On Fri, Apr 23, 2010 at 20:13, Julien Jehannet
<[email protected]> wrote:
>> * Aurélien Campéas <[email protected]> [23-avr-2010 19:48]:
>> Le vendredi 23 avril 2010 à 19:41 +0200, Sandro Tosi a écrit :
>> > On Wed, Mar 24, 2010 at 19:12, Julien Jehannet
>> > <[email protected]> wrote:
>> > >> * Sylvain Thénault <[email protected]> [22-mar-2010 12:18]:
>> > >> On 22 mars 11:58, Sandro Tosi wrote:
>> > >> > > I rather think the current test about distutils version is 
>> > >> > > inadequate.
>> > >> > > Because '--install-layout=dev' doesn't seem to exist for python 
>> > >> > > 2.5.2-3
>> > >> > > (lenny) but *does* for python 2.5.4-1 (distutils version still the 
>> > >> > > same).
>> > >> >
>> > >> > Yep, you're right: it was introduced in 2.5.3-1~exp1 so it might be
>> > >> > better to check python version than distutils one.
>> > >>
>> > >> Julien, could you do the proper fix in lgc'setup.py please ?
>> > > It's fixed in repo but I'm thinking about a new way now.
>> >
>> > The fix (as released in 0.50.0) is bugged:
>> >
>> > --- logilab-common~/setup.py    2010-04-20 11:00:32.000000000 +0200
>> > +++ logilab-common/setup.py     2010-04-23 19:37:02.697413120 +0200
>> > @@ -158,7 +158,7 @@
>> >      try:
>> >          if not USE_SETUPTOOLS:
>> >              # install-layout option was introduced in 2.5.3-1~exp1
>> > -            if sys.versioninfo < (2, 5, 4):
>> > +            if sys.version_info < (2, 5, 4):
>> >                  sys.argv.remove('--install-layout=deb')
>> >                  print "W: remove '--install-layout=deb' option"
>> >      except:
>> >
>> > because:
>> >
>> > $ python -c "import sys ; print sys.version_info"
>> > (2, 5, 5, 'final', 0)
>>
>> I may be a bit tired but I fail to see a bug here.
>>
>> FWTW:
>>
>> >>> (2, 5, 5, 'final', 0) < (2, 5, 4)
>> False
>> >>> (2, 5, 4) < (2, 5, 5, 'final', 0)
>> True
>
> Quite strange because it was already corrected but the tag
> logilab-common-version-0.50.0 was not set correctly in the hg repo.
>
> Whatever I'm not in favor of having specific distribution code into our
> setup.py. Moreover, I think that code should be drop out since it is now
> managed in debian/rules in using python.mk variables.

Agreed: that's something not for setup.py to mess with, but for the
distribution "stuff" to.

Regards,
-- 
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi
_______________________________________________
Python-Projects mailing list
[email protected]
http://lists.logilab.org/mailman/listinfo/python-projects

Reply via email to