<@james:fedora.im>
16:00:50
!startmeeting fpc
<@james:fedora.im>
16:00:56
!topic Roll Call
<@meetbot:fedora.im>
16:00:59
Meeting started at 2024-09-12 16:00:50 UTC
<@meetbot:fedora.im>
16:00:59
The Meeting name is 'fpc'
<@tibbs:fedora.im>
16:01:16
Hey folks.
<@limb:fedora.im>
16:01:18
!hi
<@zodbot:fedora.im>
16:01:19
Gwyn Ciesla (limb) - she / her / hers
<@james:fedora.im>
16:01:30
!hi
<@zodbot:fedora.im>
16:01:35
James Antill (james)
<@tibbs:fedora.im>
16:01:52
In addition to having my house torn up, I am also sick.
<@james:fedora.im>
16:02:36
The joys of getting older ... :(
<@tibbs:fedora.im>
16:03:04
I think it's just a cold but it's tough to be sure.
<@james:fedora.im>
16:03:44
Been hit with outside allergy stuff the last month or so, which is unusual.
<@decathorpe:fedora.im>
16:05:36
!hi
<@zodbot:fedora.im>
16:05:37
Fabio Valentini (decathorpe) - he / him / his
<@decathorpe:fedora.im>
16:05:54
hello o/
<@carlwgeorge:matrix.org>
16:06:45
!hi
<@zodbot:fedora.im>
16:06:47
Carl George (carlwgeorge) - he / him / his
<@james:fedora.im>
16:08:49
Okay, nothing new ... but trying to get some closure on old things, so...
<@james:fedora.im>
16:09:00
!topic FPC PR#1014 https://pagure.io/packaging-committee/pull-request/1014
<@james:fedora.im>
16:09:47
Fabio Valentini 👀
<@tibbs:fedora.im>
16:11:56
I don't know enough about go, but with old things like this I have to wonder if it's still necessary.
<@james:fedora.im>
16:12:39
Yeh, it can't merge due to conflicts ... but I'm not sure that means it's fixed.
<@decathorpe:fedora.im>
16:13:13
hum hum
<@tibbs:fedora.im>
16:13:18
Conflicts probably means little more than something getting line breaks
<@decathorpe:fedora.im>
16:13:30
I don't think this change is even correct
<@decathorpe:fedora.im>
16:13:44
probably we should ping gotmax23 who now maintains the Go macros.
<@james:fedora.im>
16:14:32
Fabio Valentini: I'm pretty happy to follow your lead here. If you think we should just close it I'll do it, obviously pinging anyone is fine ... just rather it not remain open for another 2 years :)
<@carlwgeorge:matrix.org>
16:14:42
with no reply from the submitter to the comments so far for four years, i think we should just close it. someone that wants to address it can open a new pr.
<@decathorpe:fedora.im>
16:14:46
yeah
<@decathorpe:fedora.im>
16:15:17
it's a problem that should still be looked at (go compiling C stuff without inheriting proper flags), but I don't think it's as easy as setting CGO_FLAGS to CFLAGS
<@decathorpe:fedora.im>
16:15:41
probably some mechanism similar to extension_cflags that's used by Python will be needed, IIUC
<@james:fedora.im>
16:16:09
Makes sense.
<@james:fedora.im>
16:16:16
I closed it.
<@james:fedora.im>
16:16:28
!topic FPC PR#947 https://pagure.io/packaging-committee/pull-request/947
<@decathorpe:fedora.im>
16:17:12
hum, this sounds mostly like a no-brainer to me.
<@james:fedora.im>
16:17:13
I think we just close this and have someone open a new PR if there's anything that still needs changing ... but wanted to make sure.
<@decathorpe:fedora.im>
16:18:31
"Therefore any occurence of the old macros (...) MUST be replaced" is a bit strong for something that just deprecates things, but I agree with the spirit of the changes
<@james:fedora.im>
16:18:37
Some things were done, AIUI ... was the warning added?
<@james:fedora.im>
16:19:31
I don't see it at: https://docs.fedoraproject.org/en-US/packaging-guidelines/AutoProvidesAndRequiresFiltering/
<@decathorpe:fedora.im>
16:19:59
hm. yeah it looks like the docs for the old-style generators / filtering were just dropped
<@decathorpe:fedora.im>
16:20:05
so no warning necessary?
