<@james:fedora.im>
16:00:38
!startmeeting fpc
<@meetbot:fedora.im>
16:00:40
Meeting started at 2025-10-16 16:00:38 UTC
<@meetbot:fedora.im>
16:00:40
The Meeting name is 'fpc'
<@james:fedora.im>
16:00:44
!topic Roll Call
<@limb:fedora.im>
16:00:50
!hi
<@zodbot:fedora.im>
16:00:53
Gwyn Ciesla (limb) - she / her / hers
<@churchyard:fedora.im>
16:01:22
!hi
<@james:fedora.im>
16:01:22
Hey
<@zodbot:fedora.im>
16:01:22
Miro Hrončok (churchyard) - he / him / his or they / them / theirs
<@churchyard:fedora.im>
16:01:27
I am around, reading the Go guidelines
<@james:fedora.im>
16:02:19
You thinking about changing them, or just looking things up as you do something to a Go package?
<@churchyard:fedora.im>
16:03:31
I mean the proposed ones by gotmax23 https://pagure.io/packaging-committee/pull-request/1498
<@james:fedora.im>
16:04:54
Ahh, fair enough. I'll tag that with meeting.
<@salimma:fedora.im>
16:09:24
!hi
<@zodbot:fedora.im>
16:09:25
Michel Lind (salimma) - he / him / his
<@decathorpe:fedora.im>
16:13:02
!hi
<@zodbot:fedora.im>
16:13:03
Fabio Valentini (decathorpe) - he / him / his
<@decathorpe:fedora.im>
16:13:15
sorry just had to run to catch a bus
<@james:fedora.im>
16:13:25
No problem. That makes 5 of us.
<@james:fedora.im>
16:13:35
In theory we can do all the things now ;)
<@limb:fedora.im>
16:14:28
Wow. No pressure.
<@salimma:fedora.im>
16:14:54
RFC: Year of the Linux Desktop
<@james:fedora.im>
16:15:25
!topic FPC PR#1497 https://pagure.io/packaging-committee/issue/1497
<@salimma:fedora.im>
16:16:06
huh 404
<@james:fedora.im>
16:16:16
A PR by Stephen Gallagher would be nice ;) ... but I think after the question/answers everyone is on board with the change?
<@decathorpe:fedora.im>
16:16:25
private ticket ...
<@james:fedora.im>
16:16:49
Hmm ... and not everyone can see it? I'm not sure how to fix that.
<@decathorpe:fedora.im>
16:16:50
"on board" well ... it's annoying as hell but sure I guess this needs to be done
<@james:fedora.im>
16:16:59
I'm guessing a fas change ... let me look.
<@decathorpe:fedora.im>
16:17:08
Michel Lind UTC+1 is not a group member IIRC?
<@salimma:fedora.im>
16:17:23
I joined last and never had to see a private ticket so it's possible
<@sgallagh:fedora.im>
16:17:41
James: I’ll send a PR once it’s approved, yes.
<@limb:fedora.im>
16:18:47
I'd rather see the PR first as is customary but I'm +1 generally and won't block on the PR
<@decathorpe:fedora.im>
16:20:37
same here
<@james:fedora.im>
16:21:39
Ahh, I see: https://github.com/praiskup/norpm/blob/main/rawhide-tests/expected.yaml
<@james:fedora.im>
16:22:20
Ahh, ruby, never change.
<@churchyard:fedora.im>
16:23:16
I would like to see the guideline proposed. If it needs to be in a private ticket, no need for a PR yet. But the devil is in the details. What *exactly* is supposed to be allowed and what not
<@decathorpe:fedora.im>
16:25:52
yeah. at least some common patterns like the gsub macro really need to be supported (and apparently already are?), but this needs to be documented
<@churchyard:fedora.im>
16:27:03
and the forgemeta thing -- not allowing that would explode in our faces
<@decathorpe:fedora.im>
16:27:41
that too ... so that affects both go and fonts packages?
<@decathorpe:fedora.im>
16:28:06
*and* all packages that use the forge macros outside of those use cases I guess
<@sgallagh:fedora.im>
16:28:11
I’m working with Pavel to figure these exceptional cases out.
