<@humaton:fedora.im>
14:00:14
!startmeeting Git Forge Meeting
<@meetbot:fedora.im>
14:00:15
Meeting started at 2025-02-19 14:00:14 UTC
<@meetbot:fedora.im>
14:00:15
The Meeting name is 'Git Forge Meeting'
<@humaton:fedora.im>
14:00:19
!topic init
<@humaton:fedora.im>
14:00:19
!info this is meeting about the Fedora git forge replacement this meeting template can be found at https://codeberg.org/fedora/gitforge-migration
<@nphilipp:fedora.im>
14:00:20
!hi
<@zodbot:fedora.im>
14:00:22
Nils Philippsen (nphilipp) - he / him / his
<@smilner:fedora.im>
14:00:22
!hi
<@zodbot:fedora.im>
14:00:24
None (smilner)
<@humaton:fedora.im>
14:01:32
How is it on your side of the cable? Its beautifully sunny here but -8C
<@nhanlon:beeper.com>
14:02:11
!hi
<@zodbot:fedora.im>
14:02:12
Neil Hanlon (neil) - he / him / his
<@nhanlon:beeper.com>
14:02:20
(just lurking today mostly)
<@smilner:fedora.im>
14:02:33
23C ... we get a few nice days a year ๐
<@lenkaseg:fedora.im>
14:02:36
!hi
<@zodbot:fedora.im>
14:02:37
Lenka Segura (lenkaseg)
<@nphilipp:fedora.im>
14:03:34
We had -7.5C this morning, but itโs up to +4C already and supposed to hit 18C in the next days
<@nphilipp:fedora.im>
14:03:56
(and itโs sunny ๐)
<@humaton:fedora.im>
14:04:36
so yeah its + on the sun it just does not shine over here :D
<@dherrera:fedora.im>
14:04:58
!hi
<@zodbot:fedora.im>
14:04:59
Diego Herrera (dherrera) - he / him / his
<@humaton:fedora.im>
14:05:51
its 5 after lets start with the agenda
<@humaton:fedora.im>
14:06:51
!topic tracking of work
<@amoloney:fedora.im>
14:06:53
Im here, but not really :/ In a talk/lab in Cork office
<@humaton:fedora.im>
14:07:26
You might seen some tickets getting created by Ryan, he was playing with forgejo to dog-food it.
<@humaton:fedora.im>
14:07:41
We met today and decided to continue using pagure.
<@humaton:fedora.im>
14:08:00
I would like to ask you to label your assigned tickets so we get it synced to jira
<@humaton:fedora.im>
14:08:28
Or they will be labeled by me to low later this week ๐
<@humaton:fedora.im>
14:10:28
mobing on
<@humaton:fedora.im>
14:10:29
!topic Tickets and Updates
<@humaton:fedora.im>
14:10:47
!link https://pagure.io/fedora-infra/forgejo-deployment/issues
<@humaton:fedora.im>
14:10:47
!info go over tickets and updates
<@humaton:fedora.im>
14:10:47
<@humaton:fedora.im>
14:10:47
<@humaton:fedora.im>
14:11:05
!topic UI changes
<@humaton:fedora.im>
14:11:18
!link https://pagure.io/fedora-infra/forgejo-deployment/issue/8
<@humaton:fedora.im>
14:11:42
Yaash did a good job documenting where he is in the ticket
<@humaton:fedora.im>
14:12:05
I am meeting with him tomorrow to sync up and publish the updated changes
<@lenkaseg:fedora.im>
14:12:07
mmm...I seem not to have perms to change the tickets metadata in pagure/fedora-infra/forgejo-deployment
<@humaton:fedora.im>
14:12:57
huh try now? I have added fedora-infra group as colaborator
<@Zlopez:matrix.org>
14:12:58
!hi
<@zodbot:fedora.im>
14:13:01
Michal Konecny (zlopez)
<@lenkaseg:fedora.im>
14:13:24
now yes, thanks!
<@humaton:fedora.im>
14:13:38
back to the topic of UI changes
<@humaton:fedora.im>
14:14:11
I am thinking about how should we introduce the changes, should we apply patches to the source?
