[bug#47171,00/10] Move some Bioconductor packages to (gnu packages bioconductor).

Message ID 20210315180714.8233-1-zimon.toutoune@gmail.com
Headers show
Series Move some Bioconductor packages to (gnu packages bioconductor). | expand

Message

Simon Tournier March 15, 2021, 6:07 p.m. UTC
Hi,

Janitor [1] patch set.  It moves Bioconductor packages from (gnu packages
bioinformatics) to (gnu packages bioconductor).  Far to be complete; still 4
pacth set like this one are coming...

1: <https://lists.gnu.org/archive/html/guix-commits/2020-09/msg00416.html>


To transfer the Copyright, I use:

   git --no-pager log --date=format:"%Y" --format="%ad %an %s" --grep=r-<name>

Then, after the move, I run:

   ./pre-inst-env guix build r-<name> --no-grafts
   ./pre-inst-env guix build r-<name> --no-grafts --check

and at the end of the series, I run:

    make as-derivation

to check that "guix pull" is not broken; just in case.

Last, I have not carefully checked the imported modules, so it could
be possible to perhaps remove couple of #:use-module (gnu packages <useless>)
in gnu/packages/bioinformatics.scm.

All the best,
simon

zimoun (10):
  gnu: r-limma: Move to (gnu packages bioconductor).
  gnu: r-xvector: Move to (gnu packages bioconductor).
  gnu: r-genomicranges: Move to (gnu packages bioconductor).
  gnu: r-biobase: Move to (gnu packages bioconductor).
  gnu: r-annotationdbi: Move to (gnu packages bioconductor).
  gnu: r-biomart: Move to (gnu packages bioconductor).
  gnu: r-biocparallel: Move to (gnu packages bioconductor).
  gnu: r-biostrings: Move to (gnu packages bioconductor).
  gnu: r-rsamtools: Move to (gnu packages bioconductor).
  gnu: r-delayedarray: Move to (gnu packages bioconductor).

 gnu/packages/bioconductor.scm   | 305 ++++++++++++++++++++++++++++++++
 gnu/packages/bioinformatics.scm | 305 --------------------------------
 2 files changed, 305 insertions(+), 305 deletions(-)


base-commit: 266d55dc3080475544bf45e72359c9b9bbcecd53

Comments

Ricardo Wurmus March 15, 2021, 10:33 p.m. UTC | #1
Hi,

> zimoun (10):
>   gnu: r-limma: Move to (gnu packages bioconductor).
>   gnu: r-xvector: Move to (gnu packages bioconductor).
>   gnu: r-genomicranges: Move to (gnu packages bioconductor).
>   gnu: r-biobase: Move to (gnu packages bioconductor).
>   gnu: r-annotationdbi: Move to (gnu packages bioconductor).
>   gnu: r-biomart: Move to (gnu packages bioconductor).
>   gnu: r-biocparallel: Move to (gnu packages bioconductor).
>   gnu: r-biostrings: Move to (gnu packages bioconductor).
>   gnu: r-rsamtools: Move to (gnu packages bioconductor).
>   gnu: r-delayedarray: Move to (gnu packages bioconductor).

Thanks, I’ve pushed them all.
Leo Famulari March 15, 2021, 11:56 p.m. UTC | #2
On Mon, Mar 15, 2021 at 11:33:11PM +0100, Ricardo Wurmus wrote:
> Hi,
> 
> > zimoun (10):
> >   gnu: r-limma: Move to (gnu packages bioconductor).
> >   gnu: r-xvector: Move to (gnu packages bioconductor).
> >   gnu: r-genomicranges: Move to (gnu packages bioconductor).
> >   gnu: r-biobase: Move to (gnu packages bioconductor).
> >   gnu: r-annotationdbi: Move to (gnu packages bioconductor).
> >   gnu: r-biomart: Move to (gnu packages bioconductor).
> >   gnu: r-biocparallel: Move to (gnu packages bioconductor).
> >   gnu: r-biostrings: Move to (gnu packages bioconductor).
> >   gnu: r-rsamtools: Move to (gnu packages bioconductor).
> >   gnu: r-delayedarray: Move to (gnu packages bioconductor).
> 
> Thanks, I’ve pushed them all.

These commits caused `guix pull` to fail, as described here:

https://git.savannah.gnu.org/cgit/guix.git/commit/?id=8ec0ca8faff62f19426f22aeb1bd59a8950ca05a

I think we just needed to import the bioconductor module in the cran
module, but I wasn't sure so I reverted them in order to fix the problem
quickly.
Simon Tournier March 16, 2021, 12:27 p.m. UTC | #3
Hi Leo,

On Tue, 16 Mar 2021 at 00:56, Leo Famulari <leo@famulari.name> wrote:
> On Mon, Mar 15, 2021 at 11:33:11PM +0100, Ricardo Wurmus wrote:

> > > zimoun (10):
> > >   gnu: r-limma: Move to (gnu packages bioconductor).
> > >   gnu: r-xvector: Move to (gnu packages bioconductor).
> > >   gnu: r-genomicranges: Move to (gnu packages bioconductor).
> > >   gnu: r-biobase: Move to (gnu packages bioconductor).
> > >   gnu: r-annotationdbi: Move to (gnu packages bioconductor).
> > >   gnu: r-biomart: Move to (gnu packages bioconductor).
> > >   gnu: r-biocparallel: Move to (gnu packages bioconductor).
> > >   gnu: r-biostrings: Move to (gnu packages bioconductor).
> > >   gnu: r-rsamtools: Move to (gnu packages bioconductor).
> > >   gnu: r-delayedarray: Move to (gnu packages bioconductor).
> >
> > Thanks, I’ve pushed them all.
>
> These commits caused `guix pull` to fail, as described here:
>
> https://git.savannah.gnu.org/cgit/guix.git/commit/?id=8ec0ca8faff62f19426f22aeb1bd59a8950ca05a
>
> I think we just needed to import the bioconductor module in the cran
> module, but I wasn't sure so I reverted them in order to fix the problem
> quickly.