<@sgallagh:fedora.im>
16:28:32
We will put together a PR and send it. I don’t think that part needs to be in private.
<@gotmax:matrix.org>
16:28:46
What's the issue with the forge macros now?
<@james:fedora.im>
16:29:09
Michel Lind UTC+1: Okay, You've been added to the FAS group `packaging-committee` now. Might need a few minutes for pagure to sync.
<@gotmax:matrix.org>
16:29:30
FAS Groups don't sync over to Pagure.io by the way
<@decathorpe:fedora.im>
16:29:30
I don't think that syncs? the pagure.io project has a separate list of members.
<@james:fedora.im>
16:29:44
Ahh, good times.
<@james:fedora.im>
16:29:55
Manual syncing is fun.
<@churchyard:fedora.im>
16:30:19
they might or might not be parseable by https://github.com/praiskup/norpm
<@salimma:fedora.im>
16:30:54
yeah logging out and logging back in and refreshing does not get me access
<@gotmax:matrix.org>
16:30:57
They probably won't, unless someone wants to reimplement them in Python...
<@sgallagh:fedora.im>
16:33:40
But I understand the convenience factor.
<@james:fedora.im>
16:34:35
Michel Lind UTC+1: Okay, you are in the pagure group now too.
<@james:fedora.im>
16:37:06
I always had mixed feelings ... it doesn't seem enough on it's own, but I was hoping someone would do a thing where you could throw a patch at a package and then everything including the version/build would automatically get done and then you could test a package.
<@james:fedora.im>
16:37:17
I know ... I have weird dreams ;)
<@sgallagh:fedora.im>
16:38:08
The *version* field I specifically feel like should have some intentionality to it.
<@sgallagh:fedora.im>
16:38:29
In general, I wouldn't want that to change unexpectedly, and having to set it manually helps that.
<@sgallagh:fedora.im>
16:38:37
But that's not really relevant to the topic
<@salimma:fedora.im>
16:38:43
Thanks. I am pulled into another meeting but will check later
<@gotmax:matrix.org>
16:39:41
The latest change in the forge macros was to make it possible to generate Version because the Guidelines want git snapshot information to be included in Version, not magically added to Release like they were before
<@james:fedora.im>
16:39:55
Okay ... I'm going to assume Michel, or anyone else not here, won't randomly hate the idea in new ways. So if Stephen Gallagher wants to start a PR, we can probably merge it fairly soon.
<@conan_kudo:matrix.org>
16:40:15
!hi
<@zodbot:fedora.im>
16:40:17
Neal Gompa (ngompa) - he / him / his
<@conan_kudo:matrix.org>
16:40:43
I shouldn't be here, because I'm feeling half-dead, but I think this new restriction is unworkable
<@james:fedora.im>
16:41:13
haha ... just in time ;)
<@conan_kudo:matrix.org>
16:42:06
tbh, I've always thought rpmautospec was badly done, and I think the fact that Stephen Gallagher is thinking of even new restrictions on spec parsing for it reinforces it
<@james:fedora.im>
16:42:20
Conan Kudo 😷🤒🤧: Do you think it needs a few more exceptions added ... or it's just not workable?
<@gotmax:matrix.org>
16:42:23
But requiring specfiles to be parse-able by some tool other than RPM that doesn't support all of its constructs (if that's what's being discussed) seems backwards
<@conan_kudo:matrix.org>
16:42:51
we should go the other way: instead of attempting to parse anything, just erase the Release field and write it again
<@sgallagh:fedora.im>
16:43:02
gotmax23: NoRPM is a stopgap measure; RPM upstream is also considering incorporating the same functionality. It'll just take years to land.
<@james:fedora.im>
16:43:52
Stephen Gallagher: NoRPM = https://pypi.org/project/norpm/1.0/ ?
