Hello Junio,
On 2024-04-17 17:27, Junio C Hamano wrote:
Phillip Wood <phillip.wood123@xxxxxxxxx> writes:
Playing devil's advocate for a minute, is this really common enough to
justify a new option when the user can use "--subject-prefix='PATCH
RESEND'" instead?
The same applies to "--rfc", but the justification goes like this.
* When you are working on a single subsystem in a larger project,
your patches would want to carry the subsystem name. You'd use
"--subject-prefix='PATCH frotz'" (and more likely it comes from
format.subjectPrefix in a working repository dedicated to work on
the frotz subsystem) for that.
* In the context of working on that subsystem, sometimes you would
need to mark your patch as a RFC patch, i.e., "[RFC PATCH frotz]",
and that is done per-invocation basis (i.e., you are not always
constantly sending an RFC) with "--rfc".
Having orthogonal two mechanisms whose results are concatenated
together is handy than having to specify the whole thing.
I somehow thought that during the review of the "--rfc" option a few
ideas were brought up to deal with adornments other than but similar
to RFC. I still think the approach to make "--rfc" take an optional
value, e.g., "--rfc=WIP" from the repository working in "frotz"
subsystem would produce "[WIP PATCH frotz v2 2/4]" a reasonable one.
With all due respect, "--rfc=WIP" looks like a kludge, simply
because "--rfc" should, IIUC, be some kind of a fixed shorthand.
Perhaps a new option should be added for that purpose, but I'm
not really sure how it could be called.
cf. https://lore.kernel.org/git/xmqqbkepep9k.fsf@gitster.g/
Thanks.
