[Debian-med-packaging] Bug#768730: Bug#768730: python-mne REMOVED from testing

[Yaroslav Halchenko]

On Thu, 11 Dec 2014, Andreas Tille wrote:

> Hi,

> I have not fully understood your conversation whether release team was
> contacted about 0.8.6 or not.  I also failed to understand the issue
> with failed tests in connection with pandas.  So I do not know whether
> you expect me to do something after I'm slowly recovering to normal
> operation after vacation.

> On Wed, Dec 10, 2014 at 10:21:59PM +0100, Alexandre Gramfort wrote:
> > hi,

> > thanks Yarik for your help make v0.8.6 available on neurodebian.

> > Regarding the patch I sent it via email as I don't know the procedure.
> > Currently the master branch is packaged and it's already at 0.8.6.
> > I am not sure if I should branch at old version, commit the patch,
> > and push in maint/0.8.4 branch or something.

> I'd recommend a branch called jessie featuring the very minimum changes
> in a quilt patch against the version that was in Jessie before it was
> removed.  The upload should be discussed in advance with
> debian-release at lists.debian.org.  I personally think that chances are
> very low to convince the release team to care for a low popcon package.
> An extra obstactle is the newer version in unstable so we can not really
> simply upload to unstable without an epoch.  I'm personally tempted to
> give up with python-mne in Jessie and hope that all involved people
> have learned the lession that a removal warning is an honst issue and
> we need to follow the release policy in a short time frame.

I have had discussed with debian-release on IRC about feasibility of
fixing the RC on top of the 0.8.6 in sid.  Here is the transcript:

09:14  yoh: I am to fix up a RC (tests failures) for python-mne package recently removed from jessie -- would it still be ok/allow python-mne to migrate?
09:16  jmw: yoh: if it wasn't removed so long ago that we go 'waaah', yes
09:16  KiBi: 2d-ish
09:16  yoh: yeah -- just a week or so ago.  Thanks!
09:16  jmw: oh, fine
09:17  yoh: another question: silly me already uploaded to sid a new upstream release (yet to be patched)  which had docs + 1 non-RC bug fix. diff is really small:  https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=768730#45
09:17  yoh: could I still go through sid or should prepare upload to testing-proposed-updates?
09:18  jmw: sid, please
09:18  jmw: assuming there's nothing in the way
09:18  jmw: er, I should read your question first..
09:18  yoh: jmw: [reply cut] ;)
09:19  jmw: speaking for myself, I'd take the small additional change and go via sid. others may vary.
09:20  jmw: but please make sure to include details in the unblock request
09:21  yoh: sure -- will do.  Here is my perspective (prepared for t-p-u  but would prefer going through sid so ok to redo) fix:  http://www.onerussian.com/tmp/mne.patch
09:23  jmw: seems reasonable
09:23  yoh: cool.  thanks... I will upload to sid / send unblock shortly.  Cheers!

and it is only then I realized that sid also got pandas which
"interferes" (pandas version was uploaded before freeze but never made
it since some tests were failing on exotic archs -- so I am not at fault
father ;) )

The fix Alex pointed me to is also minimalistic so I will take 0.8.6 with prev
patch (to resolve issue of RO $HOME), file FTBFS report against sid, add that
patch for sid version, will upload to sid and request unblock.

So nothing to worry for anyone else about it if this all works ;)

-- 
Yaroslav O. Halchenko, Ph.D.
http://neuro.debian.net http://www.pymvpa.org http://www.fail2ban.org
Research Scientist,            Psychological and Brain Sciences Dept.
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik