Minutes March 10 2021 - math-comp/math-comp GitHub Wiki

Creation of a code owner file:

• motivation = ease the process of finding reviewers and shorten the process time of a PR.
• problem supposed to be solved by the creation of code owners = stimulate the work on reviews
• problems =
  • we are a few people.
  • may be the PR procedure is slowing down the process?
  • what is the expected for a PR to be merged.
  • are we welcoming enough to external contributors? if so what would help?
• proposed solutions
  • code owner = automated assignee (and may be reviewer too)
  • set a deadline for each PR, if assignee does start then PR is added to the agenda of the next mathcomp dev meeting

Enrico make more comment to the PR process, and prompts people to think of the status of the PR process

• can we use more often the "merge myself" option?
• are we caring too much for code quality for merged code?

Cyril: the proposal for code owners was made to change the situation where he would be both the main author of many (most?) PRs and the main author or reviews. But is it the main problem at all?

Cyril's proposal: make a census of the problems (solutions) people think MathComp is facing wrt to its evolution as a software. Enrico created a pad for collecting ideas.

We did not discuss the point about the list of results. Reynald says that there might be an existing zulip chan dedicated to the topic. Will check. [NB(rei): I was thinking about this.]