Hi,
[ Quick reply during a final local test suite run. ]
intrigeri <intrigeri@???> (2020-03-27):
> By and large, yes.
>
> One thing that would change is that you would get raw data from Git +
> GitLab issues/MRs, without the rephrasing step currently done by the
> RM. I don't know how much this rephrasing work currently impacts your
> work: I guess it depends primarily on the target audience the RM has
> in mind, and on their tech writing skills.
>
> Personally I'd rather see us put this "express clearly what we're
> doing and why" effort into issue/MR titles and commit messages than on
> the RM-at-release-time's shoulders.
4.5~rc1 made it clear this can be a very interesting maze to navigate:
this included very-long-lived branches (dating back several years),
containing reverts, and revert of reverts, etc.; so our usual script
collecting all relevant git commit messages didn't generate a directly
useful output.
In a bunch of cases, I had to go back and check the final git diff
output from the previous release, to get a better understanding of the
final results.
This is also why I decided to go for an “abstract” view of most changes,
instead of documenting them in extenso as I've done for most if not
all other releases.
Having had a top-level summary in the relevant MRs for the few “parent
tickets” (#6560, #8415) would have been way more straightforward for me
(as the RM du jour) but quite possibly for the technical writers as
well, since they wouldn't have had to rely on my paraphrasing/summing up
those changes (even if anonym seemed to think I didn't do a bad job
there): they would have had information directly from developers.
Cheers,
--
Cyril 'kibi' Brulebois (ckb@???)
