| Message ID | pull.1495.git.git.1681329955.gitgitgadget@gmail.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | cocci: codify authoring and reviewing practices | expand |
On Wed, Apr 12, 2023 at 1:05 PM Glen Choo via GitGitGadget <gitgitgadget@gmail.com> wrote: > > Here's the followup to the discussion in [1]. Sorry for the delay. > > I've tried to incorporate most of the responses from that thread as well as > suggest some guidelines that I think would make the authoring + reviewing > process smoother. I've opted for stronger wording to make the guidelines > easier to follow, but I don't feel strongly about the specifics. > > [1] > https://lore.kernel.org/git/kl6l7cuycd3n.fsf@chooglen-macbookpro.roam.corp.google.com > > Glen Choo (2): > cocci: add headings to and reword README > cocci: codify authoring and reviewing practices > > contrib/coccinelle/README | 33 +++++++++++++++++++++++++++++---- > 1 file changed, 29 insertions(+), 4 deletions(-) > > > base-commit: f285f68a132109c234d93490671c00218066ace9 > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-git-1495%2Fchooglen%2Fpush-lsxuouxyokwo-v1 > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-git-1495/chooglen/push-lsxuouxyokwo-v1 > Pull-Request: https://github.com/git/git/pull/1495 > -- > gitgitgadget I read through both patches, and I generally like them. I'm a little unsure on the "To give a Reviewed-by" bit of patch 2. For example, it could be possible that .cocci changes might apply a few different kinds of changes, and say only 2 of the 3 are reflected in the current tree, and those 2 types are handled correctly but the third type of change is buggy. The .cocci files would then be a bug waiting to happen. Maybe that's just a risk we take and it's okay for folks to give a Reviewed-by even being unfamiliar with cocci. Maybe the wording should instead be "It's okay to give a Reviewed-by: on a series that also contains cocci changes when you are unfamiliar with coccinelle; just state that your Reviewed-by is limited to the other bits". Or maybe the instructions should just be to give an Acked-by. You should probably have someone familiar enough with coccinelle that they know what is worth worrying about weigh in on that aspect. But you can have my Acked-by on the other bits. :-)
Elijah Newren <newren@gmail.com> writes: > .... Maybe > the wording should instead be "It's okay to give a Reviewed-by: on a > series that also contains cocci changes when you are unfamiliar with > coccinelle; just state that your Reviewed-by is limited to the other > bits". Or maybe the instructions should just be to give an Acked-by. > You should probably have someone familiar enough with coccinelle that > they know what is worth worrying about weigh in on that aspect. > > But you can have my Acked-by on the other bits. :-) The value of Reviewed-by takes two sides to determine. Even if we reserve a Reviewed-by to "I have reviewed the entirety of this patch, and the patch is something I can stand behind" (as opposed to "my understanding of this patch is iffy in this and that area, but all the other parts of the patch is something I can stand behind"), the value of such a Reviewed-by is conditional to "how well does the reviewer actually know the area?" A drive-by "Reviewed-by:" thrown into a review discussion thread by a total stranger would not carry much weight, until we know how much they are familiar with and how good a taste they have. And honest qualifying comments like "my understanding of this and that area is iffy so I cannot endorse these parts" helps build trust by others in the reviewer who gives such a partial review and we should encourage such behaviour. I agree "Acked-by:" with comments is a good idea. Thanks.