<@james:fedora.im>
16:01:25
!startmeeting fpc
<@meetbot:fedora.im>
16:01:27
Meeting started at 2024-10-31 16:01:25 UTC
<@meetbot:fedora.im>
16:01:27
The Meeting name is 'fpc'
<@james:fedora.im>
16:01:29
!topic Roll Call
<@limb:fedora.im>
16:01:48
!hi
<@zodbot:fedora.im>
16:01:49
Gwyn Ciesla (limb) - she / her / hers
<@james:fedora.im>
16:02:03
!hi
<@zodbot:fedora.im>
16:02:04
James Antill (james)
<@james:fedora.im>
16:02:31
Gwyn Ciesla (she/her): Ready for the joy of DST this wekend?
<@limb:fedora.im>
16:03:46
OMG I'M SO EXCITED I
<@limb:fedora.im>
16:03:49
no.
<@carlwgeorge:matrix.org>
16:06:39
!hi
<@zodbot:fedora.im>
16:06:41
Carl George (carlwgeorge) - he / him / his
<@salimma:fedora.im>
16:08:39
!hi
<@zodbot:fedora.im>
16:08:41
Michel Lind (salimma) - he / him / his
<@salimma:fedora.im>
16:08:55
eh it's not as bad as losing one hour, but also not a fan
<@conan_kudo:matrix.org>
16:09:16
!hi
<@zodbot:fedora.im>
16:09:17
Neal Gompa (ngompa) - he / him / his
<@james:fedora.im>
16:09:59
Is that because you dread all change? ;)
<@conan_kudo:matrix.org>
16:10:08
lol no
<@salimma:fedora.im>
16:11:11
just non-Wayland changes
<@conan_kudo:matrix.org>
16:11:14
it's because I already went through one time change while I was in the netherlands earlier this week
<@conan_kudo:matrix.org>
16:11:22
I hate experiencing it twice
<@salimma:fedora.im>
16:11:24
ugh yeah Neal got to hit both
<@james:fedora.im>
16:13:40
Ahh, yeh, that does suck. I did that one year, and don't recommend double jetlagg plus double DST change.
<@conan_kudo:matrix.org>
16:13:58
yep
<@james:fedora.im>
16:15:03
!topic Open Floor
<@james:fedora.im>
16:15:16
I think we can just close this one: https://pagure.io/packaging-committee/issue/1307
<@james:fedora.im>
16:20:15
This seems like someone could do a minor PR and we could close it out: https://pagure.io/packaging-committee/issue/1159
<@james:fedora.im>
16:20:42
Title now is actually more like: Have policy about using configure in %conf.
<@salimma:fedora.im>
16:23:03
should it extend to autoreconf and similar tools too?
<@conan_kudo:matrix.org>
16:23:16
yes
<@conan_kudo:matrix.org>
16:23:41
we should also probably tell people to move `%cmake` and `%meson` to `%conf` as well
<@james:fedora.im>
16:23:50
I updated the title, and just mentioned configure, but yeh
<@carlwgeorge:matrix.org>
16:24:01
does anyone know offhand how far back `%conf` works in epel rpm versions?
<@james:fedora.im>
16:24:26
I think no for el7, but yes for anything live now.
<@carlwgeorge:matrix.org>
16:25:49
yeah easy if all branches for sure work with it, but i'm suspicious that's the case
<@carlwgeorge:matrix.org>
16:26:19
even if it doesn't work all the way back to epel8, we can account for that by making the guideline a SHOULD
<@conan_kudo:matrix.org>
16:26:34
conf section was introduced in RPM 4.18
<@conan_kudo:matrix.org>
16:26:36
so RHEL 9+
<@carlwgeorge:matrix.org>
16:26:49
there you go
<@conan_kudo:matrix.org>
16:26:57
oh well, actually no
<@conan_kudo:matrix.org>
16:26:59
RHEL 10+
<@conan_kudo:matrix.org>
16:27:02
RHEL 8 is RPM 4.16
<@conan_kudo:matrix.org>
16:27:06
RHEL 8 is RPM 4.14
<@conan_kudo:matrix.org>
16:27:09
RHEL 9 is RPM 4.16
<@james:fedora.im>
16:27:35
wow
<@conan_kudo:matrix.org>
16:28:01
unless of course %conf gets backported
<@james:fedora.im>
16:28:18
I would have bet a decent amount that epel9 would be new enough.
