Le dimanche 30 décembre 2018 à 05:14:34+0000, Nick Morrott a écrit :
> On Sun, 30 Dec 2018 at 02:12, Pierre-Elliott Bécue <peb@debian.org> wrote:
> >
> > Le dimanche 30 décembre 2018 à 02:01:28+0100, Pierre-Elliott Bécue a écrit :
> > > [snip]
> >
> > ## python-project-generator-definitions: uploaded
> >
> > 1) I uploaded a bit fast. It's not a big issue but could you consider
> > removing the explicit python3 dependencies from the binary package's
> > Depends field? Normally debhelper finds the appropriate list from
> > python3:depends meta-dependency, using install_requires field of
> > setup.py.
>
> Updated
>
> > ## python-project-generator:
> >
> > 1) I wonder why you removed the alternative entry point.
>
> I dropped it to be consistent with
> python3-project-generator-definitions which only provides a single
> entrypoint in the same "namespace". I also thought the longer
> entrypoint name was somewhat cumbersome.
>
> > 2) Please remove all specific entries for the binary package Depends,
> > except python3-project-generator-definitions (<< 0.3.0):
> > python3:depends drags the others properly.
>
> Updated
Uploaded, thanks!
> > ## valinor:
> >
> > 1) You should ask upstream to provide a changelog, even though your
> > changelog.upstream.md is a perfectly fine temporary solution.
>
> https://github.com/ARMmbed/valinor/issues/29
>
> > 2) Same about the Depends field of this package, keep gdb of course, as
> > no :depends entry will drag it.
>
> Updated
>
> > 3) d/watch: could you use version=4 here or is there an issue?
>
> Bumped to version 4 (I had cribbed from another d/watch file using pypi)
>
> > 4) Is there a good reason for your export PYBUILD_AFTER_INSTALL=mv
> > {destdir}/usr/bin/valinor {destdir}/usr/share/valinor/run in d/rules
> > plus the link for /usr/bin?
>
> The package build attempts to install the script "valinor" into
> /usr/share/valinor/valinor/, which it can't because of the naming
> conflict. This occurs in a few of the other packages too I think.
Ack, I'm surprised by the mentioned behaviour but in any case, this choice
you made isn't irrelevant so let's stick to it for now.
If you find some time to investigate what causes such behaviour, please give
me some feedback!
> > ## python-mbed-ls:
> >
> > 1) Same as the others regarding the Depends field of the binary package.
>
> Updated
>
> > 2) For the Doxygen patch, you removed the genlatex, I guess it's because
> > you don't want to drag texlive for the building part? I'm fine with
> > this.
>
> Yes
I made a small mistake, you had to keep python3-pkg-resources and
python3-colorlog here, but I forgot to mention them. I made a fixing commit.
Uploaded.
> > ## python-mbed-host-tests:
> >
> > 1) Same as the others regarding the Depends field, prospectively keeping
> > the versioned dependencies
>
> Updated
I restored python3-pkg-resources as a dependency. I missed that one.
Uploaded.
> > ## mbed-test-wrapper:
> >
> > 1) Still the same Depends question. :) keep binutils-arm-none-eabi, of
> > course.
>
> In this instance, the source setup.py file does not declare anything
> in install_requires and so the dependencies must be added manually.
>
> I've asked upstream to consider adding explicit dependencies for
> mbed-ls and mbed-host-tests:
>
> https://github.com/autopulated/mbed-test-wrapper/issues/3
Ack, my bad, I must've been a little tired, I missed the setup.py emptyness.
Uploaded.
> > ## python-hgapi
> >
> > 1) You should suggest your patches to upstream.
>
> The patches were take from upstream's repository and links to the PRs
> are in the patch headers.
Ah, good to know!
> > ## yotta
> >
> > 1) Same old Depends field. You must keep valinor and mbed-test-wrapper at
> > least for now (I think that they'll be matched by the dh magic when
> > they'll be in the archive).
>
> Updated
>
> > 2) I wonder why you added python3-openssl and python3-idna to the
> > dependencies, they seem not needed according to setup.py? If you need
> > them, please add them manualy to the Depends field as they won't be
> > dragged by the python3:depends variable. In that case maybe it would be
> > worth do att a README.source in the debian/ directory explaining why
> > these packages are needed. If upstream forgot to add these dependencies,
> > please notify them with a bug report.
>
> I *think* I originally added them because they are suggested deps of
> requests. I can't see anything in the source that suggests they are
> needed, so I will drop them.
Ack, and uploaded, too.
> > Summary:
> >
> > > > -> yotta: x
> > > > -> python-hgapi: ./
> > > > -> mbed-test-wrapper: x
> > > > -> python-mbed-host-tests: x
> > > > -> python-mbed-ls: x
> > > > -> valinor: x
> > > > -> python-project-generator: x
> > > > -> python-project-generator-definitions: ./ (!)
> >
> > The other packages will have to wait, I'd rather finish this batch first.
>
> That's awesome. Thank you for the review.
>
> I've pushed changes to salsa, (hopefully) addressing all of the above
> issues. Please let me know if I've missed any or if any follow up is
> required.
I think it's all good now. Sorry for my own mistakes.
Summary: all uploaded, some already got through new, the others should
follow.
I shall give some time to other python packages waiting for sponsorship,
then I'll come back at you for the second batch.
> > As my remarks are more like nitpicking, I'd like to insist on the quality of
> > your work. Those packages are really neat.
> >
> > Thanks for your great work!
>
> Thanks for your kind words,
You're very welcome.
--
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528 F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.
Attachment:
signature.asc
Description: PGP signature