Hi Boris,
On 1/24/2024 2:46 PM, Borislav Petkov wrote:
> On Wed, Jan 24, 2024 at 01:31:01PM -0800, Reinette Chatre wrote:
>> I do not know the motivation for that requirement. From what I can tell the
>> change [1] that added that check went in as first version without discussion.
>> [1] starts by saying that the format is "preferred" so I assume there is
>> some history that I am not familiar with.
>
> My main goal with commit messages, code comments and every other *text*
> you have in the code is to be as succinct and understandable as possible
> for time considerations, clarity, etc.
>
> If I see a 12-char sha1 followed by a title, to me that is a commit. No
> need to say "commit" too. But as already said, if you prefer to have
> "commit" there, I'll add it - no biggie.
I totally understand your sentiment. I am just the messenger here.
Without the "commit" this patch triggers a loud:
ERROR: Please use git commit description style 'commit <12+ chars of sha1> ("<title line>")' - ie: 'commit 92bd5a139033 ("x86/resctrl: Add interface to write mbm_total_bytes_config")'
#15:
92bd5a139033 ("x86/resctrl: Add interface to write mbm_total_bytes_config")
I usually (unless, for example, following checkpatch.pl advice causes a change
to fall out of place with surrounding code) try to format my patches to get a
clean slate from checkpatch.pl with the goal to eliminate obstacles to the
patch getting included.
Since you are the one that decides the rules for inclusion you can make this
check to be one where checkpatch.pl can be ignored. No objection from me if
you choose to do so (and I will note the precedent for future patches).
Reinette
