[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[zimoun]

* doc/contributing.texi (Submitting Patches): Add item about 'git-format-patch
--base'.
---
 doc/contributing.texi | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/doc/contributing.texi b/doc/contributing.texi
index af3601442e..5ea3cb1899 100644
--- a/doc/contributing.texi
+++ b/doc/contributing.texi
@@ -932,6 +932,12 @@ Before submitting a patch that adds or modifies a package definition,
 please run through this check list:
 
 @enumerate
+@cindex @code{git format-patch}
+@cindex @code{git-format-patch}
+@item
+We recommend to use the command @code{git format-patch --base} to
+include the commit where your patch applies.
+
 @item
 If the authors of the packaged software provide a cryptographic
 signature for the release tarball, make an effort to verify the

base-commit: e746d30ec15852ec50e4247ea0a07ab2a9a69aa8
-- 
2.28.0

bug#43946: [PATCH] doc: Add item to "Submitting Patches" section.

[Oleg Pykhalov]

[-- Attachment #1: Type: text/plain, Size: 969 bytes --]

Hi,

zimoun <zimon.toutoune@gmail.com> writes:

> * doc/contributing.texi (Submitting Patches): Add item about 'git-format-patch
> --base'.
> ---
>  doc/contributing.texi | 6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/doc/contributing.texi b/doc/contributing.texi
> index af3601442e..5ea3cb1899 100644
> --- a/doc/contributing.texi
> +++ b/doc/contributing.texi
> @@ -932,6 +932,12 @@ Before submitting a patch that adds or modifies a package definition,
>  please run through this check list:
>  
>  @enumerate
> +@cindex @code{git format-patch}
> +@cindex @code{git-format-patch}
> +@item
> +We recommend to use the command @code{git format-patch --base} to
> +include the commit where your patch applies.

“git format-patch” already mentioned in the documentation in the first
paragraph https://guix.gnu.org/manual/en/html_node/Submitting-Patches.html

Do I miss something?  Fill free to reopen the issue ;-)

Oleg.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 861 bytes --]

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[zimoun]

Dear,

Thank you for the review.

On Sat, 24 Oct 2020 at 11:01, Oleg Pykhalov <go.wigust@gmail.com> wrote:
> zimoun <zimon.toutoune@gmail.com> writes:
>
> > * doc/contributing.texi (Submitting Patches): Add item about 'git-format-patch
> > --base'.
> > ---
> >  doc/contributing.texi | 6 ++++++
> >  1 file changed, 6 insertions(+)
> >
> > diff --git a/doc/contributing.texi b/doc/contributing.texi
> > index af3601442e..5ea3cb1899 100644
> > --- a/doc/contributing.texi
> > +++ b/doc/contributing.texi
> > @@ -932,6 +932,12 @@ Before submitting a patch that adds or modifies a package definition,
> >  please run through this check list:
> >
> >  @enumerate
> > +@cindex @code{git format-patch}
> > +@cindex @code{git-format-patch}
> > +@item
> > +We recommend to use the command @code{git format-patch --base} to
> > +include the commit where your patch applies.
>
> “git format-patch” already mentioned in the documentation in the first
> paragraph https://guix.gnu.org/manual/en/html_node/Submitting-Patches.html

The point of the patch is:
 - the addition of the cindex
 - the addition of the option --base
and AFAITC it is not already in the manual.  If it is the wrong place,
please let me know where it could be better.

All the best,
simon

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[Ludovic Courtès]

Hi Simon!

zimoun <zimon.toutoune@gmail.com> skribis:

> * doc/contributing.texi (Submitting Patches): Add item about 'git-format-patch
> --base'.
> ---
>  doc/contributing.texi | 6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/doc/contributing.texi b/doc/contributing.texi
> index af3601442e..5ea3cb1899 100644
> --- a/doc/contributing.texi
> +++ b/doc/contributing.texi
> @@ -932,6 +932,12 @@ Before submitting a patch that adds or modifies a package definition,
>  please run through this check list:
>  
>  @enumerate
> +@cindex @code{git format-patch}
> +@cindex @code{git-format-patch}
> +@item
> +We recommend to use the command @code{git format-patch --base} to
> +include the commit where your patch applies.

I’m not entirely convinced TBH, in part because I know I often pile a
couple of WIP branches on top of one another, “knowing what I’m doing”
(actually hoping that I do), and so the base commit would be useless in
this case.

Closing?  :-)

Thanks,
Ludo’.

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[zimoun]

Hi Ludo,

On Fri, 15 Jan 2021 at 14:30, Ludovic Courtès <ludo@gnu.org> wrote:

> > +We recommend to use the command @code{git format-patch --base} to
> > +include the commit where your patch applies.
>
> I’m not entirely convinced TBH, in part because I know I often pile a
> couple of WIP branches on top of one another, “knowing what I’m doing”
> (actually hoping that I do), and so the base commit would be useless in
> this case.

I am not buying your argument. :-)
It is not because <name> does not use this information that it cannot
be a recommendation, i.e., a suggestion or advice on what seems
helpful.
Other said, it is not because it is "useless in this case" that it is
useless in other cases.

For example, this information about which known commit that patch
applies is helping for the automation of testing patches.  Well, see
[1,2] for instance.  Discussions of such tooling happened in #44625
[3] and Emacs helper [4].


