Re: RFS: mu-editor and yotta dependencies

2018-12-30 Thread Nick Morrott
On Sun, 30 Dec 2018 at 18:42, Pierre-Elliott Bécue  wrote:
>
> Re,
>
> Actually, nudatus is a dependency of uflash.

See below for my reasoning for adding python3-nudatus as a Recommends
for python-uflash, rather than a hard Depends.

> ## python-nudatus
>
> Uploaded.

Awesome

> ## python-uflash
>
> Uploaded, though installing it will be hard, as
> firmware-microbit-micropython is missing and apt tries to fetch the
> recommends too.

nudatus is really an optional dependency of uflash, as minification is
not essential. The uflash code checks to see if nudatus is available
at runtime, and will disable minify if it is not available.

> ## python-guizero
>
> Uploaded.

Awesome

> ## mu-editor
>
> Hmmm.
>
> In the dependencies of the project are mentionned and imported:
>
> gpiozero

Please see the request to provide the gpiozero package on non-armhf
architectures at
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=856413

> pigpio

Please see the RFP discussion for packaging pigpio at
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=908787

> pynsist

pynsist is a tool to build Windows installers

> + some deps useful for testing.

i) pytest-cov measures test coverage, and I am not sure if that's a
useful statistic for packing?

ii) pytest-random-order would be useful, especially for reproducibility.

*However*, since upstream Mu added the dependency back in 2018/03,
upstream pytest-random-order 1.0.0+ disabled its "random by default"
behaviour (2018/10). As such, the mu-editor tests are not currently
randomised. This is a testing regression and I will report it, so that
the forthcoming 1.0.2 release of Mu contains it.

pytest-random-order is not actually packaged yet for Debian, so I could ITP it.

> I don't see any of those three packages in debian right now. Is there a
> reason you consider we don't need these?

I have added notes about gpiozero/pigpio to d/control. pynsist does
not seem necessary.

Are there any others I have missed?

Thanks,
Nick



Re: RFS: mu-editor and yotta dependencies

2018-12-30 Thread Pierre-Elliott Bécue
Re,

Actually, nudatus is a dependency of uflash.

## python-nudatus

Uploaded.

## python-uflash

Uploaded, though installing it will be hard, as
firmware-microbit-micropython is missing and apt tries to fetch the
recommends too.

## python-guizero

Uploaded.

## mu-editor

Hmmm.

In the dependencies of the project are mentionned and imported:

gpiozero
pigpio
pynsist

+ some deps useful for testing.

I don't see any of those three packages in debian right now. Is there a
reason you consider we don't need these?

-- 
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.


signature.asc
Description: PGP signature


Re: RFS: mu-editor and yotta dependencies

2018-12-30 Thread Pierre-Elliott Bécue
Le dimanche 30 décembre 2018 à 05:14:34+, Nick Morrott a écrit :
> On Sun, 30 Dec 2018 at 02:12, Pierre-Elliott Bécue  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 

Re: Please add my DD account to DPMT

2018-12-30 Thread Piotr Ożarowski
> Please add my DD account to DPMT group on salsa: vvidic
done :)



Please add my DD account to DPMT

2018-12-30 Thread Valentin Vidic
Hello,

Please add my DD account to DPMT group on salsa: vvidic

I am already a member as vvidic-guest and agree to follow
the team policy.

-- 
Valentin


signature.asc
Description: PGP signature