Hi there On Sat, 17 Aug 2024, at 5:45 PM, Daniel Gröber wrote: > Hi Dave, > > On Sat, Aug 17, 2024 at 05:17:58PM +0100, Dave Hibberd wrote: >> Is there any progress or another set of hands required on this? > > As per usual I've gotten side-tracker by other things, so I haven't gotten > started on this yet :]
I know that all too well - I need this for a fork of a forked project. I must be 20 forks down and somehow this is more interesting than just setting up mailman 3 (or discourse!). > Feel free to jump in. Next steps are to reconcile Michael's changes/git > history with Chris' repo. Either by just overwriting files and > re-committing or by re-using the history and dropping the irellevant tags. I stuck my nose in this tonight here [1]. It's not ready for others and comes with some caveats and notes. For my own understanding and perhaps clarity for others, I have started a fresh repo, cherrypicking from both repos (they are referenced here as branches and I'm just checking out the bits I want onto my debian/latest branch). This repo is in DEP-14[2] format as I like how things are laid out and how the branch names read. It largely reflects my brain and will help me bugfix. Once things are building and we've got some consensus, I am happy to reformat all of this work into something that can be merged atop what already exists in the mlmmj repo. I'm not particularly religious about the branch names etc. * Actions taken * Since this discussion occurred, there has been another point release (1.4.7), so I started with that as my base. I have extensively rewritten d/copyright into what I feel is a more legible, user-informative format from the git commits [3]. Hopefully this helps - does it line up with what you've done MIchael? The rest is just me playing with computers! * Outcomes * So far, I have not been able to make this 1.4.7 toy *or* mjeanson's 1.4.6 & 1.4.5 releases build in my `sbuild` environment. Michael - how are you building this? I did a quick dirty backport to bookworm and it still failed for me. It fails due to failing tests [4] - this is true of 1.4.7-1 too, and disconcertingly there are different tests failing on salsa [5]. I think we need to work out *if* there is an issue here and if there is a regression between sid and bookworm. Trying to working around that, it fails if I I deploy the nocheck profile [6] - see line 1511. Package can't find, libatf-dev, which tracks - that's a <!nocheck> dependency and we're building without it. This will need to be fixed if we take the dh13 packaging approach and is probably worth reporting upstream as a bug if someone else can verify it. Interesting aside - these tests took minutes on my arm macbook, pegging one core at 100% but seconds on x86! I wonder what's going on here. Once things build I might have a play on a porterbox to dig into it. * Next steps* I think I'll come at this from a different angle and try to stimulate a build of 1.4 series using Chris' foundations. I've successfully built 1.3 tonight, somewhat unsurprisingly! While I like Debhelper and what I feel is the clarity it brings, i like software I can build more! > > Since Chris still hasn't added us to the mlmmj-team as requested I'd > suggest just pushing this to your personal namespace for now and adding me, > Michael and Chris (@dxld @mjeanson @krait) there so we can work together on > this more easily. Once i've got a better understanding of a) why things aren't building and b) If I can be any help, I'll get that done! > > Feel free to ping me on OFTC (dxld) if you need any help. > Noted, you'll find me as hibby. [1] https://salsa.debian.org/hibby/hibby-mlmmj [2] https://dep-team.pages.debian.net/deps/dep14/ [3] https://salsa.debian.org/hibby/hibby-mlmmj/-/raw/debian/latest/debian/changelog [4] https://paste.debian.net/1327045/ [5] https://salsa.debian.org/hibby/hibby-mlmmj/-/jobs/6156232 [6] https://paste.debian.net/1327047/ -- Hibby Debian Developer Packet Radioist MM0RFN
