v2SubmittingPatches: release-notes entry experiment

The "What's cooking" report lists the topics in flight, with a short paragraph descibing what they are about.

Once written, the description is automatically picked up from the "What's cooking" report and used in the commit log message of the merge commit when the topic is merged into integration branches. These commit log messges of the merge commits are then propagated to the release notes.

It has been the maintainer's task to prepare these entries in the "What's cooking" report. Even though the original author of a topic may be in the best position to write the initial description of a topic, we so far lacked a formal channel for the author to suggest what description to use. The usual procedure has been for the author to see the topic described in "What's cooking" report, and then either complain about inaccurate explanation and/or offer a rewrite.

Let's try an experiment to optionally let the author propose the one paragraph description when the topic is submitted. Pick the cover letter as the logical place to do so, and describe an experimental workflow in the SubmittingPatches document.

Signed-off-by: Junio C Hamano <gitster@pobox.com>
---
 * An experimental procedure for a topic author to propose the topic
   description to be used in "What's cooking" report and in the
   release notes have been added to the SubmittingPatches document.

 The above is an example that follows this protocol for a
 single-patch series.

    >> Would it be beneficial to request some specific heading, phrase, or
    >> other structured text such that this summary is obvious, or even easily
    >> extracted with some sort of script? Or is that perhaps overkill for now?
    >
    > ... the rule might end up
    > to be as simple as "When the first paragraph of the message looks
    > like an entry in the Release Notes, it is used as such".

Range-diff:
1:  83f8b69ab9 ! 1:  86b861255b SubmittingPatches: release-notes entry experiment
      ## Documentation/SubmittingPatches ##
     @@ Documentation/SubmittingPatches: an explanation of changes between each iteration can be kept in
    @@ Documentation/SubmittingPatches: an explanation of changes between each iteratio
     +paragraph summary that appears in the "What's cooking" report when it
     +is picked up to explain the topic.  If you choose to do so, please
     +write 2-5 lines of a paragraph that will fit well in our release notes
    -+(see Documentation/RelNotes/* directory for examples), and put it in
    -+the cover letter, clearly marked as such.  For a single-patch series,
    ++(see Documentation/RelNotes/* directory for examples), and make it
    ++the first paragraph of the cover letter.  For a single-patch series,
     +use the space between the three-dash line and the diffstat, as
     +described earlier.
     +

 Documentation/SubmittingPatches | 11 +++++++++++
 1 file changed, 11 insertions(+)

Hi Junio

On Mon, Mar 25, 2024 at 5:21 PM Junio C Hamano <gitster@pobox.com> wrote:

One very minor grammar note: "you can propose *a* one paragraph summary".

Otherwise, this patch looks good to me. Thanks for considering this.

"Brian Lyles" <brianmlyles@gmail.com> writes:

Oh, yeah, of course.  Thanks for carefully reading.

On 2024-03-26 00:37, Brian Lyles wrote:

Actually, it should read "a one-paragraph summary", to be precise. :)

On 2024-03-25 23:21, Junio C Hamano wrote:

Looking good to me.

Reviewed-by: Dragan Simic <dsimic@manjaro.org>

Hi Junio

On 25/03/2024 22:21, Junio C Hamano wrote:

Maybe "please write a 2-5 line paragraph"?

I think this is a good idea - one question though, how do you want patch authors to indicate that the first paragraph should be used as the summary?

Best Wishes

Phillip

Phillip Wood <phillip.wood123@gmail.com> writes:

Very true.

I want to start this as a light-weight process for contributors, and leave the automation for later, because we do not know how this will be useful in practice. We may end up talking in inconsistent voices if the author-supplied summary is used verbatim, so automation has its limit---the result always need to be copy-edited.

In an case, taking an example from what eventually became 9187b276 (Merge branch 'pw/diff-no-index-from-named-pipes', 2023-07-17), let's illustrate how the current process works and the proposed new process would have worked.

The commit log message of the topic reads:

    Merge branch 'pw/diff-no-index-from-named-pipes'

    "git diff --no-index" learned to read from named pipes as if they
    were regular files, to allow "git diff <(process) <(substitution)"
    some shells support.

    * pw/diff-no-index-from-named-pipes:
      diff --no-index: support reading from named pipes
      t4054: test diff --no-index with stdin
      diff --no-index: die on error reading stdin
      diff --no-index: refuse to compare stdin to a directory

The three-line paragraph summary were written by me back then first in the draft of "What's cooking" being prepared when the topic was first merged to 'seen', and then the integration process [*1*] copied the description to the merge commit message. Such a merge commit with the summary are made every time the integration cycle runs, including the time the topic gets merged to 'next' and more importantly to 'master', at which point, it also gets distributed into sections of the draft version of RelNotes.

The topic is listed in the release notes for Git 2.42 as one of the "UI, Workflows & Features":

    * "git diff --no-index" learned to read from named pipes as if they
      were regular files, to allow "git diff <(process) <(substitution)"
      some shells support.

Its cover letter <cover.1688586536.git.phillip.wood@dunelm.org.uk> started like so:

    In some shells, such as bash and zsh, it's possible to use a command
    substitution to provide the output of a command as a file argument to
    another process, like so:

      diff -u <(printf "a\nb\n") <(printf "a\nc\n")

    However, ...

but you could have started it like so:

    * "git diff --no-index" learned to read from named pipes as if they
      were regular files, to allow "git diff <(process) <(substitution)"
      some shells support.

    In some shells, such as bash and zsh, it's possible to use a command
    substitution to provide the output of a command as a file argument to
    another process, like so:
    ...

and I suspect it would be sufficient to notice that the paragraph wants to be the topic description. We may even feed it to automation if we decide to do so later [*2*]. Until then we

 - identify three-place indented first paragraph that is 2-5 lines
   long whose first line is indented with " * ";

 - somehow use it when adding the topic to "What's cooking" draft;

and then the integration process merges the topic to 'seen' and uses it in the merge commit log message. We *can* still copy-edit what I keep in the draft of "What's cooking" which I send to the list about twice a week. When the topic eventually hits 'master', the integration process would extract these merge log messages from "git log --first-parent master" output for the batch, and I rearrange them into sections of RelNotes, while doing the final proofreading.

[Footnotes]

 *1* The Reintegrate script and the other files from the 'todo'
     branch of my git repository are checked out in an untracked
     Meta subdirectory of my primary working area for Git
     development.  It knows how to take topic descriptions from the
     draft of "What's cooking" (also checked out in Meta/) among
     other tricks.

 *2* Teaching "am" to do something useful with the cover letters is
     something I have been wanting to do for quite some time.  Ideas
     other than the topic description we are disussing here include
     allowing the patch submitter to pick a branch name for the
     topic and creating an empty commit at the tip (not the bottom)
     of the topic branch that records the contents of the cover
     letter.