<@conan_kudo:matrix.org>
16:28:46
nope
<@conan_kudo:matrix.org>
16:28:57
RPM 4.18 barely missed it, it was released Sept 2022
<@conan_kudo:matrix.org>
16:29:14
just like we missed RPM 4.20 and declarative builds for RHEL 10
<@conan_kudo:matrix.org>
16:29:29
it really sucks that we get held back by stuff like this
<@conan_kudo:matrix.org>
16:29:39
and then the rpm developers get upset because we don't use the features :(
<@conan_kudo:matrix.org>
16:29:48
and it's a vicious cycle of awfulness
<@carlwgeorge:matrix.org>
16:30:45
that's unnecessary exaggeration, this is just how rhel works
<@conan_kudo:matrix.org>
16:31:02
no, it's not unnecessary, because the rpm developers have complained about this multiple times
<@james:fedora.im>
16:31:03
Oh, well, I guess we wait another few years then
<@conan_kudo:matrix.org>
16:31:20
they complained with tilde versioning, they complained with rich deps, etc.
<@carlwgeorge:matrix.org>
16:31:33
i for one fully support wanting things to bake a bit more in fedora before landing them in rhel *cough* modularity *cough*
<@conan_kudo:matrix.org>
16:31:52
it is a real problem that fedora doesn't use features it asks for because of rhel
<@conan_kudo:matrix.org>
16:32:22
it is a lot less bad now than it was 7 years ago, but it is a problem still
<@conan_kudo:matrix.org>
16:33:05
we know that even if we update the packaging guidelines to prefer the new stuff, if it's not supported in any epel branch, a very large contingent of people will just not use it anyway
<@conan_kudo:matrix.org>
16:33:41
and that has the effect of making it harder to get that feedback that they want
<@conan_kudo:matrix.org>
16:34:06
that's why tilde versioning got backported all the way back to rhel 6
<@conan_kudo:matrix.org>
16:34:23
and why sourcelicense is now backported to rhel 9
<@carlwgeorge:matrix.org>
16:34:40
like i said, easy to handle with a SHOULD. many packages are fedora only and would be able to follow it. adoption of any new thing is going to be slow, even with the guidelines change. as a general rule, people don't read docs or guidelines.
<@salimma:fedora.im>
16:34:48
the sourcelicense thing... going forward, I wish RPM would just gain the ability to be told to ignore certain tags
<@salimma:fedora.im>
16:35:05
<@salimma:fedora.im>
16:35:05
that would solve the problem for things that are not critical, like SourceLicense
<@salimma:fedora.im>
16:35:05
wouldn't help us with %conf of course
<@conan_kudo:matrix.org>
16:35:07
I understand why it doesn't work that way, even if it is annoying
<@salimma:fedora.im>
16:35:42
any particular reason? certainly we don't want to let it just silently ignore anything it does not understand, but having an allowlist would be nice
<@salimma:fedora.im>
16:36:01
just ship it as a config file so we can adjust accordingly when there is a new field
<@conan_kudo:matrix.org>
16:36:05
because they influence available variables
<@conan_kudo:matrix.org>
16:36:17
which can change how macros behave and such
<@conan_kudo:matrix.org>
16:36:37
it's not safe to ignore unknown tags because we don't know how they're used
<@salimma:fedora.im>
16:40:41
yeah.. probably not worth doing then
<@james:fedora.im>
16:40:56
I think the main reason for this is because it was so visible (anything even looking at the built package needed to understand it).
<@james:fedora.im>
16:41:15
Anyway ... unless anyone has anything else I'll close in a couple of minutes.
<@conan_kudo:matrix.org>
16:44:28
got nothing from me
<@james:fedora.im>
16:47:14
!endmeeting