Bug#868977: Confirm that I'm able to verify the problem

[Yaroslav Halchenko]

On Thu, 27 Jul 2017, Andreas Tille wrote:

> > would need to dig in, may be later, meanwhile pushed my slight tune ups, 
> > feel
> > welcome to fix up.  You do know where to get the .orig.tar.gz ;)

> Sure I know and I used it for building.  What about simplifying things
> for your team mates and follow policy which does not require them
> downloading the source manually? :-P

> As I said, I would have added pristine-tar if this would not have messed
> up the repository for reasons I do not understand.  It would be really
> cool if you would consider this for next version.

apparently it is a problem in recent pandas, which isn't yet resolved or
worked around in statsmodels
https://github.com/statsmodels/statsmodels/issues/3841#issuecomment-318630239
so we might just need to wait a bit more

-- 
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
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

Bug#868977: Confirm that I'm able to verify the problem

[Andreas Tille]

Hi Yaroslav,

On Thu, Jul 27, 2017 at 10:21:35AM -0400, Yaroslav Halchenko wrote:
> 
> would need to dig in, may be later, meanwhile pushed my slight tune ups, feel
> welcome to fix up.  You do know where to get the .orig.tar.gz ;)

Sure I know and I used it for building.  What about simplifying things
for your team mates and follow policy which does not require them
downloading the source manually? :-P

As I said, I would have added pristine-tar if this would not have messed
up the repository for reasons I do not understand.  It would be really
cool if you would consider this for next version.

Kind regards

  Andreas.

-- 
http://fam-tille.de

Bug#868977: Confirm that I'm able to verify the problem

[Andreas Tille]

Hi Yaroslav,

On Thu, Jul 27, 2017 at 09:44:10AM -0400, Yaroslav Halchenko wrote:
> Is there any demand/requirement for debhelper 10 for this package?
> Could we settle on 9 (to ease backportability to old jessie and alike)?

Just to answer this question:

  1. There is no demand - I just choose the latest compat level.
  2. debhelper 10 is backported to Jessie

Feel free to override my decision.

Kind regards

   Andreas. 

-- 
http://fam-tille.de

Bug#868977: Confirm that I'm able to verify the problem

[Yaroslav Halchenko]

On Thu, 27 Jul 2017, Andreas Tille wrote:

> Hi,

> I was hoping that taking over the version from experimental to unstable
> might help to fix this bug but I can only confirm that I can reproduce

eh -- plain rebuild doesn't suffice -- tests still fail
and some new failures happen:


==
ERROR: statsmodels.tsa.vector_ar.tests.test_var.TestVARResults.test_pickle
--
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/nose/case.py", line 197, in runTest
self.test(*self.arg)
  File 
"/build/statsmodels-0.8.0/debian/python-statsmodels/usr/lib/python2.7/dist-packages/statsmodels/tsa/vector_ar/tests/test_var.py",
 line 463, in test_pickle
res_unpickled = self.res.__class__.load(fh)
  File 
"/build/statsmodels-0.8.0/debian/python-statsmodels/usr/lib/python2.7/dist-packages/statsmodels/base/wrapper.py",
 line 77, in load
return load_pickle(fname)
  File 
"/build/statsmodels-0.8.0/debian/python-statsmodels/usr/lib/python2.7/dist-packages/statsmodels/iolib/smpickle.py",
 line 32, in load_pickle
return cPickle.load(fin)
ValueError: non-string names in Numpy dtype unpickling


and possibly more

would need to dig in, may be later, meanwhile pushed my slight tune ups, feel
welcome to fix up.  You do know where to get the .orig.tar.gz ;)
but just in case
http://http.debian.net/debian/pool/main/s/statsmodels/statsmodels_0.8.0.orig.tar.gz

-- 
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
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

Bug#868977: Confirm that I'm able to verify the problem

[Yaroslav Halchenko]

On Thu, 27 Jul 2017, Yaroslav Halchenko wrote:

> Is there any demand/requirement for debhelper 10 for this package?
> Could we settle on 9 (to ease backportability to old jessie and alike)?

I would guess that the answers are no and yes, fixed up, building
now, will upload whenever done

Cheers and thanks!
-- 
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
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


signature.asc
Description: PGP signature

Bug#868977: Confirm that I'm able to verify the problem

[Yaroslav Halchenko]

On July 27, 2017 9:23:16 AM EDT, Andreas Tille  wrote:
>Hi,
>
>I was hoping that taking over the version from experimental to unstable
>might help to fix this bug but I can only confirm that I can reproduce
>the problem here.  So it also happens for version 0.8.0.  I've commited
>my attempt to Git and hope that somebody else might have a better clue.
>
>BTW, I failed to use gbp to build the package since the repository is
>lacking a pristine-tar branch and when I tried to import the existing
>tarball I've got so many conflicts with the Git repository that I
>simply
>gave up.
>
>Kind regards
>
>  Andreas.

I can do it in 10 minutes.

Bug#868977: Confirm that I'm able to verify the problem

[Yaroslav Halchenko]

On Thu, 27 Jul 2017, Andreas Tille wrote:

> Hi,

> I was hoping that taking over the version from experimental to unstable
> might help to fix this bug but I can only confirm that I can reproduce
> the problem here.  So it also happens for version 0.8.0.  I've commited
> my attempt to Git and hope that somebody else might have a better clue.

> BTW, I failed to use gbp to build the package since the repository is
> lacking a pristine-tar branch and when I tried to import the existing
> tarball I've got so many conflicts with the Git repository that I simply
> gave up.

Is there any demand/requirement for debhelper 10 for this package?
Could we settle on 9 (to ease backportability to old jessie and alike)?

-- 
Yaroslav O. Halchenko
Center for Open Neuroscience http://centerforopenneuroscience.org
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

Bug#868977: Confirm that I'm able to verify the problem

[Andreas Tille]

Hi,

I was hoping that taking over the version from experimental to unstable
might help to fix this bug but I can only confirm that I can reproduce
the problem here.  So it also happens for version 0.8.0.  I've commited
my attempt to Git and hope that somebody else might have a better clue.

BTW, I failed to use gbp to build the package since the repository is
lacking a pristine-tar branch and when I tried to import the existing
tarball I've got so many conflicts with the Git repository that I simply
gave up.

Kind regards

  Andreas.

-- 
http://fam-tille.de