)
* cpan_testers: skip mirrorlist tests on !MageiaThierry Vignaud2020-01-171-0/+2
| | | | | Rationale: It needs a Mageia specific patch that introduces Time::ZoneInfo->current_zone()"
* fix testsuite errors on CentOs (NVR vs NVRA)Thierry Vignaud2020-01-177-13/+13
| | | | | | where rpmquery default to outputing NVRA (vs NVR on Mageia) reuse check_installed_fullnames() where needed instead of introducing %{NVR}
* skip test if /proc is not mountedThierry Vignaud2020-01-171-0/+1
| | | | | no need to show red in cpan-testers matrix if local chroot's /proc isn't mounted :-)
* explain build processThierry Vignaud2020-01-171-0/+4
|
* fix failing scriptlets test with rpm < 4.13.0Thierry Vignaud2020-01-171-1/+6
| | | | | it got broken when wee added support for rpm 4.13.0 in commit cdcb6c6359e8ee2a590c3c40aa3b4d57ec3b1e80
* fix "Undefined subroutine &urpm::select::_rpm_version"Thierry Vignaud2020-01-081-0/+1
| | | | | | | Undefined subroutine &urpm::select::_rpm_version called at t/superuser--media_info_dir.t line 55. which wents undetected by the testsuite for quite some time... sigh... bug introduced in commit 60b5009cc4679d4b55fb0490d4d273b5cc6f220c
* skip tests if we have no supported downloaderThierry Vignaud2020-01-081-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | as seen in CPAN reports: <<< ...retrieving failed: no webfetch found, supported webfetch are: curl, wget, prozilla, aria2 no metadata found for medium "various" >>> Whic results in: <<< Test Summary Report ------------------- t/superuser--http.t (Wstat: 2560 Tests: 11 Failed: 10) Failed tests: 2-11 Non-zero exit status: 10 Files=40, Tests=3775, 953 wallclock secs ( 1.19 usr 0.08 sys + 309.33 cusr 52.13 csys = 362.73 CPU) Result: FAIL Failed 1/40 test programs. 10/3775 subtests failed. make: *** [Makefile:1155: test_dynamic] Error 255 (/usr/bin/make test exited with 512) CPAN::Reporter: Test result is 'fail', One or more tests failed. >>> And indeed: $ rpm -q curl wget prozilla aria2 package curl is not installed package wget is not installed package prozilla is not installed package aria2 is not installed Should have been done in commit 360f05b8db2029cb62c11f7eddb7fb57f3b774f1
* (need_downloader) split it for next commitThierry Vignaud2020-01-072-12/+17
| | | | also switching one character back from UTF to ASCII
* skip network tests if we don't have a downloaderThierry Vignaud2018-09-201-0/+11
|
* make testsuite work with rpm < 4.14.2 againThierry Vignaud2018-09-191-1/+2
|
* fix testsuite with rpm-4.14.2Thierry Vignaud2018-09-171-1/+1
|
* remove commented out debug lineThierry Vignaud2018-07-021-1/+0
|
* perl_checker cleanupsThierry Vignaud2018-07-022-5/+5
|
* fix buggy rpms test with rpm-4.14.2Thierry Vignaud2018-07-021-1/+2
| | | | | | | explanations: rpm-4.14.2 spits more error messsages which can mess up Test::* if eg some space or other characters happen before "ok X" so just redirect all rpm output to /dev/null
* adapt testsuite to rpm-4.14.2 RC1Thierry Vignaud2018-07-023-3/+5
| | | | | | we need to define %_pkgverify_level to none else locally build packages will derail the testsuite as some pkg installations will fail with eg: package nls-1.0-2.i386 does not verify: no digest
* adapt testsuite to rpm-4.13.0~rc2Thierry Vignaud2016-10-201-1/+11
| | | | ...which fixed %preun not aborting package erasure
* fix a testsuite warningThierry Vignaud2015-08-041-1/+1
| | | | | | | | Explicitely adding libnss3.so fixes installing that package: warning: %post(rpm-query-in-scriptlet-1-1.x86_64) scriptlet failed, exit status 127 Which is b/c of: error: Failed to initialize NSS library
