Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 943294 - Policy for pushing changes for python eclasses
Summary: Policy for pushing changes for python eclasses
Status: CONFIRMED
Alias: None
Product: Documentation
Classification: Unclassified
Component: Policy Guide (show other bugs)
Hardware: All Linux
: Normal normal
Assignee: Python Gentoo Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-11-11 20:28 UTC by Arthur Zamarin
Modified: 2024-11-11 20:28 UTC (History)
3 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Arthur Zamarin archtester Gentoo Infrastructure gentoo-dev Security 2024-11-11 20:28:34 UTC
I want to open this discussion in hopes we understand the current policy on pushing changes to python eclasses in Gentoo.

From my side, I know some "informal" rules, which are:

1. Discuss it with project members in IRC. We are mostly active up there, so quite logical.

2. Open a PR with the changes, mention the eclass name in the PR title, and ping @python project for review. Post in IRC. Wait for review.

3. (formal) Post patches to ML - same for all eclass changes - wait for comments.

4. Accumulate them until an undefined point, and then push - to mitigate cache metadata regen (FYI, the only mention of this in official documentation is the short explanation at [1])



For me, it mostly worked, since I myself don't like to perform eclass changes, so I mostly wrote ideas and requests, mgorny implemented them and handled it for me. I never had anything urgent so easy life.

Now, I want us to see some issues with this workflow, more for people who are fine with authoring eclass changes, like Sam and Eli.

- How much time to wait for GitHub review. Does missing review makes it close to an ACK (I limit this question to core @python members).

- How much time to accumulate? We could wait until something urgent comes, but this is too long and depressing as an author. I also don't think we should just wait for a single day (unless urgent).
I myself don't know the full cost of metadata regen - from my view it is mostly mitigated by `git pull --depth=1` which most users do. I also think that rsync emerge sync is quite rare, and git is more common (no proof). So please let's have discussion on this cost.




I'm requesting the participant I CCed to raise more questions on the policy of python eclass changes, and answer as they can, and raise a discussion if needed. I don't want the progress stopping, or devs feeling hurt, or doing stupid things that hurt Gentoo users.
Therefore understanding the situation and decisions will leave all happy I hope.

I also really request the reader to come with good faith, and to attempt to ignore any feeling of being attacked/misjudged - we all come from different parts of the world, and our words sometimes change meaning when read by the other. Assume we all are friends :)


[1] https://devmanual.gentoo.org/eclass-writing/index.html#adding-and-updating-eclasses