<@tibbs:fedora.im>
16:20:34
I'm surprised the old stuff even works at all.
<@james:fedora.im>
16:20:59
Yeh, I tried searching and we just refer to them and don't talk about them at all.
<@james:fedora.im>
16:21:16
I'll just close this as done.
<@tibbs:fedora.im>
16:22:17
It would be possible to redefine the underlying macros to warn or error out, I guess, but I think eventually RPM will just make it "not work".
<@james:fedora.im>
16:22:26
!topic FPC#1053 https://pagure.io/packaging-committee/issue/1053
<@james:fedora.im>
16:22:54
Yeh, I'm happy to do nothing and let people blame rpm ;)
<@james:fedora.im>
16:25:37
I don't know much about this ... but if things still haven't changed maybe we just close this and wait for a PR?
<@decathorpe:fedora.im>
16:26:31
I don't think anything about the situation has changed.
<@tibbs:fedora.im>
16:27:07
I still like my last suggestion (error and eventually fail the build if there's something to check but the validator wasn't installed).
<@decathorpe:fedora.im>
16:27:30
- the tool for validation is based on the library that's exclusively used for generating repository metadata
<@decathorpe:fedora.im>
16:27:30
- the tool that is *not* used for validation is based on the library that's now used by *all* software centers
<@decathorpe:fedora.im>
16:27:49
tibbs: yeah that's an interesting solution
<@decathorpe:fedora.im>
16:28:09
that would be a BRP script?
<@decathorpe:fedora.im>
16:28:30
I can file a Change Proposal to formalize this idea
<@tibbs:fedora.im>
16:28:32
Yes, I think it's entirely doable.
<@james:fedora.im>
16:28:40
But someone mentioned that the validator that is used is stricter than the one proposed (and reading the data) ... so it's all fine?
<@tibbs:fedora.im>
16:28:46
Assuming that you can easily determine that there's a file which needs validation.
<@decathorpe:fedora.im>
16:29:05
they're both doing some checks that the other one doesn't. so it's not a strict subset relationship
<@james:fedora.im>
16:29:14
ahh, of course.
<@decathorpe:fedora.im>
16:29:50
as far as I can tell, yes, it's just files in `%{_metainfodir}` and the old location (`%{_datadir}/appdata`)
<@james:fedora.im>
16:33:45
I added a comment about the BRP change proposal.
<@tibbs:fedora.im>
16:34:01
Can't promise anything in my current state but I will try to look over the existing BRP scripts and see whether there is anything simple that can be done. Stuff can always be added later.
<@tibbs:fedora.im>
16:34:44
One nice thing about doing it that way is that you can largely isolate packagers from having to care about which tool actually does the validation.
<@james:fedora.im>
16:35:04
!topic FPC#1049 https://pagure.io/packaging-committee/issue/1049
<@james:fedora.im>
16:35:29
Last old one for this week ... one of the many things I don't know much about.
<@decathorpe:fedora.im>
16:36:38
oh, another one I filed
<@decathorpe:fedora.im>
16:36:43
:sad face:
<@decathorpe:fedora.im>
16:37:12
I think we just need to write some basic guidelines. GIR belongs in -devel subpackage, .vapi too, .typelib belongs in main library package, etc.
<@decathorpe:fedora.im>
16:37:18
I can try to work on this.
<@carlwgeorge:matrix.org>
16:37:42
i ran into a similar thing this week, there was an rpmlint error in review about an explicit requires on libadwaita, which has the gi typelib file.
<@carlwgeorge:matrix.org>
16:38:07
https://bugzilla.redhat.com/show_bug.cgi?id=2311833#c2
<@decathorpe:fedora.im>
16:38:11
and it annoyed rpmlint because the package name has prefix "lib*" supposedly?
<@decathorpe:fedora.im>
16:38:37
we might also want to add dependency generators for typelib / gir / vapi stuff too while we're at it
<@decathorpe:fedora.im>
16:38:41
other distros have it
<@carlwgeorge:matrix.org>
16:38:41
i'm assuming so, because the same thing was true for the dep on gtk4, but it didn't complain about that one
<@decathorpe:fedora.im>
16:39:07
yeah I think rpmlint is stupid and just matches against (lib.*|.*-libs)
<@decathorpe:fedora.im>
16:39:16
yeah I think rpmlint is stupid and just matches against `(lib.*|.*-libs)`
<@james:fedora.im>
16:39:28
Can we copy and paste things?
