ferringb

Any reason not to use sources or PPA? ...he says, having entirely forgot about that until seeing your PR. ;-)

I checked into the PPA/backport angle; TL;DR: that's a very painful idea due to GH's modified 24.04 that is 'latest'. Anyone have any major objections to if we were to...

This is still reproducible w/ that SHA in pkgcheck 0.10.37. The odd part is the feed going into that function *is* stable, as is the vars in that code path,...

Just confirming: is it intentional that if `--cache=no` is passed, the eclass check itself just doesn't run? That seems like a bug.

@TheK0tYaRa can you share the ebuilld, or at least the `pkg_preinst`? As @ulm said, this really should be in `src_install` if possible. Also, if that flies with portage that'd be...

@ulm you'll like this one... and I'm ripping my hair out that this even exists since it's triggering some flashbacks... Repro is below, but the take away is dosym is...

@TheK0tYaRa thanks for the link and confirmation. > Ulm > IIUC Portage doesn't make any distinction between phases but simply exposes all commands. Yeah, this confirms it. ``` ferringb@hrair /var/db/repos/ferringb/dev-util/testing...

> Maybe it is easier to restrict things on the QA level, e.g. warn about do* before src_install? That still breaks the users of any PM (well, pkgcore) implementing the...

Fair warning; pkgcore's now primarily being used as a data API layer, and repository manipulation API. Pquery (pull ebuild metadata), pkgcheck (QA), pkgdev (QA + commit/manifest manipulation). I believe @radhermit...

Was pkgcore installed via portage, or manually installed? EPREFIX is *only* relevant in an ebuild context- it tells the manager where to shove shit and prefix it accordingly. Pkgcore uses...