| gnu_srs | Seems like we need a version 3.2.9-10~beowulf1 to be able to install eudev on beowulf. Will create that after -10 has been available for ? days. | 10:52 |
|---|---|---|
| gnu_srs | Need help explaining the question-mark above. | 10:52 |
| bb|hcb | gnu_srs: I believe this is also a process with manual approval (like stable proposed updates in Debian). If I am not wrong it should be safe to build it without waiting; its important to push them in order in the repo, not to break the upgrade path | 11:30 |
| LeePen | I can do builds for chimaera and beowulf in a couple of days. | 11:38 |
| bb|hcb | LeePen: Is this the normal process? (just curious) | 12:01 |
| LeePen | Well, only a few folk can trigger builds for -proposed-updates. | 12:08 |
| LeePen | Whereas anybody with write access to git.devuan.org/devuan can trigger builds for unstable. | 12:09 |
| LeePen | So it is managed and controlled in a slightly different way to debian. | 12:10 |
| LeePen | See https://git.devuan.org/devuan/documentation/src/branch/master/maintainers/JenkinsAutobuild.md | 12:12 |
| LeePen | As far as chimaera goes, it is only necessary in this case because the latest unstable build has picked up a dependency on | 12:14 |
| LeePen | libkmod2-udeb (>= 29) which is not satisfiable in chimaera. | 12:15 |
| LeePen | See https://pkgmaster.devuan.org/britney/chimaera/excuses.html | 12:15 |
| LeePen | So, it will need to be rebuilt on chimaera to get the correct dependencies. | 12:16 |
| LeePen | And the same goes for beowulf. | 12:16 |
| bb|hcb | Thanks for the explanation :) | 13:08 |
| bb|hcb | What is the convention for pu - e.g. 3.2.9-10~beowulf1 vs 3.2.9-10~devuan1 is ok by chance for beowulf and chimaera but will break with excalibur; isn't it better to use 3.2.9-9+beowulf1 and backport the fixes from 3.2.9-10 there? | 13:28 |
| gnu_srs | LeePen: So i should update branch beowulf-proposed-updates and push -10~beowulf1 and create a branch for chimaera-proposed-updates and push -10~chinaera1? | 13:35 |
| bb|hcb | gnu_srs: Yes, that is OK because eudev is not a forked package | 13:37 |
| bb|hcb | I didn't explain well above :( - I see a problem with forked packages, where Debian has a ~bpoXX or +debXXuY. e.g. ~bpoXX is > than +devuanX and +devXXuY is > than ~beowulfX | 13:44 |
| bb|hcb | LeePen: now I see that what I describe above will affect upgrades of gpsd: beowulf is 3.22-4~bpo10+1 while ceres/chimaera is 3.22-4+devuan2 (in case beowulf-backports is enabled) | 14:04 |
| fsmithred | bb|hcb, 3.22-4~bpo* is lower version than 3.22-4 or 3.22-4+anything | 14:06 |
| LeePen | gnu_srs: That would be great, thanks. | 15:34 |
| LeePen | bb|hcb: Using gpsd in beowulf as an example, I think it would be 3.22-4~bpo10+1+devuan1. | 15:36 |
| LeePen | It is a bit messy, but works correctly. | 15:36 |
| LeePen | So for forked packages always the debian version with +devuanN. | 15:37 |
| LeePen | For Devuan rebuilds for older suites, the debian version with ~suiteN appended. | 15:38 |
| LeePen | (Ignore the last line and I will rewrite it correctly) | 15:38 |
| LeePen | For Devuan rebuilds for older suites, the *Devuan* version with ~suiteN appanded. | 15:39 |
| bb|hcb | :) Got it; It seems I can't compare without dpkg --compare-versions :( | 16:23 |
| LeePen | bb|hcb: Me neither! ;) | 18:39 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!