Re: RFC: using supersedes: trailer to indicate patch/series revision flow

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Thursday, November 7, 2019 9:43:49 PM CET Konstantin Ryabitsev wrote:
> Hi, all:
> 
> The only mechanism we currently have for patch/series versioning is 
> subject suffixes. I think it would be useful to have a way to more 
> explicitly mark that a series obsoletes a previous version, and I 
> propose this is done with a `supersedes:` trailer at the end of the 
> cover letter or in the first patch of the series:

All such things are a pain for patches generated by quilt, say.

> E.g.:
> 
> Initial patch:
> 
> ---8<---
> From: Dev Eloper <dev.eloper@example.com>
> Message-Id: <1572991351-86061-1-git-send-email-dev.eloper@example.com>
> Subject: [PATCH] Change foo
> 
> Foo is no good. Use Bar.
> 
> Signed-off-by: Dev Eloper <dev.eloper@example.com>
> ---
> 
>   foo | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/foo b/foo
> ...
> 
> base-commit: 23fdb198ae81f47a574296dab5167c5e136a02ba
> --
> 2.24.0
> ---8<---
> 
> Follow-up patch:
> 
> ---8<---
> From: Dev Eloper <dev.eloper@example.com>
> Message-Id: <1572991352-86062-2-git-send-email-dev.eloper@example.com>
> Subject: [PATCH,v2] Change foo
> 
> Foo is no good. Use Bar. Also use baz.
> 
> Signed-off-by: Dev Eloper <dev.eloper@example.com>
> ---
> 
>   foo | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/foo b/foo
> ...
> 
> supersedes: <1572991351-86061-1-git-send-email-dev.eloper@example.com>
> base-commit: 23fdb198ae81f47a574296dab5167c5e136a02ba
> --
> 2.24.0
> ---8<---
> 
> Questions:
> 
> 1. Should this be exposed via git-format-patch flags, or just used by
>     specialized tooling?

I'd vote for specialized tooling (if anything).

> 2. Should supersedes: link to the previous version of the patch, or the
>     first ever version of the patch? I am leaning towards the latter,

And then how do you know that version 2 was superseded by version 3?

>     even though in this case the message-id largely becomes identical in
>     usage to Gerrit's Change-Id.
> 
> 3. Should the supersedes trailer have:
>     a. message-id without brackets
>     b. message-id with brackets
>     c. https://lore.kernel.org/r/message-id
>     My preference is b, to match with The Message-Id header usage.

An id of the exact patch (or series) superseded by this one IMO.

Cheers!