aboutsummaryrefslogtreecommitdiffstats
path: root/modules/buildsystem/templates
Commit message (Collapse)AuthorAgeFilesLines
* Switch links to https: where possibleDan Fandrich2024-01-253-3/+3
| | | | | | | All these point to valid https: resources, but there is a small chance that some unusual interaction will cause it not to work. Some of these changes also won't take effect until the server is restarted, so we'll need to keep this in mind if failures occur long from now.
* Cleanup duplicate src.rpmPascal Terjan2024-01-031-2/+15
| | | | | | | Sometimes someone moves a package manually and we end up with several versions of a package. In this case we should delete the src.rpm and handle its binary packages like if it was gone.
* Fix cleaner.rbPascal Terjan2023-01-082-17/+23
| | | | | | It did not really wait for 7d for removed subpackages. Also remove an incorrect condition, src.rpm is usually slightly older that binary rpms, and fix the tests.
* Cleanup packages no longer built on an archPascal Terjan2022-09-301-4/+23
|
* Remove debug mesagesPascal Terjan2022-09-251-3/+0
|
* Add an option to use upload lockPascal Terjan2022-09-251-3/+34
|
* Fix the test to really verify the fixPascal Terjan2022-09-231-1/+1
| | | | | | | That test would pass even without the fix as it would be considered that the package hasn't been rebuilt for this arch yet. The bug could only happen when there was also another subpackage remaining.
* Fix packages being deleted too early during uploadPascal Terjan2022-09-222-9/+34
|
* Also clean tainted and nonfree by defaultPascal Terjan2022-08-271-38/+39
|
* Delete leftover noarch without waiting for 7dPascal Terjan2021-01-192-12/+24
|
* Fix cleanup to correctly wait for 7dPascal Terjan2020-12-312-6/+6
| | | | | | It was checking if there is another version of the src, but using the name of the package instead of the name of the src, and those are usually different in the case we care most about (libraries).
* Fix apache configuration for autobuildNicolas Lécureuil2020-12-011-1/+1
|
* Fix getting timestamp of non noarch packagesPascal Terjan2020-09-301-3/+3
| | | | | | | urpmf --qf '%filename:%buildarchs:%buildtime' shifts fields when buildarchs is empty: python-sphinxcontrib-svg2pdfconverter-0.1.0-2.mga8.src.rpm:noarch:1564917727 python-argon2_cffi-19.1.0-1.mga8.src.rpm:1564918990:
* Delete again non noarch packagesPascal Terjan2020-09-301-2/+5
|
* Clean pure noarch packages toPascal Terjan2020-09-261-4/+14
|
* More repository vhost fixesPascal Terjan2020-06-291-5/+5
| | | | This time it really looks like what I had tested...
* More repository vhost fixesPascal Terjan2020-06-291-4/+4
|
* Simplify repository vhostPascal Terjan2020-06-281-8/+10
| | | | | | | | | We use the same access list for all versions of the distribution on the HTPP server (our internal machines) so there is no reason to duplicate it per distro. This also allows to access files at the top level directory (like mageia_timestamp).
* remove --ignorearch option for urpmiThierry Vignaud2020-04-181-2/+0
| | | | | | | | | | rationale: it badly broke installing some packages with rpm-4.16 which add provides for every generated package to the SRPM See https://rpm.org/wiki/Releases/4.16.0 Basically reverting commit cc1079609497da5d5700e27f1076d57bf2b9282c and its 2 followup fixups
* Allow larger log sizesPascal Terjan2020-03-201-0/+1
|
* Fix syntax error in mirror cleanerPascal Terjan2020-02-261-1/+1
|
* vhost_webstatus: set dynamic NameWidthThomas Backlund2020-02-191-0/+2
|
* Add a --auto to mga-clean-distribPascal Terjan2019-10-271-9/+14
| | | | I think it is now safe enough to run in cron
* Make the cleanup script smarterPascal Terjan2019-10-232-92/+174
| | | | | | | | | Also, refactor it and add tests Now subpackages of a package not yet rebuilt on an architecture do not get clean on that architecture (even the noarch ones from that package) and packages get deleted only if the src package is totally gone or if the new version of it is at least 7 days old.