<@humaton:fedora.im>
14:14:26
Or should we inject them during the container build?
<@humaton:fedora.im>
14:14:41
Any ideas?
<@gwmngilfen:fedora.im>
14:14:52
maintaining a branch of the source to cherry-pick / rebase from seems more transparent
<@Zlopez:matrix.org>
14:14:57
What is more transparent?
<@gwmngilfen:fedora.im>
14:15:07
easier to follow for contributors
<@humaton:fedora.im>
14:15:27
well that depends on the devel it will be container build for some and code for others
<@gwmngilfen:fedora.im>
14:15:56
perhaps I'm just showing my bias then ๐
<@humaton:fedora.im>
14:15:56
Both are equally transparent if we document the process
<@humaton:fedora.im>
14:16:14
the thing about injecting is that we dont have to maintain fork at all...
<@gwmngilfen:fedora.im>
14:17:20
is there an effect on testing changes?
<@humaton:fedora.im>
14:17:40
we will have "just" repo with assets and templates that will be copied in during container build
<@gwmngilfen:fedora.im>
14:17:57
I mean that from either path - trying to be objective, or we'll all just argue for our own defaults ๐
<@nphilipp:fedora.im>
14:18:33
Mhm if we want to test UI, the container probably should be that test target
<@humaton:fedora.im>
14:18:33
actually this is interesting question
<@nphilipp:fedora.im>
14:18:47
Mhm if we want to test UI, the container probably should be the test target
<@humaton:fedora.im>
14:19:16
I prefer targeting the service for testing i.e the container
<@humaton:fedora.im>
14:19:27
testing on the source level is done by upstream
<@humaton:fedora.im>
14:19:35
or should be done :)
<@humaton:fedora.im>
14:19:50
but forgejo folks do good job with testing what is covered
<@gwmngilfen:fedora.im>
14:20:46
sure, but sooner or later we'll have a conflict between upstream and our patches, presumably that would be easy to catch too? Obviously it's easy when rebasing, as the rebase fails ๐
<@humaton:fedora.im>
14:21:07
hopefully we will not carry on patches
<@humaton:fedora.im>
14:21:31
The UI changes are done transparently by adding new theme
<@humaton:fedora.im>
14:21:37
and configuring it in UI
<@gwmngilfen:fedora.im>
14:21:51
you said patches here ๐
<@humaton:fedora.im>
14:22:17
ok my bad i mean code changes,
<@gwmngilfen:fedora.im>
14:22:17
if it's just themes and assets then I have far less worries
<@humaton:fedora.im>
14:22:32
I read you patches as RPM style patches on the side of the source
<@gwmngilfen:fedora.im>
14:23:31
so if we're just adding net-new files that cannot conflict with upstream, then I will shut up, either way can work ๐
<@humaton:fedora.im>
14:23:59
Ok, we will look into both and see. Because as you said later on we might have more changes...
<@humaton:fedora.im>
14:24:46
Avoiding source conflicts and easiness of maintenance will be the criteria.
<@humaton:fedora.im>
14:25:30
!topic Explore current Forgejo deployment methods
<@humaton:fedora.im>
14:25:33
!link https://pagure.io/fedora-infra/forgejo-deployment/issue/2
<@humaton:fedora.im>
14:25:58
ah I see David around
<@dkirwan:fedora.im>
14:26:14
hey, we have the instance available now, and you can sign in via FAS
<@dkirwan:fedora.im>
14:26:29
https://forgejo.apps.ocp.stg.fedoraproject.org/
<@humaton:fedora.im>
14:26:54
!link https://forgejo.apps.ocp.stg.fedoraproject.org/explore
<@dkirwan:fedora.im>
14:27:02
feel free to log in, poke around, but dont expect too much atm, I might need to tear it down and redeply multiple times as we're figuring out how to deploy with custom configurations
<@nhanlon:beeper.com>
14:27:53
a few of us in commops had a good time doing some testing a week ago or so. excellent job!
