Dear KDE maintainers, [tl;dr: I don't think xorg-server and/or mesa caused the regression in kwin]
On 18-06-18 15:25, Paul Gevers wrote: > On 18-06-18 15:17, Paul Gevers wrote: >> Because all but the last source package appear to come from the kde >> suite, I checked the last log⁴. The installation was clean, in the sense >> that no fall-backs occurred and only packages from xorg-server were >> installed from unstable. It appears that xorg-server is introducing >> regressions in the autopkgtest of kwin. How bad is this? > > May this be related to > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=900352 > ? If yes, I'll shut up. I ran a test where I tried to filter out and prevent xorg* and mesa related packages from unstable¹. It still failed. This either means that the KDE stack introduced this regression itself, or one of the still not filtered packages from unstable cause the issue. If your interested, could you check which packages from unstable in that log you would want to have from testing instead to have a fair test? (Right at the top you'll see which packages I specified that should be taken from unstable if they are going to be installed.) In the mean time, I have started creating a more up-to-date set of references for all packages in testing. As a result it may *appear* that there are now more packages in the KDE stack that seem to regress with the latest KDE stack upload (e.g. krunner, kservice and plasma-framework). However, I suspect that that may be because these packages were only tested during the uploading phase, when not all packages were build yet. I have (and will continue to if new regressions appear) retriggered those tests. Thanks for considering. Paul ¹ https://ci.debian.net/data/autopkgtest/testing/amd64/k/kwin/489052/log.gz
signature.asc
Description: OpenPGP digital signature