<@sgallagh:fedora.im>
16:43:56
Conan Kudo 😷🤒🤧: It is parseable so that we can figure out how long ago it last changed, in order to compute the release number
<@conan_kudo:matrix.org>
16:44:05
because you're using the wrong input
<@conan_kudo:matrix.org>
16:44:17
don't ask the spec file, ask the build system
<@conan_kudo:matrix.org>
16:44:31
the build system has all that data in a fully safe form
<@sgallagh:fedora.im>
16:45:10
James: Yes, that's the one
<@churchyard:fedora.im>
16:45:41
Conan Kudo 😷🤒🤧: I don't think this is helpful. I know your opinion on rpmautospec.
<@decathorpe:fedora.im>
16:45:47
you wouldn't even need to ask, that number would just be another input for the build
<@conan_kudo:matrix.org>
16:45:53
yeah
<@conan_kudo:matrix.org>
16:46:30
we could even make it simpler and just add a %buildrelease macro to the disttag so that if the value is set, it will expand
<@sgallagh:fedora.im>
16:46:42
Conan Kudo 😷🤒🤧: That only works if every build of that package always goes through the same central system. It doesn't account for local builds, mock or COPR.
<@churchyard:fedora.im>
16:46:47
<@churchyard:fedora.im>
16:46:47
> The requested change is "executable code in the Epoch, Version, or Release is forbidden when %autorelease is in use".
<@churchyard:fedora.im>
16:46:47
For reference for those who do not see the private ticket:
<@conan_kudo:matrix.org>
16:47:22
COPR is already broken with the existing system
<@gotmax:matrix.org>
16:48:01
I assume "executable code" includes both `%(...)` shell and `%{lua:...}` expansions?
<@churchyard:fedora.im>
16:48:04
I also noticed now that the proposed rule is contradicting itself as %autorelease is in use in Release and is executable code :D
<@conan_kudo:matrix.org>
16:48:10
yup
<@churchyard:fedora.im>
16:48:34
pretty much. but standard macros implementedin Lua (think %gsub) should be allowed
<@conan_kudo:matrix.org>
16:48:58
but they can't be if rpmautospec uses norpm
<@gotmax:matrix.org>
16:49:02
I think gsub is a builtin macro that calls lua in C
<@conan_kudo:matrix.org>
16:49:08
it is
<@sgallagh:fedora.im>
16:49:15
Conan Kudo 😷🤒🤧: That statement is false.
<@churchyard:fedora.im>
16:49:23
Conan Kudo 😷🤒🤧: norpm reimplements gsub in python
<@sgallagh:fedora.im>
16:49:31
As mhroncok says.
<@sgallagh:fedora.im>
16:49:50
NoRPM will reimplement some common ones to make this transition easier.
<@churchyard:fedora.im>
16:49:57
but my concern in the ticket was that I don't want a rule that replies on the implementation details
<@gotmax:matrix.org>
16:49:59
And also it's not supported in some RHELs so packages that need to do common string operations like gsub in a backward compatible way need bash or lua expansions...
<@sgallagh:fedora.im>
16:50:13
Because some (like gsub) don't need to invoke a complete language
<@sgallagh:fedora.im>
16:50:27
gotmax23: What's not supported in RHEL?
<@decathorpe:fedora.im>
16:50:38
%gsub is fairly new
<@churchyard:fedora.im>
16:50:39
IIRC I cannot use gsub as even Zuul is too old to understand it, so I need to use Lua or Bash
<@gotmax:matrix.org>
16:50:41
And is there a reason for requiring this? Why is it tied together with using rpmautospec?
<@gotmax:matrix.org>
16:50:55
gsub and the lua wrappers added along with it
<@gotmax:matrix.org>
16:51:02
%gsub and the lua wrappers added along with it
<@sgallagh:fedora.im>
16:51:06
gotmax23: That's in the private ticket. I can't say here.
<@churchyard:fedora.im>
16:51:10
gotmax23: that's kidna why the ticket is private
<@decathorpe:fedora.im>
16:51:47
well it's kind of silly that this is a private issue but oh well
<@gotmax:matrix.org>
16:51:48
Development discussions like this shouldn't be private
<@gotmax:matrix.org>
16:51:57
This isn't RHEL
<@churchyard:fedora.im>
16:52:09
there is no conspiracy here, the privacy is justified
<@sgallagh:fedora.im>
16:52:33
gotmax23: I assume you can't see the ticket, but think about why I (an outspoken member of the "everything should be public" crew) would have made it private.