<@zodbot:fedora.im>
14:28:05
neil has already given cookies to dkirwan during the F41 timeframe
<@humaton:fedora.im>
14:28:38
We need to look into how to seed the deployment
<@zodbot:fedora.im>
14:28:44
nphilipp gave a cookie to dkirwan. They now have 13 cookies, 5 of which were obtained in the Fedora 41 release cycle
<@zodbot:fedora.im>
14:28:56
lenkaseg gave a cookie to dkirwan. They now have 14 cookies, 6 of which were obtained in the Fedora 41 release cycle
<@zodbot:fedora.im>
14:29:21
dherrera gave a cookie to dkirwan. They now have 15 cookies, 7 of which were obtained in the Fedora 41 release cycle
<@dkirwan:fedora.im>
14:30:37
currently just looking inside the helm chart weve forked, presumably we'll use the fedora branch
<@dkirwan:fedora.im>
14:31:20
build an image containing these charts, and we can pass in any variables to customise anything else needed then
<@humaton:fedora.im>
14:31:52
David Kirwan: sync with lenkaseg she did the image builds from source last week.
<@dkirwan:fedora.im>
14:32:27
will do!
<@lenkaseg:fedora.im>
14:32:54
Here one of the images: https://quay.io/repository/konflux-fedora/fedora-infra-tenant/forgejo-oci-images?tab=tags&tag=latest
<@humaton:fedora.im>
14:33:09
!topic Package Forgejo as vendored golang RPM
<@humaton:fedora.im>
14:33:12
!link https://pagure.io/fedora-infra/forgejo-deployment/issue/7
<@humaton:fedora.im>
14:33:23
did you have any progress nils ?
<@nphilipp:fedora.im>
14:33:34
Any? Yes
<@nphilipp:fedora.im>
14:33:51
Meanwhile I got it to build, but it was reluctant
<@nphilipp:fedora.im>
14:34:20
I restarted my effort, starting with another attempt of go2rpm
<@nphilipp:fedora.im>
14:34:48
I.e. first I tried to get the make/Makefile route to work (how upstream builds it). That was a lot of pain
<@humaton:fedora.im>
14:35:25
are you ready to share your efforts? should we create ti repo on codeberg?
<@nphilipp:fedora.im>
14:35:28
Now Iโm building things using go generate/build/... and it works better โ probably this will bite me in the behind on packaging assets
<@nphilipp:fedora.im>
14:36:43
We can do that, but my efforts are very chaoticโฆ at least have been so far. Perhaps I should put each in a separate branch? I donโt want to have that chaos as a constant reminder of past failure ๐
<@nphilipp:fedora.im>
14:37:11
Anyway, I can do that, somehow ๐
<@humaton:fedora.im>
14:37:32
oh I understand that, its ok to wait and share something that you are more sure about...
<@gwmngilfen:fedora.im>
14:38:11
๐ถ rebase your sorrows away ๐ถ
<@nphilipp:fedora.im>
14:38:20
Iโm fine sharing things, just donโt want the depressing history
<@gwmngilfen:fedora.im>
14:38:35
(only works in your own repos ofc)
<@nphilipp:fedora.im>
14:38:36
Gwmngilfen: yeah, thatโs what I was trying to prepare the audience for
<@zodbot:fedora.im>
14:39:25
jflory7 gave a cookie to dkirwan. They now have 16 cookies, 8 of which were obtained in the Fedora 41 release cycle
<@humaton:fedora.im>
14:39:50
nils: do you have account at codeberg?
<@nphilipp:fedora.im>
14:39:59
So I can upload the spec file with comments (e.g. how to create the vendor tarball)
<@humaton:fedora.im>
14:40:01
I can add you to the org and you should be able to add new repo
<@nphilipp:fedora.im>
14:40:05
I have, yes
<@nphilipp:fedora.im>
14:40:38
Iโd like to keep it under my name for the time being, and move it to the org once Iโve gotten it into shape
<@humaton:fedora.im>
14:41:52
nils: if you can give me your username I will add you and you can do that whenever you want....
<@zodbot:fedora.im>
14:42:09
t0xic0der gave a cookie to dkirwan. They now have 17 cookies, 9 of which were obtained in the Fedora 41 release cycle
<@humaton:fedora.im>
14:42:30
ha found you!