Sorry for the annoyance.  I did "make as-derivation" and it reports
nothing.  But indeed, "./pre-inst-env guix pull -p /tmp/new
--url=$(pwd) --disable-authentication --branch=test" breaks.
I do not know if I do something wrong (run make in a worktree for the branch).

Well, thanks for the report and I confirm that adding the bioconductor
import in cran fixes "guix pull".   But in my memories, this import
was adding circular dependency.  Ricardo, WDYT?


Cheers,
simon
Ricardo Wurmus March 16, 2021, 1:20 p.m. UTC | #4
zimoun <zimon.toutoune@gmail.com> writes:

> Well, thanks for the report and I confirm that adding the bioconductor
> import in cran fixes "guix pull".   But in my memories, this import
> was adding circular dependency.  Ricardo, WDYT?

The point of having (gnu packages cran) and (gnu packages bioconductor)
be separate modules is to separate independent packages.

I’d like to keep it that way and not let (gnu packages cran) depend on
(gnu packages bioconductor).
Simon Tournier March 16, 2021, 1:30 p.m. UTC | #5
Hi,

On Tue, 16 Mar 2021 at 14:20, Ricardo Wurmus <rekado@elephly.net> wrote:
> zimoun <zimon.toutoune@gmail.com> writes:
>
> > Well, thanks for the report and I confirm that adding the bioconductor
> > import in cran fixes "guix pull".   But in my memories, this import
> > was adding circular dependency.  Ricardo, WDYT?
>
> The point of having (gnu packages cran) and (gnu packages bioconductor)
> be separate modules is to separate independent packages.
>
> I’d like to keep it that way and not let (gnu packages cran) depend on
> (gnu packages bioconductor).

Therefore, the fix is to move the CRAN r-bisquerna and r-absfiltergsea
packages from (gnu packages cran) to (gnu packages bioconductor)
because they depend on the Bioconductor package r-biobase (and
r-limma).  Right?
How do I proceed?  Submit a patch moving these 2 packages and then
another patch reverting the revert?

Thanks,
simon
Ricardo Wurmus March 16, 2021, 4:08 p.m. UTC | #6
zimoun <zimon.toutoune@gmail.com> writes:

> Hi,
>
> On Tue, 16 Mar 2021 at 14:20, Ricardo Wurmus <rekado@elephly.net> wrote:
>> zimoun <zimon.toutoune@gmail.com> writes:
>>
>> > Well, thanks for the report and I confirm that adding the bioconductor
>> > import in cran fixes "guix pull".   But in my memories, this import
>> > was adding circular dependency.  Ricardo, WDYT?
>>
>> The point of having (gnu packages cran) and (gnu packages bioconductor)
>> be separate modules is to separate independent packages.
>>
>> I’d like to keep it that way and not let (gnu packages cran) depend on
>> (gnu packages bioconductor).
>
> Therefore, the fix is to move the CRAN r-bisquerna and r-absfiltergsea
> packages from (gnu packages cran) to (gnu packages bioconductor)
> because they depend on the Bioconductor package r-biobase (and
> r-limma).  Right?

Correct.  We don’t want any CRAN packages to depend on Bioconductor
things.  When this is unavoidable, though, we move them to (gnu packages
bioconductor) with a comment.

> How do I proceed?  Submit a patch moving these 2 packages and then
> another patch reverting the revert?

Sounds good.  (You could do this all in the same issue.)
Simon Tournier March 16, 2021, 4:36 p.m. UTC | #7
On Tue, 16 Mar 2021 at 17:08, Ricardo Wurmus <rekado@elephly.net> wrote:

> > Therefore, the fix is to move the CRAN r-bisquerna and r-absfiltergsea
> > packages from (gnu packages cran) to (gnu packages bioconductor)
> > because they depend on the Bioconductor package r-biobase (and
> > r-limma).  Right?
>
> Correct.  We don’t want any CRAN packages to depend on Bioconductor
> things.  When this is unavoidable, though, we move them to (gnu packages
> bioconductor) with a comment.

Ok.


> > How do I proceed?  Submit a patch moving these 2 packages and then
> > another patch reverting the revert?
>
> Sounds good.  (You could do this all in the same issue.)

By the same issue, do you mean the same guix-patches number?  Or the
same commit?


Cheers,
simon
Ricardo Wurmus March 16, 2021, 10:04 p.m. UTC | #8
zimoun <zimon.toutoune@gmail.com> writes:

>> > How do I proceed?  Submit a patch moving these 2 packages and then
>> > another patch reverting the revert?
>>
>> Sounds good.  (You could do this all in the same issue.)
>
> By the same issue, do you mean the same guix-patches number?  Or the
> same commit?

I meant the same guix-patches issue number; separate commits.
Simon Tournier March 17, 2021, 8:15 p.m. UTC | #9
Hi Ricardo,

On Tue, 16 Mar 2021 at 23:04, Ricardo Wurmus <rekado@elephly.net> wrote:

> I meant the same guix-patches issue number; separate commits.

Done in v2.  I have double-checked but who knows. :-)


Thanks,
simon