<@carlwgeorge:matrix.org>
16:39:39
indeed i think a `Requires: typelib(Adw)` would have solved the problem here, if that was provided by libadwaita
<@decathorpe:fedora.im>
16:39:58
probably copy-paste would work - Conan Kudo knows this better than I do
<@decathorpe:fedora.im>
16:40:06
I'll put it on my list! :D
<@carlwgeorge:matrix.org>
16:40:11
even better if generated and not explicit
<@decathorpe:fedora.im>
16:41:08
generating Provides should be easy, I think. Requires, not so much, you'd need to parse Python code for `from gi.repository import Gtk_4.0` or something like that
<@carlwgeorge:matrix.org>
16:41:42
yeah was just about to say the generated dep might not be feasible, but the generated provides should work
<@james:fedora.im>
16:42:00
Mmmm, sweet sweet regexps ;)
<@james:fedora.im>
16:43:02
!topic Open Floor
<@james:fedora.im>
16:43:19
Okay, not trying to do all the old things at once ... but is there anything anyone wants to talk about?
<@decathorpe:fedora.im>
16:44:22
hey, ticket <del>archaeology</del> triage is fun
<@decathorpe:fedora.im>
16:45:49
I have just today had some ideas for improving the Versioning guidelines, but I don't want to step on anybody's toes since it's already in flux
<@carlwgeorge:matrix.org>
16:46:46
ping me about it after the meeting and i'll let you know if it's something i'm already incorporating, or if it makes sense to
<@carlwgeorge:matrix.org>
16:47:13
i started with the snapshot part, but it quickly grew to a total refresh as i couldn't unsee phrasing that didn't make sense
<@decathorpe:fedora.im>
16:47:34
(TL;DR is incorporating %autorelease usage, which is not documented yet, and adding more "do this in this case" style examples)
<@carlwgeorge:matrix.org>
16:48:46
%autorelease is already covered
<@carlwgeorge:matrix.org>
16:48:50
https://docs.fedoraproject.org/en-US/packaging-guidelines/Versioning/#_release_tag
<@decathorpe:fedora.im>
16:49:05
it is? hum, I must have missed that. I had an old TODO item for this
<@decathorpe:fedora.im>
16:49:38
it's just basic info, no examples for snapshot versioning etc. though
<@carlwgeorge:matrix.org>
16:51:06
i think that will only be needed for the traditional versioning style (which i'd like to move to an appendix)
<@decathorpe:fedora.im>
16:51:20
true!
<@carlwgeorge:matrix.org>
16:51:58
with the preferred snapshot info in the version, %autorelease is dead simple. but you're right that examples of the advanced usage should exist somewhere in the guidelines.
<@conan_kudo:matrix.org>
16:53:05
!hi
<@decathorpe:fedora.im>
16:53:05
yeah, that situation seems to be what throws off most people
<@zodbot:fedora.im>
16:53:07
Neal Gompa (ngompa) - he / him / his
<@james:fedora.im>
16:55:40
Well, now Neal is here I guess we can end the meeting ;)
<@conan_kudo:matrix.org>
16:57:28
we can and should consider pulling in the openSUSE generator into Fedora for this
<@conan_kudo:matrix.org>
16:57:38
it's kind of silly that we haven't had this for 12+ years
<@carlwgeorge:matrix.org>
16:57:59
works for me, are you volunteering? 😀
<@conan_kudo:matrix.org>
16:58:04
sure why not
<@zodbot:fedora.im>
16:59:58
neil has already given cookies to ngompa during the F40 timeframe
<@zodbot:fedora.im>
17:00:02
james gave a cookie to ngompa. They now have 142 cookies, 29 of which were obtained in the Fedora 40 release cycle
<@james:fedora.im>
17:00:40
Anything else anybody wants to volunteer Neal for?
<@james:fedora.im>
17:01:11
Technically we are out of time, so I should close anyway ... but I think the next slot is free if we want to talk an extra couple of minutes.
<@nhanlon:beeper.com>
17:01:51
i'd love to be a fly on the wall and provide support if ya need
<@james:fedora.im>
17:06:03
Okay, thanks for coming everyone
<@james:fedora.im>
17:06:05
!endmeeting