Thread (20 messages) 20 messages, 4 authors, 2021-08-21

Re: [RFC PATCH 1/5] checkpatch: improve handling of revert commits

From: Denis Efremov <efremov@linux.com>
Date: 2021-08-21 07:35:22
Also in: linux-kselftest, lkml


On 8/21/21 10:12 AM, Joe Perches wrote:
On Sat, 2021-08-21 at 09:47 +0300, Denis Efremov wrote:
quoted
On 8/20/21 1:17 AM, Joe Perches wrote:
quoted
And I'm not sure if this particular ERROR is that useful overall.
I find it useful to check commit-id and that it matches a title.
It's easy to make a typo in commit-id and get an invalid one.
That's true, but I meant requiring the sha1 hash to contain both
the word "commit" and use ("title").

Looking at checkpatch's errors produced by this GIT_COMMIT_ID
test makes the required form seem a bit too inflexible to me.

For instance: a sha1 hash may be repeated in a commit message where
the first instance has the correct form but the second use is just
the hash and the warning is still produced.
I agree with you. There is also another example with list of commits:
 - commit <id-1> ("Title1")
 - commit <id-2> ("Title2")
...

I see no reason in writing "commit" on each line.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help