On Thu, Jan 29, 2026 at 11:50:04PM +0100, Ludovic Courtès wrote: (...) > Andreas Enge <[email protected]> skribis: > > > We have part of this with the cuirass-bot, but without the upper limits > > on package builds, and the build results are thrown away. > > An update: the ‘devel’ branch of Cuirass (future 1.4.0) implements > per-evaluation build limits¹ and isolated evauations² (the latter > meaning that substituting from pulls.ci would be acceptable). > > Ludo’. > > ¹ https://codeberg.org/guix/cuirass/pulls/110 > ² https://codeberg.org/guix/cuirass/pulls/74 (...)
Widening out, what is needed for 'cuirass-bot' to become close to what QA did? I assume we need to find a way to scale it's limited to a single VM right? It also seems to lose evaluations after a period of time right? If I click the evaluation (https://codeberg.org/guix/guix/pulls/5307) it doesn't have anything. The main thing I'd be missing from the 'QA' experience is a 'page' where you can see a list of builds it's done so it's easy to merge and/or follow-up on failed PRs. QA did additional checks from 'guix lint' AFAIK, so that's something we'd want to add. In an ideal world can we go beyond 'QA' and automatically merge leaf package updates. This would really help us improve package freshness. It would be a great goal to aim for! Steve / Futurile
