[pymvpa] Weave code broken on lenny

[Yaroslav Halchenko]

uff... weave is 'cool':

    # it's nice to let the users know when anything gets compiled, as the
    # slowdown is very noticeable.
    print '<weave: compiling>'

so there is no 'native' way to suppress the crap it prints...  I wish
the authors weren't that 'nice' to us ;-) bugreport filed
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=500815

the issue Michael has encountered seems to be relevant to:

/usr/lib/python2.5/site-packages/scipy/weave/blitz/blitz/mathfunc.h:45: error: 'labs' is not a member of 'std'
and
/usr/lib/python2.5/site-packages/scipy/weave/blitz/blitz/funcs.h:509: error: call of overloaded 'abs(int&)' is ambiguous
/usr/include/c++/4.3/cmath:99: note: candidates are: double std::abs(double)
/usr/include/c++/4.3/cmath:103: note:                 float std::abs(float)
/usr/include/c++/4.3/cmath:107: note:                 long double std::abs(long double)

and we are not alone who hit it
http://lists.debian.org/debian-devel/2008/08/msg00730.html

I am trying to add externals checking for weave but besides being too
talkative weave's failure to build fails to restore sys.argv which it
modifies.... So, again bugreport with a patch
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=500814

So after some pain I seemed to figure out the way to 'patch' it in our
code ;-)
Now testing of weave external is working, but we might also like to tune
up weave invocation in distance.py so it doesn't spit out its 'nice'
messages

There is though a price associated with all this: import of the whole suite
becomes even longer due to weave external verification

NB 1 part of modification is *nix specific (adding redirections
>/dev/null ;-)), so someone might want to fix it so it works on windows
as well

PS besides that I fixed some some externals checkings so now we again
can run tests without failure if any optional dependency is missing ;-)

On Wed, 01 Oct 2008, Michael Hanke wrote:

> On Wed, Oct 01, 2008 at 09:43:48AM -0400, Yaroslav Halchenko wrote:
> > heh -- well we have workaround already I think -- I've done pure python
> > implementation. We just need proper check for weave in externals I guess
> I don't think that is done via a test -- i have everything installed (I
> think).

> Michael
-- 
Yaroslav Halchenko
Research Assistant, Psychology Department, Rutgers-Newark
Student  Ph.D. @ CS Dept. NJIT
Office: (973) 353-5440x263 | FWD: 82823 | Fax: (973) 353-1171
        101 Warren Str, Smith Hall, Rm 4-105, Newark NJ 07102
WWW:     http://www.linkedin.com/in/yarik