<@gotmax:matrix.org>
16:52:44
I didn't mean to imply a conspiracy
<@decathorpe:fedora.im>
16:52:57
to be honest the more I learn about this the less I like the proposed solution. let's just change rpmautospec to not parse spec files. and let's throw out epoch:version-release tags from changelogs too while we're at it ...
<@churchyard:fedora.im>
16:53:00
I am not saying it MUST have been private, but I understand the resoning behind it
<@james:fedora.im>
16:53:30
Is there any way to let gotmax see it without adding him to the FPC group?
<@decathorpe:fedora.im>
16:53:45
not after the fact I think
<@sgallagh:fedora.im>
16:53:53
James: I can reach out to him via a secure message. I trust him.
<@gotmax:matrix.org>
16:54:00
I would guess something is parsing untrusted specifles in a trusted environment but don't really want to speculate
<@decathorpe:fedora.im>
16:54:47
is it time for the "neither confirm nor deny" joke?
<@james:fedora.im>
16:56:03
Somewhat related ... I saw these recently: https://www.amazon.com/gp/product/B0FC9VFS49/
<@conan_kudo:matrix.org>
16:56:27
we could actually drop EVR tags from changelog entries and require people to specify version number in the text in version bumps
<@james:fedora.im>
16:56:30
But not sure I have enough spare money to buy enough for everyone ;)
<@conan_kudo:matrix.org>
16:56:37
technically, it's already accepted, just not mandatory
<@sgallagh:fedora.im>
16:57:29
Conan Kudo 😷🤒🤧: I'm in favor of eliminating EVR from changelog entries, FWIW.
<@sgallagh:fedora.im>
16:57:50
(I'm also in favor of eliminating them entirely, since git history exists, but...)
<@conan_kudo:matrix.org>
16:58:01
yeah no, not that
<@churchyard:fedora.im>
17:00:25
I am in favor of eliminating computers
<@churchyard:fedora.im>
17:00:34
it was a mistake
<@james:fedora.im>
17:00:48
Maybe AI will do that for us
<@sgallagh:fedora.im>
17:00:51
mhroncok++
<@zodbot:fedora.im>
17:00:53
sgallagh gave a cookie to churchyard. They now have 179 cookies, 12 of which were obtained in the Fedora 42 release cycle
<@limb:fedora.im>
17:01:07
You could make the same case for agricultural societies.
<@churchyard:fedora.im>
17:01:11
ok, let's move on for now?
<@james:fedora.im>
17:02:10
But FWIW I also find rpm changelog to be an artifact from 1995 that would be very different if done now, and maybe the best thing to do is make it go away and let someone tie in something better.
<@james:fedora.im>
17:02:34
Well, technically we are 2m over time. So we can all go home.
<@james:fedora.im>
17:02:43
!topic Open Floor
<@james:fedora.im>
17:03:07
Anything we need to speak about before the day before halloween?
<@churchyard:fedora.im>
17:03:18
gotmax23: I tried to provide some input tot he go guidelines proposal. note that it is very opinionated. happy to discuss it synchronously as well
<@gotmax:matrix.org>
17:04:07
Thank you! I'll try to take another look after I eat lunch :)
<@james:fedora.im>
17:04:07
If anyone has time ... can you take a quick look at the R PR: https://pagure.io/packaging-committee/pull-request/1494
<@james:fedora.im>
17:04:17
Hopefully we can get that merged.
<@james:fedora.im>
17:05:16
Seems simple, so unless anyone objects this week I'll probably hit merge tomorrow.
<@james:fedora.im>
17:06:24
Saying that I know basically nothing about R, so if anyone knows anything please take a look.
<@churchyard:fedora.im>
17:06:48
the only thing I know about R is that it's hard to google it
<@churchyard:fedora.im>
17:07:14
which is even not true anymore :D
<@james:fedora.im>
17:08:35
On that note...
<@james:fedora.im>
17:08:37
!endmeeting