matplotlib-devel Mailing List for matplotlib

On Saturday 08 July 2006 10:28 am, Darren Dale wrote:
> On Friday 07 July 2006 6:45 pm, Fernando Perez wrote:
> > On 7/7/06, Jeff Whitaker <js...@fa...> wrote:
> > > John:  setupext.py in svn 2545 uses numpy.get_include(), which
> > > apparently is not available in 0.9.8.  In 0.9.8, it's
> > > get_numpy_include().
> >
> > And in current numpy, get_numpy_include fires a deprecation warning (I
> > just fixed this in my code).  Such are the joys of living on the
> > bleeding edge :)
>
> Sorry, I was the one who changed get_numpy_include to get_include. I guess
> I was thinking that anyone who is using svn mpl would also be using svn
> numpy. Should I change it back to get_numpy_include?

This is fixed in svn. I wrapped get_include in a try statement for the time 
being, get_numpy_include gets called in the event of an attribute error.

Darren

On Friday 07 July 2006 6:45 pm, Fernando Perez wrote:
> On 7/7/06, Jeff Whitaker <js...@fa...> wrote:
> > John:  setupext.py in svn 2545 uses numpy.get_include(), which
> > apparently is not available in 0.9.8.  In 0.9.8, it's
> > get_numpy_include().
>
> And in current numpy, get_numpy_include fires a deprecation warning (I
> just fixed this in my code).  Such are the joys of living on the
> bleeding edge :)


Sorry, I was the one who changed get_numpy_include to get_include. I guess I 
was thinking that anyone who is using svn mpl would also be using svn numpy. 
Should I change it back to get_numpy_include?

On 7/8/06, Travis Oliphant <oli...@ie...> wrote:
> Fernando Perez wrote:
> > Hi all,
> >
> > I was trying to track down a problem with current numpy/scipy from
> > svn, and realized that mpl doesn't build anymore:
> >
> >
> Get the latest SVN of matplotlib.   I fixed it there to use
> numpy/oldnumeric.h
>
> It compiles and runs for me.

Mmh, odd.  I must have caught things again in-between updates, because
I /was/ using SVN mpl for the test at the moment.

Thanks for the info, I'll update again.

Cheers,

f

Hi all,

I was trying to track down a problem with current numpy/scipy from
svn, and realized that mpl doesn't build anymore:

...

compile options:
'-I/home/fperez/tmp/local/lib/python2.4/site-packages/numpy/core/include
-I/usr/local/include -I/usr/include -I. -I/usr/include/python2.4 -c'
extra options: '-DSCIPY=1'
gcc: src/_ns_cntr.c
src/_ns_cntr.c: In function 'build_cntr_list_v2':
src/_ns_cntr.c:1376: warning: unused variable 'point'
src/_ns_cntr.c: In function 'Cntr_init':
src/_ns_cntr.c:1617: error: 'PyArray_SBYTE' undeclared (first use in
this function)
src/_ns_cntr.c:1617: error: (Each undeclared identifier is reported only once
src/_ns_cntr.c:1617: error: for each function it appears in.)
src/_ns_cntr.c: In function 'build_cntr_list_v2':
src/_ns_cntr.c:1376: warning: unused variable 'point'
src/_ns_cntr.c: In function 'Cntr_init':
src/_ns_cntr.c:1617: error: 'PyArray_SBYTE' undeclared (first use in
this function)
src/_ns_cntr.c:1617: error: (Each undeclared identifier is reported only once
src/_ns_cntr.c:1617: error: for each function it appears in.)
error: Command "gcc -pthread -fno-strict-aliasing -DNDEBUG -g -O3
-Wall -Wstrict-prototypes -fPIC
-I/home/fperez/tmp/local/lib/python2.4/site-packages/numpy/core/include
-I/usr/local/include -I/usr/include -I. -I/usr/include/python2.4 -c
src/_ns_cntr.c -o build/temp.linux-i686-2.4/src/_ns_cntr.o -DSCIPY=1"
failed with exit status 1


There have been changes to the C API in numpy which are probably
causing this (See Travis' recent messages on that list).

Cheers,

f