* Build %dist based on other macrosPascal Terjan2018-11-241-0/+1
| | | | | This means temporary src.rpm generated on the main build node will have correct suffix.
* Do not set arch in infra media.cfg for armv7hlPascal Terjan2018-06-131-0/+6
|
* Allow libguestfs to access network during buildPascal Terjan2017-10-131-0/+7
|
* Do not forget about remote builds after 10h, 100h looks more realisticPascal Terjan2016-10-151-0/+1
|
* Fix typoNicolas Lécureuil2016-05-161-1/+1
|
* Fix typo, missing semi-colonNicolas Lécureuil2016-05-161-1/+1
|
* Ignore architecture when installing build chrootPascal Terjan2016-05-161-0/+2
|
* Remove hack, armv5tl now exists on normal mirrorPascal Terjan2016-03-241-3/+0
|
* Kill ssh connections after they have been dead for 30sPascal Terjan2016-03-031-0/+1
|
* Also switch to Apache 2.4 auth config for pkgsubmitPascal Terjan2016-02-071-2/+2
|
* Switch to Apache 2.4 auth config for repository vhostPascal Terjan2016-02-071-8/+19
|
* buildsystem: fix youri host.conf template to use an array as well for src nodeOlivier Blin2015-11-301-1/+1
| | | | | Otherwise it failed on Mageia 5 with: Detail: undefined method `sort' for "valstar":String
* iurt: Enlarge your timeout (x 2) on ARMOlivier Blin2015-11-301-1/+1
|
* iurt: Move build timeout config to puppet fileOlivier Blin2015-11-301-7/+4
|
* Expose armv5tl bootstrap as mirror on internal repositoryPascal Terjan2015-11-011-0/+3
|
* fix up maintdb breakageThomas Backlund2015-10-201-1/+1
|
* lint fixes for buildsystemThomas Backlund2015-10-205-56/+56
|
* buildsystem: maintdb allow *-team maintainers.Colin Guthrie2014-09-191-2/+2
| | | | | | | | | | | *-team maintainers behave just like 'nobody' but can be used to indicate packages that actually are maintained, but are handled by a team rather than an individual. This doesn't integrate *-team anywhere else in the stack it just allows the maintainers to be set. Anything that assumes a maintainer in the maintdb is a user (modulo 'nobody') would need to be updated to handle this change.
* buildsystem: maintdb slight modernisation+use bash explicitly.Colin Guthrie2014-09-191-28/+15
| | | | | | This is slightly subjective but most scripts do it this way these days without the extra letters on string comparisons and collapsing the if+then onto one line.
* buildsystem: maintdb tabs->spacesColin Guthrie2014-09-191-33/+32
|
* buildsystem: Future proof SRPM check URL for backports.Colin Guthrie2014-08-121-2/+2
|
* Add a super simple API for checking whether an update SRPM exists.Colin Guthrie2014-08-121-0/+9
| | | | | This will be useful for advisory publication checks to make sure the SRPMs actually exist prior to assigning an advisory ID
* add backports tree to allowed listThomas Backlund2014-01-231-1/+1
|
* remove unused media_types definition (we rely on media_type)Thomas Backlund2013-11-021-5/+0
|
* cinnamon and mate are mga4 accepted featuresThomas Backlund2013-07-211-9/+0
|
* buildsystem::distros: remove iurt_use_bootstrap optionNicolas Vigier2013-05-311-4/+0
| | | | This option is no longer needed with the no_mirror option.
generated by cgit v1.2.1 (git 2.21.0) at 2024-10-22 08:26:42 +0000