* speedup build by skipping debuginfoThierry Vignaud2015-08-041-0/+3
|
* do check that %%pretrans do fail (rpm-4.10+)Thierry Vignaud2015-08-041-5/+1
| | | | | | | basically reverting commit 5d7414cd5dc0c7dbdfe5314df4253cb14f86e9b2 and inversing the test expected result needs rpm-4.12.90-2.mga6 or newer
* convert trigger scriptlets from lua to shThierry Vignaud2015-08-0414-42/+42
| | | | | needed for next commit because: lua's exit(1) is ignored since rpm-4.9.0
* install sh as dep for failing-scriplets pkgsThierry Vignaud2015-08-041-4/+7
|
* introduce new helper funcThierry Vignaud2015-08-041-0/+5
| | | | needed for next commits
* add /bin/sh providerThierry Vignaud2015-08-041-0/+12
| | | | (needed for next commit)
* kill useless Group: definitionsThierry Vignaud2015-08-04298-299/+0
| | | | uneeded since rpm-4.6.0
* kill useless Url: definitionsThierry Vignaud2015-08-04298-298/+0
|
* kill useless buildroot definitionsThierry Vignaud2015-08-04279-279/+0
| | | | (uneeded since rpm-4.6+ ...)
* disable debuginfoThierry Vignaud2015-08-041-0/+3
| | | | thus fixing noise in testsuite output
* use new option nameThierry Vignaud2014-09-071-1/+1
|
* check that --no-suggests alias still worksThierry Vignaud2014-09-071-0/+2
|
* s!--no-suggests!--no-recommends! in testsuiteThierry Vignaud2014-09-071-5/+5
|
* s/Suggests:/Recommends:/ in testsuiteThierry Vignaud2014-09-077-8/+8
|
* add 2 other possible transaction ordersThierry Vignaud2014-02-191-0/+2
| | | | needed due to perl-5.18 random keys order
* add 2 other possible transaction ordersThierry Vignaud2013-10-041-0/+2
| | | | | | needed due to perl-5.18 random keys order it could alternatively be workarounded by sorting packages in genhdlist2
* factorize possible transaction sets definitionThierry Vignaud2013-10-041-19/+18
| | | | (needed for next commit)
* simplify: use member()Thierry Vignaud2013-10-041-1/+1
| | | | (needed for next commit)
* adapt to perl-5.18's random hash keys orderingThierry Vignaud2013-10-041-2/+2
| | | | | | | | | | | | | | | | | | | | (translate_why_removed_one) ensure we got the right reason (testsuite) adapt to new URPM predictible behaviour it was previously relying on the following fixed hash order: $ perl -MData::Dumper -e 'warn Dumper [ keys {'rr1', 'rr1', 'rr2', 'rr2'} ]; \ warn Dumper [ keys {'bb1', 'bb1', 'bb2', 'bb2-1'} ]' $VAR1 = [ 'rr2', 'rr1' ]; $VAR1 = [ 'bb1', 'bb2' ];
* (install_logger) keep a separate counter for erasuresThierry Vignaud2013-02-091-3/+3
| | | | | | | | | | only initialize $total_pkg: - when starting a new transaction - or on first erasure in transaction (_compute_pkg_total) kill it (no more needed) adjust testsuite accordingly
* (install_logger) unify with callback_uninst(), thus fixing counting forThierry Vignaud2013-02-051-4/+5
| | | | | | | | | | erasures (callback_uninst) restore quiet behavior when not in verbose mode (install) temporary set $urpm->{trans} adapt testsuite
* (callback_uninst) display erasure progressThierry Vignaud2013-02-031-0/+1
| | | | | | inspirated from install_logger() adjust testsuite accordingly
* wrap commentThierry Vignaud2013-02-031-3/+2
|
* fix that test when prelink is installedThierry Vignaud2013-01-311-0/+7
|
* revert my previous commits, this was a bad anywayGuillaume Rousse2012-12-105-79/+11
|
* drop executable bit, as for other testsGuillaume Rousse2012-12-083-0/+0
|
* obsoleted by pod-syntax.tGuillaume Rousse2012-12-061-5/+0
|