Nguyễn Gia Phong via "Development of GNU Guix and the GNU System distribution." <[email protected]> writes:
> Hi Rodion, > > On 2026-09-19 at 08:21+02:00, Rodion Goritskov wrote: >> Nguyễn Gia Phong <[email protected]> writes: >> > According to data.qa, ghc (3400 dependents) and webkitgtk-for-gtk3 >> > (343 dependents) are failing to build. Do you know if they are flaky? >> > >> > https://data.qa.guix.gnu.org/revision/24fa2773955522917a4a3d3c9ec911b74a9eacaa/blocking-builds >> >> As far as I can see, both ghc and webkitgtk-for-gtk3 are OK and kde-team >> merged to master by Sharlatan. > > I think as mentioned by Oleg/Sharlatan, QA's reporting were drunk > and didn't reflect the actual state which could be queried by weather. I didn't see the reference to "drunk", but I did see the reference to "stuck" here https://issues.guix.gnu.org/81416#31 Stuck is somewhat helpful, but it would be useful to know if the information regarding the branch was out of date (stuck), or whether builds weren't being submitted, or whether the build information was out of date. What do you mean by drunk? The "actual state which could be queried by weather" is a concerning statement. As far as I'm aware, the changes on kde-team affected around ~18,000 packages (potentially more). So checking the status of just ~600 doesn't really tell you anything about the effect of merging the branch? I'm pretty confident this isn't a "QA's reporting were drunk" issue, ci.guix.gnu.org agrees that there's much more on this branch than changes just affecting a few small package manifests: https://ci.guix.gnu.org/jobset/kde-team (see eval 2203893 for example, I'm not sure what the number is, but it's surely bigger than 600).
signature.asc
Description: PGP signature
