Actually there does seem to be an active discussion about this issue
at the moment: 

https://github.com/swig/swig/pull/7 

And it seems like
stack overflow has some answers:


http://stackoverflow.com/questions/8991520/python-conditional-module-object-has-no-attribute-error-with-personal-package?rq=1


Some documentation:


http://docs.python.org/2/faq/programming.html#how-can-i-have-modules-that-mutually-import-each-other


I don't have a system that actually exhibits the problem at the
moment, but it looks like 2.0.10 has at least been proposed for the next
Ubuntu release. 

My guess is something along the lines that the new
version of swig changes some imports to be absolute instead of relative
and thus we run into this namespace isn't fully initialized issue.


Thanks, 

Ali 

On 04.06.2013 14:16, Steve Reinhardt wrote: 

> I just
skimmed the swig bug descriptions on Sourceforge, and it sure sounds
>
like a swig bug to me. If that's true, and if there is no simple
>
workaround on our end, I'd support continuing to require working
versions
> of swig, even if that means that there's a period where
people can't just
> grab the default ubuntu package or whatever.
> 
> I
don't understand exactly what's not working with gem5 to say whether
>
there is a simple workaround or not.
> 
> Steve
> 
> On Tue, Jun 4, 2013
at 7:32 AM, Ali Saidi <[email protected]> wrote:
> 
>> I guess I don't
know if it's a bug or a feature. Two people have posted solutions to the
swig bug tracking system, but as soon as Ubuntu or Fedora start shipping
2.0.9 or 2.0.10 by default it's going to get real annoying really quick.
Ali On 04.06.2013 02:01, Andreas Hansson wrote: 
>> 
>>> Indeed, both
Andreas S and I stumbled on this (and the former
>> added a 
>> 
>>>
warning to the conscript for 2.0.9 hoping it was a bug that
>> would be
weeks ago if anyone was aware of a solution to the problem. So far I
>

> _______________________________________________
> gem5-dev mailing
list
> [email protected]
> http://m5sim.org/mailman/listinfo/gem5-dev
[5]

 

Links:
------
[1]
http://overt.org/http<div>

&gt;&nbsp;forge.net/p/swig/bugs/1297/&quot;&gt;http://sour
</div>/swig/bugs/1297/</a>[2]Ali_______________________________________________gem5-devmailinglist<ahref=
[2]
http://m5sim.org/mailman/<div>

&gt;&nbsp;left:5px;&nbsp;border-left:#1010ff&nbsp;2px&nbsp;solid;&nbsp;margin-left:5px;&nbsp;width:100%&quot;&gt;--&nbsp;IMPORTANT&nbsp;NOTICE:&nbsp;The&nbsp;contents&nbsp;of&nbsp;this&nbsp;email&nbsp;and&nbsp;any&nbsp;attachments&nbsp;are&nbsp;confidenti
</div>sobeprivileged.Ifyouarenottheintendedrecipient,pleasenotifythesenderimmediatelyanddonotdisclosethecontentstoanyotherperson,useitforanypurpose,orstoreorcopytheinformationinanymedium.Thankyou._______________________________________________<blockquotetype=
[3]
http://m5sim.org/mail<div>

&gt;&nbsp;hr
</div>urceforge.net/p/swig/bugs/1293/
[4]
http://sourceforge.net/p/swig/bugs/1297/
[5]
http://m5sim.org/mailman/listinfo/gem5-dev
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to