Alad please also see my inline comments to Eli's later reply. I touched on a couple of these package comments in there. I'll just skip those in this email. On Tue, 2021-06-22 at 15:49 +0200, alad via aur-general wrote:
* asterisk
- Services should be managed by the user, not an .install script. - You'll need to assign a GUID for repo usage, cf. https://wiki.archlinux.org/title/DeveloperWiki:UID_/_GID_Database
Yup yup, I'm aware of that change that would need to happen before bringing it to the repos.
* csvkit
- Two missing dependencies. That's a big no-no.
NOT MY FAULT! I don't know what happened here but those packages were in the AUR and even have a long version history. Now they were silently gone. I just ran: ssh aur@aur.archlinux.org restore python-agate-sql ssh aur@aur.archlinux.org restore python-agate-dbf Now they are back, but I can't find any record of their removal. I tried searching the aur-requests list but don't see anything. I found the notification email that I was made co-maintainer on 3/3/20, but nothing about the deletion!
* gitlab-pages
- L33-36 looks pretty hard to audit. Does the -help output change often?
Yes. Every freaking release. They add and deprecate options like the world is ending tomorrow.
If not, consider adding a file to the repository.
Considered, and I'd much rather the scripted approach that keeps it up to date. The alternative is a script in a comment somewhere that I have to rememer to run on every bump outside of the regular routine.
* programmers-dvorak
- Might be worth convincing upstream to have a workable install script.
I've approached the upstream several times over the years and they haven't been responsive. I would review this situation again and consider forking it myself to fix it upstream anyway. I have a derivative that I maintain anyway, programmer-turkish-f, that uses some of the same bits, so it might make sense for me to go ahead and fork the upstream.
* zerobane-studio I can't seem to find this one in AUR?
Another "what the heck" moment here. I am just seeing this has been deleted, but I don't see any notification as to when or why. I would undelete/repost it right now, but it happens to be caught up in the wxLua issues that are ongoing. The upstream half fixed it based in my report but it still doesn't actually compile using current gcc. I'll repost this when I get it working locally but it would be nice to have answers for why it was deleted with no notice.
* zettlr
- Seems very similar to marktext, why add both?
Good question. And the answer is, maybe not. But they are both poppular and have quite a bit of respective momentum. I don't use either for personal use (prefering NeoVIM of course) but I work with several publishing companies and do tech support for others that do. These are both used in the workflow for people I support, hence I also have to be able to run them both. They could both or neither be candidates for [community], that would take some more evaluation.
- I'd include the auxiliary files (.desktop, mime) in the repository. - Vendored pandoc is sad
Yes, very sad and for the record I object. I'm the co-maintainer of this and the other maintainer added that—again. Note it works fine without, it just isn't "upstream's expectation". I would like to remove it again and go back to using the system Pandoc *only*. (as opposed to a UI option for system Pandoc + vendored version. I have suggested that if they want to follow upstream's expectations they should work on a zettlr-bin package instead (which exists anyway) but that hasn't been agreeable for some reason.
I've considered switching to an AAAAAA layout. Seems perfect for Internet discussions.
I just searched to see if that is a real thing. I probably deserved what I (didn't) find. Caleb