1: <https://git.kernel.org/pub/scm/utils/b4/b4.git/tree/README.rst>
2: <https://docs.kyleam.com/piem/Using-b4-to-apply-patches.html>
3: <http://issues.guix.gnu.org/44625>
4: <https://inbox.kyleam.com/piem/20201115061518.22191-1-kyle@kyleam.com/>

> Closing?  :-)

As you want. :-)
But, I think "qui peut le plus peut le moins".  Even if maybe my
wording is not the good one, I still think that the "base-commit"
where the patch applies should be provided, at least suggested.

Cheers,
simon

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[zimoun]

Hi Ludo,


On Fri, 15 Jan 2021 at 15:00, zimoun <zimon.toutoune@gmail.com> wrote:
> On Fri, 15 Jan 2021 at 14:30, Ludovic Courtès <ludo@gnu.org> wrote:
>
>> > +We recommend to use the command @code{git format-patch --base} to
>> > +include the commit where your patch applies.
>>
>> I’m not entirely convinced TBH, in part because I know I often pile a
>> couple of WIP branches on top of one another, “knowing what I’m doing”
>> (actually hoping that I do), and so the base commit would be useless in
>> this case.

[...]

>> Closing?  :-)
>
> As you want. :-)
> But, I think "qui peut le plus peut le moins".  Even if maybe my
> wording is not the good one, I still think that the "base-commit"
> where the patch applies should be provided, at least suggested.

I still think that recommending to provide the commit on which it is
known that the patch (or patch set) applies is a good recommendation.
Especially when the submission rate is greater than the review rate and
the tree is moving really quickly (yeah!).

It is no extra work for the submitter and really helps for the reviewer.
They applies at base-commit, checks, rebases and resolves conflicts if
they are.  Otherwise, the patch is useless or it needs to be rewritten
by hand (or please indicate me how to do :-)).

BTW, it helps automation tools.

Sometimes, this base-commit is useless for the reviewer workflow but
having it does not interfere.  Having an information does not mean it
must be used.  However, not having an information implies it cannot be
used. ;-)


Cheers,
simon

[bug#43946] [PATCH] doc: Add item to "Submitting Patches" section.

[zimoun]

Hi Ludo,

On Fri, 15 Jan 2021 at 14:30, Ludovic Courtès <ludo@gnu.org> wrote:
> zimoun <zimon.toutoune@gmail.com> skribis:
>
>> * doc/contributing.texi (Submitting Patches): Add item about 'git-format-patch
>> --base'.
>> ---
>>  doc/contributing.texi | 6 ++++++
>>  1 file changed, 6 insertions(+)
>>
>> diff --git a/doc/contributing.texi b/doc/contributing.texi
>> index af3601442e..5ea3cb1899 100644
>> --- a/doc/contributing.texi
>> +++ b/doc/contributing.texi
>> @@ -932,6 +932,12 @@ Before submitting a patch that adds or modifies a package definition,
>>  please run through this check list:
>>  
>>  @enumerate
>> +@cindex @code{git format-patch}
>> +@cindex @code{git-format-patch}
>> +@item
>> +We recommend to use the command @code{git format-patch --base} to
>> +include the commit where your patch applies.
>
> I’m not entirely convinced TBH, in part because I know I often pile a
> couple of WIP branches on top of one another, “knowing what I’m doing”
> (actually hoping that I do), and so the base commit would be useless in
> this case.

Could you explain more?  Here [#,@], I argument:

--8<---------------cut here---------------start------------->8---
It is not because <name> does not use this information that it cannot be
a recommendation, i.e., a suggestion or advice on what seems helpful.
Other said, it is not because it is "useless in this case" that it
isuseless in other cases.

For example, this information about which known commit that patch
applies is helping for the automation of testing patches.  Well,
see [1,2] for instance.  Discussions of such tooling happened in
#44625 [3] and Emacs helper [4].

1: https://git.kernel.org/pub/scm/utils/b4/b4.git/tree/README.rst
2: https://docs.kyleam.com/piem/Using-b4-to-apply-patches.html
3: http://issues.guix.gnu.org/44625
4: https://inbox.kyleam.com/piem/20201115061518.22191-1-kyle@kyleam.com/

[..]

I still think that recommending to provide the commit on which it
isknown that the patch (or patch set) applies is a good recommendation.
Especially when the submission rate is greater than the review rate
andthe tree is moving really quickly (yeah!).

It is no extra work for the submitter and really helps for the
reviewer.  They applies at base-commit, checks, rebases and resolves
conflicts if they are.  Otherwise, the patch is useless or it needs to
be rewritten by hand (or please indicate me how to do :-)).

BTW, it helps automation tools.

Sometimes, this base-commit is useless for the reviewer workflow but
having it does not interfere.  Having an information does not mean it
must be used.  However, not having an information implies it cannot be
used. ;-)
--8<---------------cut here---------------end--------------->8---

#: <http://issues.guix.gnu.org/issue/43946#5>
@: <http://issues.guix.gnu.org/issue/43946#6>


Seeing the number of patches in the patch tracker, most of them does not
apply anymore.  It is not encouraging to review and decrease the queue
when first the reviewer has to guess onto which commit the patch
applies.


Cheers,
simon

bug#43946: [PATCH] doc: Add item to "Submitting Patches" section.

[Ludovic Courtès]

Hi,

zimoun <zimon.toutoune@gmail.com> skribis:

> Seeing the number of patches in the patch tracker, most of them does not
> apply anymore.  It is not encouraging to review and decrease the queue
> when first the reviewer has to guess onto which commit the patch
> applies.

Yeah in hindsight I think I agree with you.  I went ahead and (finally!)
applied the patch.

Thanks for persevering, and apologies for my stubbornness!  :-)

Ludo’.