Hi all,
please find the meeting minutes in rendered form here:
https://md.archlinux.org/s/WeAd-Jd20#
The plain markdown is attached below:
# 2022-08-31 repod meeting
Date: 2022-08-31T17:00:00Z - 2022-08-31T18:30:00Z
Location: Jitsi
Scribe:
## Attendees
* dvzrv
* wCPO
* polyzen
* Segaja
* sradjoker
* bittin
* Prawn
## Endeavor OS
- sradjoker from Endeavor OS is describing their setup
- uses own tooling (e.g. Makefile/ scripts based) to maintain packages
- most packages are reused from Arch Linux / Arch Linux ARM
- confused by initial use/ what to expect from repod. adding further
high level information might help first-time users
([#108](https://gitlab.archlinux.org/archlinux/repod/-/issues/108))
## Demo by David
David demoed importing a package and how the file structure is created.
David presented the current milestone:
[0.3.0](https://gitlab.archlinux.org/archlinux/repod/-/milestones/2#tab-issu….
## Evaluate the minimum base set of package attributes
We need to evaluate whether the base set of package attributes needs to
be reduced:
[#109](https://gitlab.archlinux.org/archlinux/repod/-/issues/109).
Packagers with very limited/ small setups may otherwise run into issues.
In the future the existence requirement for the various attributes could
become configurable.
--
https://sleepmap.de
Hi all,
please find the rendered meeting minutes here:
https://md.archlinux.org/s/_N4trcwIK
The raw markdown below:
# 2022-08-16 repod meeting
Date: 2022-08-16T17:00:00Z - 2022-08-16T
Location: Jitsi
Scribe: brett
## Attendees
* dvzrv
* wCPO
* luna
* polyzen
* brett
* Segaja
## Agenda
### Milestone 0.2.0
* Create integration for downstream packagers
* More DB/data management
* Rudimentary PGP verification with pacman-key
* Multi-arch improvements
* New fields in the JSON output
* buildinfo
* installed packages while building the package
* other system states while package was built
* repod will now complain when not specifying an arch when arch is ambiguous
* A few bugs left to iron out in current MRs
### Future version goals
* Cleanup in management repo logic
* Workflow scenarios
* Group package actions
* Git backend work
* Decide between using plain Git vs the frameworks available to us
* Signed package repo DBs
## Pacman changes
A few internal Pacman changes have caused a little bit of needed reworking.
Hi all,
the next meeting for repod will take place on jitsi on 2022-08-16
19:00CEST.
NOTE: This is a Tuesday, where usually it is a Wednesday (I just don't
have time next Wednesday).
https://meet.jit.si/20220816-repod
Best,
David
--
https://sleepmap.de
Hi all,
please find the rendered meeting minutes here: https://md.archlinux.org/s/vyKfwOzzY#
The raw markdown below:
# 2022-08-03 repod meeting
Date: 2022-08-03T17:00:00Z
Location: Jitsi
Scribe: brett
## Attendees
* dvzrv
* wCPO
* luna
* polyzen
* brett
## Agenda
### 0.2.0 Milestone
The [0.2.0 milestone](https://gitlab.archlinux.org/archlinux/repod/-/milestones/1#tab-… needs:
* Consolidation of the management repositories
* More mtadata parsing/management (#20) (#39)
* Debug repositories
* BUILDINFO v2
* Using sha256sum to make checking out pkgbuilds easier
### Debug repositories
The feature for tracking debug repositories of package repositories has
been merged
[!78](https://gitlab.archlinux.org/archlinux/repod/-/merge_requests/78).
It is now possible to target the debug repository of a package
repository (e.g. move data of consumed package to respective directory
in the management repository layout).
### .BUILDINFO data in management repository
Work has started for adding relevant .BUILDINFO data to the management
repository
([#20](https://gitlab.archlinux.org/archlinux/repod/-/issues/20)). While
the change is fairly trivial, adapting tests to cover this is more
involved.
### Source URL tracking in management repository
A ticket [#39](https://gitlab.archlinux.org/archlinux/repod/-/issues/39)
in the current milestone concerns itself with the tracking of the
upstream source URL (the version controlled upstream of the PKGBUILD
from which a given pkgbase has been built). It still needs further
discussion and refinement, as the workflow around this feature is not
yet established.
### Checks on package consumption
The checks done on package consumption by dbscripts are currently not
translated or implemented by repod. Levente has promised to look into
this to create tickets for repod which will track the checks required
during package consumption.
### Sphinx improvements
Polyzen offered to look into sphinx related documentation improvements
(e.g. [#77](https://gitlab.archlinux.org/archlinux/repod/-/issues/77)
and [#78](https://gitlab.archlinux.org/archlinux/repod/-/issues/78)) in
the current milestone.
### Thoughts on current and future design
While the current milestone concerns itself with the consolidation of
management and package repository data and actions, the upcoming
milestone is about implementing a git-based backend for the management
repository.
Changes in the current milestone (e.g.
[#20](https://gitlab.archlinux.org/archlinux/repod/-/issues/20)) and
upcoming changes in pacman (making `pgpsig` a non-required field in the
sync databases) show, that it will not be possible to retain the current
functionality of allowing a roundtrip from sync database to management
repository and back to sync database, as the sync database will not
contain all the data required for the management repository anymore.
### Walkthrough
Dvzrv did a code walkthrough of the current repository, outlining the
various functionality contained in it.
A short example of importing a package file into the directory structure
maintained by repod has been show-cased as well.
### Q&A
Q: "Is the plan to use git instead of subversion for packages?""
A: Yes, the eventual plan is to move package build sources and package
repository state management to gitlab.archlinux.org.
Best,
David
--
https://sleepmap.de
Dear archlinux-keyring maintainers,
I'm working on getting archlinux-keyring packaged in Debian and Ubuntu
(as it would make it easier for mkosi - https://github.com/systemd/mkosi
- to generate Arch images from Debian and Ubuntu hosts).
There are some patches I'm applying that should probably be upstreamed;
since I can't fork the repo on GitLab, please find the first one
attached here.
It fixes the permission of the keyring files; they are not executable so
should be installed with 644 rather than 755.
It's also available on my Git repo:
https://git.sr.ht/~michel-slm/archlinux-keyring/commit/aa42fc9c4a2b66a16997…
Best regards,
--
Michel Alexandre Salim
identities: https://keyoxide.org/5dce2e7e9c3b1cffd335c1d78b229d2f7ccc04f2