#1105004 resurrect README patch with info requested by upstream and by users

Package:
php-elisp
Source:
php-elisp
Submitter:
Nicholas D Steeves
Date:
2025-05-10 22:51:01 UTC
Severity:
normal
Tags:
#1105004#5
Date:
2025-05-09 19:13:40 UTC
From:
To:
Hi Manphiz,

When I was reviewing your work I noticed that you appear to have been
unable to rebase a patch from 1.25.1 onto 1.27.0.  This should take
less than five minutes with ediff (gbp pq workflow), and there have
been several discussions both on upstream's tracker and privately
about how the info in that patch is needed.  We have many years of
good rapport with upstream, and this patch was part of that
understanding and collaboration.

I'd be happy to fix up the patch header to be more informative, and I
concede that I missed this important step when I orphaned the package.

Regards,
Nicholas

#1105004#12
Date:
2025-05-10 00:32:58 UTC
From:
To:
Nicholas D Steeves <sten@debian.org> writes:

The patch only touches README.md, so I fail to see it as "a bug which
has a major effect on the usability of a package".  Changed to minor.

Do you have any reference to the discussions? I saw it marked as
"Forwarded: not-needed" and the changes seem to be Debian specific,
which seems possible that upstream may not be aware of it.

Adding links to the discussions would help.

Also, this is still a team maintained package, so feel free to make
changes as you see fit.

#1105004#19
Date:
2025-05-10 15:22:43 UTC
From:
To:
Xiyue Deng <manphiz@gmail.com> writes:

Right, that's in the subject of this bug.  Is your bias that our users
don't read README.md, and/or that documentation isn't important?

When users follow the installation instructions in the documentation and
it renders the package unusable or breaks non-related packages, then
it's more severe than "minor".  This has happened disproportionately
often with this package compared to others.  "Minor" also means trivial
to fix, which begs that question: why did you create a problem that
would have been trivial to avoid?
If in doubt, you should ask your teammates.

Help with what?  The old patch addresses the issues raised both upstream
and in private.

As the maintainer of this package, you are now responsible to our users,
and to upstream.

Also, in my checklist of your competencies, you're still missing a
couple of categories.  Being able to do a three-way merge (I recommend
ediff) and maintain a patch is one of them.  Maintaining good relations
with users and upstreams are two others.

Best,
Nicholas

#1105004#26
Date:
2025-05-10 16:06:11 UTC
From:
To:
Nicholas,

You could offer your feedback using a much more constructive tone than
the one you took in this message.

This, in particular, was not a nice thing to say at all.  Clearly Xiyue
is capable of doing those things.  But more importantly, just telling
someone they don't have certain capacities can only have negative
outcomes.  Instead, tell them how you think they could do better.

#1105004#31
Date:
2025-05-10 22:49:15 UTC
From:
To:
Nicholas D Steeves <sten@debian.org> writes:

I'm following the bug severity level suggestions[1].  The patch only
touches documents without any code change, which I think fits the
category of "minor": a problem which doesn't affect the package's
usefulness, and is presumably trivial to fix.

Also, I see that you changed the severity to "normal", so I think you
agree it's not "important".

I found your arguments vague and exaggerating without any reference.
Let's first make sure we are discussing the patch I dropped in
commit[2].  The patch mainly does 2 things: replaces upstream
instruction to install through ELPA with Debian instructions (with a
suggest prefer version in sid); and a new instruction to report bug in
the Debian way.

I have no concern with the second change.  By the first change, Are you
suggesting that installing php-mode through ELPA breaks users in anyway?
In my experience, installing addons both from ELPA and Debian work fine,
even if there were dependency relationships.  Also suggesting using sid
on stable Debian systems creates FrankenDebian, which I believe a senior
DD like you know is a bad thing.  If a newer upstream version fixes any
issues, and the newer version will not be backported to the stable
distribution, the better way is to install the newer version through
ELPA, instead of configuring a FrankenDebian.  The is also an important
reason I didn't try to rebase the patch.

And where is the link referencing the discussion?

OK.

Adding a link to the upstream issue you mentioned would help me and
other maintainers understand your concern better.

The maintainer is the Emacsen team, but yes, I do have responsibilities
as you mentioned, with help from other team members.

You assume that I dropped the patch because I didn't know how to do a
3-way merge.  I gave my reason above.

[1] https://www.debian.org/Bugs/Developer#severities
[2] https://salsa.debian.org/emacsen-team/php-elisp/-/commit/08630e52c7ee41c022522543d127af42d3afe5c2