Re: Guix Days: Patch flow discussion

[Kyle Meyer <kyle@kyleam.com>]

Hi Ricardo,

Ricardo Wurmus writes:

> Hi Josselin,

>> They both can co-exist with debbugs, and for now the patchwork instance
>> of QA is not usable for status tracking (because it is not meant to be
>> used as such for now).  One can already use both of them, but using both
>> supercedes debbugs, and gets rid of its limitations.  I've been using
>> b4/lei with the yhetil public-inbox instance, with piem.el as an
>> interface, and it's really useful.  With a properly configured b4, one
>> could simply run `b4 shazam some-msg-id` and it would automatically
>> apply the corresponding patchset.
>
> I’m interested in adopting this workflow.  Where can I find more
> information on how to configure this?

In 889a6204f8 (doc: Add some guidelines for reviewing, 2023-11-07),
Maxim added some b4 configuration to Guix's etc/git/gitconfig that
points at yhetil.org.  With that setup, running 'b4 shazam MESSAGE-ID'
from a Guix checkout should work.

You may have just be asking about configuring 'b4 shazam', but if you
(or others) are interested in configuring piem, read on :)


piem setup and use
==================

For piem, you need some custom setup in your Emacs config.

  https://docs.kyleam.com/piem/Registering-inboxes.html

piem-inboxes contains entries that provide info to help piem map from a
message (say in Notmuch or Gnus) to a public-inbox URL and to a local
checkout of the repo.

  (setq piem-inboxes
        '(("guix"
           :coderepo "/path/to/your/clone/of/guix/"
           :url "https://yhetil.org/guix/"
           :listid "bug-guix.gnu.org")))

Further setup depends on how (in Emacs) you read your messages.

  https://docs.kyleam.com/piem/Enabling-integration-libraries.html

For example, a Notmuch user would enable piem-notmuch-mode:

  (piem-notmuch-mode 1)

Or a debbugs.el user would enable piem-debbugs-mode (contributed by
Jelle Licht).  You can enable more than one integration library.

Once that's set up, the main entry point to b4 is the piem-b4-am
transient.  That transient in turn has the main command of interest,
piem-b4-am-from-mid.

  https://docs.kyleam.com/piem/Using-b4-to-apply-patches.html

Calling piem-b4-am-from-mid from, say, a Notmuch message with a patch
series will prepare the "am-ready" series, prompt you for the name of a
branch to create, and apply the series with 'git am' to the configured
repo.


piem-b4-am-from-mid vs b4 shazam
================================

piem-b4-am-from-mid (from Emacs) and 'b4 shazam' (from the shell
visiting Guix repo) are similar in spirit: use 'b4 am' to prepare an
am-ready mbox from a thread's full mbox and then use 'git am' apply it
to a repo.  A key difference is that, with piem, you start off in some
non-repository Emacs buffer containing a message and then piem knows how
to get to the repository (using the config above).

As a side note: 'b4 shazam' didn't exist when I wrote piem's b4
integration.  At some point, I plan to add 'b4 shazam' support to piem,
similar to how there is a piem command (piem-b4-am-ready-from-mid)
that's a direct interface to 'b4 am', without the extra handling of
piem-b4-am-from-mid.  This will allow callers that prefer to use shazam
invoke it from Emacs rather than the command line.


lei
===

There's also mention upthread of lei.  I don't think you're asking about
configuring that, but fwiw that's public-inbox's local command-line
client.  I gave a short (and incomplete) description of it here:

  https://yhetil.org/guix-devel/87a6izsoio.fsf@kyleam.com

piem currently has some basic support for lei, focused on searching
public-inbox instances and displaying the results in Emacs.