Le 25/01/2023 à 16:47, Julian Hyde a écrit :
A common practice in many Git-based projects is to add “Co-authored-by”
comments. To Git they are merely lines in the commit message, but they
are recognized by GitHub tooling.
Our merge script already does that with all commit authors in the PR,
AFAICT.
Regards
Antoine.
I merged such a commit to Calcite yesterday:
a326bd2d0e0b4b6b3336f10217b0ecbb79522239.png
[CALCITE-5424] Customize handling of literals based on type system ·
apache/calcite@a326bd2
<https://github.com/apache/calcite/commit/a326bd2d0e0b4b6b3336f10217b0ecbb79522239>
github.com
<https://github.com/apache/calcite/commit/a326bd2d0e0b4b6b3336f10217b0ecbb79522239>
<https://github.com/apache/calcite/commit/a326bd2d0e0b4b6b3336f10217b0ecbb79522239>
Note how GitHub says “julianhyde and olivrlee committed x hours ago”
rather than the usual “olivrlee authored and julianhyde committed x
hours ago”.
I do this only when both authors have made substantial (> 25%)
contributions. If I, as reviewer, make minor changes such as copy
editing comments or adding a unit test., that does not warrant a
co-author tag.
Julian
On Jan 25, 2023, at 1:47 AM, Antoine Pitrou <anto...@python.org> wrote:
Hi Anja,
Le 24/01/2023 à 23:19, Anja a écrit :
A new contributor might not look at this section of the documentation,
because it is written for reviewers, not PR authors.
That is true. Perhaps it would be desirable to split that document in
two: one with general guidelines for what makes a PR ready for
merging, one with specific guidelines for reviewers.
Things to consider are:
* Would it make sense to mention that a PR could be adopted, and the
circumstances in which it might be adopted, in the Lifecycle of a PR:
https://arrow.apache.org/docs/developers/guide/step_by_step/pr_lifecycle.html.
* What is a decent standard for "waiting for response to ping" (1
week? 2?
Context dependent?).
One week or two sounds fine to me. It's certainly context-dependent
(is the contribution strongly needed, is it a prerequisite for other
changes? perhaps there's a release looming and it would be a welcome
addition? is the author usually responsive and collaborative, or did
they already abandon PRs?).
* Are there community norms for credit attribution in the Squashed merge
commit, if multiple authors worked on a PR over time?
I don't think there are any. Usually I would try to keep credit to the
"main" author, which is a subjective estimate but should be correlated
to the number of lines of code changes (even if most of this code is
"boring" and someone else contributed a small critical change).
* What is the process if OP has not set "allow contributions to this
PR" on
GitHub or if the adopted is not an Arrow committer? Does someone fork
their
fork?
Personally, I would create a branch on my fork with the same changes
(but with authorship potentially lost, if I just apply them as a patch).
Regards
Antoine.
