[Deb-scipy-devel] Re: Bug#380218: segmentation fault in apply_along_axis()

[Jose Fonseca]

Mark,

On Mon, 2006-07-31 at 09:18 +0200, Mark Asbach wrote:
> Hi Jose,
> 
> > I'm not sure what's the policy concerning the bugs which should and
> > should not be fixed on debian stable, and how this bug qualifies.
> 
> Me neither ...

The current policy regarding updates to stable is given at
http://www.debian.org/doc/developers-reference/ch-pkgs.en.html#s-upload-stable

It's OK to update stable packages if they meet certain criteria.

> > Anyway, I'm quite busy ATM, but when I get some free time I'll look  
> > into
> > this issue (if one of my co-maintainers does not beat me to it).
> 
> Okay, thanks.
> 
> This bug is really rendering scipy useless for a multitude of  
> problems, because the only work around is using for() loops in python  
> which implies a huge performance penalty - and of course it means  
> that third-party python code must be rewritten when it segfaults  
> because of apply_along_axis().

So it might qualify as "truly critical functionality problem".

> 
> Yours, Mark
> 

>From the info in the scipy tracker, the bug is actually in Numeric --
what's in Scipy CVS is a work-around. 

It will be a bunch of work:
- search the change in the CVS and produce a patch
- include the patch in a new release 
- install stable debian and build the release

Please bear in mind that it might take a while or not happen at all,
since it's not a high priority issue (personally), as I don't use
stable, there are up-to-date and fixed versions of scipy in testing and
unstable, a new stable debian is planned for shortly, and there are
other bugs which also require my attention...

Sorry I can't more helpful.

Regards,

José Fonseca


		
___________________________________________________________ 
The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html