<@nphilipp:fedora.im>
14:42:38
jednorozec: https://codeberg.org/nilsph
<@nphilipp:fedora.im>
14:42:40
heh
<@humaton:fedora.im>
14:44:14
if there are no questions about the rpm build lets move on
<@nphilipp:fedora.im>
14:44:30
Iโm adding a couple comments to the spec file and upload it
<@nphilipp:fedora.im>
14:44:38
Iโm adding a couple comments to the spec file and will upload it later
<@nphilipp:fedora.im>
14:44:46
Yeah, move on
<@humaton:fedora.im>
14:45:36
!topic OCI images
<@humaton:fedora.im>
14:45:39
!link https://pagure.io/fedora-infra/forgejo-deployment/issue/1
<@humaton:fedora.im>
14:46:08
as lenkaseg pointed out in previous comment there are OCI images build by konflux that she did
<@lenkaseg:fedora.im>
14:46:42
yup, it's the forgejo-oci-images app in our konflux tenant
<@humaton:fedora.im>
14:46:48
any updates questions notes?
<@lenkaseg:fedora.im>
14:47:16
there are two components now, one builds from our fedora/forgejo fork and another one builds from the upstream forgejo
<@lenkaseg:fedora.im>
14:47:30
there was some back and forth with konflux
<@humaton:fedora.im>
14:47:38
Are you building the stable branch as well?
<@humaton:fedora.im>
14:47:44
we deploy v10 now
<@humaton:fedora.im>
14:48:19
I would suggest to build v10 from upstream
<@lenkaseg:fedora.im>
14:48:42
when I fixed the containerfiles to work with fedora:rawhide and buildah, konflux built it, but because of some reported vulnerabilities in gitea version or so, it looked like the build failed...
<@lenkaseg:fedora.im>
14:49:09
yeah, it builds from the v10 branch
<@lenkaseg:fedora.im>
14:49:43
btw I noticed our fork had the main branch as default, so changed it to "forgejo" to match the upstream
<@humaton:fedora.im>
14:49:49
so it build from v10 sources from upstrema and main branch from our fork?
<@humaton:fedora.im>
14:49:55
nope!
<@lenkaseg:fedora.im>
14:49:57
exactly
<@humaton:fedora.im>
14:49:59
It should be main
<@humaton:fedora.im>
14:50:13
Our main differs from forgejo branch
<@lenkaseg:fedora.im>
14:50:53
ah, ok! so that means I synced the upstream forgejo branch into fedora fork forgejo branch....and it should be main then?
<@lenkaseg:fedora.im>
14:51:22
ok, I will return the main branch then as a default
<@lenkaseg:fedora.im>
14:51:54
then we probably want to cherrypick stuff from the forgejo branch to main?
<@humaton:fedora.im>
14:52:25
we can rebase it
<@humaton:fedora.im>
14:52:34
actually it can be done automatically
<@lenkaseg:fedora.im>
14:52:57
I think there were some merge conflicts when I tried...
<@humaton:fedora.im>
14:53:30
then something went wrong, I was rebasing it for about month and it was building in the other pipeline in konflux
<@lenkaseg:fedora.im>
14:54:11
that will be most probably me getting tangled in git :)
<@humaton:fedora.im>
14:54:51
untangle yourself :)
<@humaton:fedora.im>
14:54:55
lets move on
<@humaton:fedora.im>
14:55:02
!topic Open floor
<@lenkaseg:fedora.im>
14:55:39
continuing on this, merging the konflux-proposed pull requests fixed that so now it's all green and building and stuff :)
<@humaton:fedora.im>
14:56:04
did you really merge them?
<@humaton:fedora.im>
14:56:11
or created patches and applied?
<@humaton:fedora.im>
14:56:30
because merge on GH will be overridden by the next sync...
<@lenkaseg:fedora.im>
14:56:37
the latter
<@humaton:fedora.im>
14:57:21
cool
<@lenkaseg:fedora.im>
14:57:39
:)
<@humaton:fedora.im>
14:57:42
Lets close the meeting and save 3mins
<@humaton:fedora.im>
14:57:48
!endmeeting