summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Add more dependency icons and high error attributes separately.user/danf/results-dbDan Fandrich2020-03-266-5/+10
|
* Show icons for some programming language dependenciesDan Fandrich2020-03-2613-5/+39
| | | | | | | | Search the package install log to find some dependencies that get turned into icons next to the package names. This should give maintainers a better clue about whether they they have the skills or interest to tackle a package build issue. Logos are displayed only for packages that are failing and have not been fixed yet.
* Update column numbersPascal Terjan2020-03-241-2/+2
| | | | | There is an untitled column full of 0, update the graphs to skip it until the data gets fixed.
* Handle new autobuild being on aarch64Pascal Terjan2020-01-271-5/+10
|
* Do not try working on FALSE file handle if "latest" link is broken, this ↵Pascal Terjan2018-04-281-19/+21
| | | | causes a lot of warnings in logs
* Detect recreate_srpm_failure getting fixedPascal Terjan2017-09-151-1/+4
|
* Load check's dependencies.rss asynchronously to avoid hangsOlivier Blin2016-10-051-11/+15
| | | | The synchronous load was hanging pkgsubmit load in case alamut was down.
* Avoid crazy durations when start is 0Pascal Terjan2016-06-211-4/+6
|
* Make the build list stablePascal Terjan2016-06-211-0/+1
|
* Fix more linksPascal Terjan2016-03-301-1/+2
|
* Fix some linksPascal Terjan2016-03-301-2/+2
|
* Handle packages submitted at the same secondPascal Terjan2016-03-301-1/+1
|
* Make the deps glob more robustPascal Terjan2016-03-301-1/+1
|
* Drop support for /etc/bs-webstatus.confOlivier Blin2016-03-041-5/+0
| | | | | | It is confusing to see that conf.php are not actually used, and the separate config file does not bring much value. Also, it was not version-controlled, while conf.php is.
* Fix one last warningPascal Terjan2016-02-101-1/+5
|
* Fix 3 more warningsPascal Terjan2016-02-101-3/+3
|
* Remove another very noisy errorPascal Terjan2016-02-101-1/+1
|
* Avoid a warning flooding logsPascal Terjan2016-02-101-5/+7
|
* Use the most recent install_deps, in case of a retryPascal Terjan2016-01-151-3/+4
|
* Fixed configuration of number of build nodesDan Fandrich2016-01-121-2/+2
|
* Add a page listed broken packages on non mandatory archesPascal Terjan2016-01-121-0/+87
|
* Add a link on failed architecturesPascal Terjan2015-11-072-22/+11
|
* Use new official domain name for madb (.mageia.org)Olivier Blin2015-11-071-1/+1
|
* Do not show builds as partial when a .fail file exists in the done directoryOlivier Blin2015-11-051-1/+3
| | | | | | | | | | With non-mandatory arch support, a .fail file may exist in the done directory while the build is still on going for ongoing arches. Thus, the build is not failed, and the status would not be "done" (partial build) either. Only consider a build as "done" (partial) when a .done file is present.
* Handle BS files that may not set a statusOlivier Blin2015-11-051-1/+5
| | | | | .fail files for non mandatory-arch do not update the status variable. Thus the status for a previous file would have been incorrectly used.
* Make noarch failure fatal as wellOlivier Blin2015-11-031-1/+1
|
* List all failed arches in status result, even if build for mandatory arches ↵Olivier Blin2015-11-032-0/+8
| | | | is successful
* Detect failures for mandatory arches only using .fail filesOlivier Blin2015-11-031-4/+17
| | | | | | | | | | Previously, we just used the failures directory to check if any failure had happened for the submit prefix. But this does not work properly when non-mandatory architectures are used, their failure should not be fatal. Instead, use .fail files and mark as a failure only if failed for one of the mandatory architectures.
* Declare mandatory arches in conf.php and pass to get_refined_packages_listOlivier Blin2015-11-033-2/+5
|
* Display those failing builds that appear to be flakyDan Fandrich2014-10-032-0/+54
| | | | | The flakiness algorithm is tuned for packages that are >20% flaky over the past ~6 months, which may turn out to be too high a threshold.
* Stack build resultsPascal Terjan2014-05-231-0/+1
|
* Fix MGA4 datePascal Terjan2014-05-231-1/+1
|
* Fix fixed packages detection and hide successful packagesPascal Terjan2014-05-231-9/+56
|
* Add MGA4 Final Release date to autobuild graphPascal Terjan2014-02-071-0/+1
|
* Small grpah improvementsPascal Terjan2013-11-261-4/+13
|
* fix link to source codeOlav Vitters2013-10-161-1/+1
|
* Remove bugzilla query for number of updates to testNicolas Vigier2013-02-231-9/+3
|
* Remove useless ifNicolas Vigier2013-02-231-20/+16
|
* Fix path to unmaintained.txtNicolas Vigier2013-02-231-1/+2
|
* Add build history of packagesPascal Terjan2013-01-222-12/+48
|
* Restore the "new" featurePascal Terjan2013-01-211-1/+4
|
* More compact list of builds (would need a non js version too :( )Pascal Terjan2013-01-211-10/+5
|
* Add a simple version creating a fake status.log with packages still brokenPascal Terjan2013-01-171-0/+50
|
* Use the time of the specific package, it may have been fixed before the end ↵Pascal Terjan2013-01-161-5/+3
| | | | of the run
* Use modification time to detect fixed packages instead of version, to ↵Pascal Terjan2013-01-161-8/+15
| | | | support when it does not get changed
* Ignore version for new breakgae detectionPascal Terjan2013-01-141-2/+5
|
* Support failure being on a more recent version than the one on the mirrors, ↵Pascal Terjan2013-01-141-1/+1
| | | | for fake autobuild
* Support absence of logs for success tooPascal Terjan2013-01-141-1/+5
|
* Use src.rpm presence to define todo status, .info remain after buildingPascal Terjan2013-01-131-1/+1
|
* Limit the table to the 10 most recent buildsPascal Terjan2013-01-051-1/+1
|