matplotlib-devel Mailing List for matplotlib

Referring to this:
>/ #include <math.h>
/32a34,37
>/ extern "C" {
/>/ int isnan(double);
/>/ }
/Jouni K Seppanen wrote:
>I suppose that would work, but then you will take a slight performance
>hit when calling isnan. It's probably not too bad; depends on how
>often it gets called.
> 
>
The code that makes use of isnan is not called enough to be worried 
about a minor speed hit.
>Apparently, the upshot is that isnan is a C99 feature and C++ does not
>incorporate C99.
>
So, does that mean the above should work on most C++ compilers? (Do 
they implement C99 mode when in extern "C" mode?) If so, I'm all for 
that approach as abandoning the isnan64 macro I borrowed from numarray. 
I'd rather take the minor speed hit than have all that intricate C 
coding currently in the MPL_isnan.h file.

Tom Loredo <lo...@as...> writes:
> http://www.ssl.berkeley.edu/pipermail/boinc_dev/2004-October/000529.html
>
> It involves including math.h and defining a prototype for isnan
> explicitly.
I suppose that would work, but then you will take a slight performance
hit when calling isnan. It's probably not too bad; depends on how
often it gets called.
Apple's cmath include file (my version of it, anyway, in
/usr/include/gcc/darwin/3.3/c++/) first defines a template function
that expands the macro, then undefines the macro and makes the
function available in the standard namespace (I've omitted a lot of
stuff here):
 namespace __gnu_cxx
 {
 template<typename _Tp>
 int 
 __capture_isnan(_Tp __f) { return isnan(__f); }
 }
 #undef isnan
 namespace __gnu_cxx
 {
 template<typename _Tp>
 int 
 isnan(_Tp __f) { return __capture_isnan(__f); }
 }
 namespace std
 {
 using __gnu_cxx::isnan;
 } 
So this ensures that when you call std::isnan, it is expanded at
compile-time to a call to __capture_isnan, which in turn is expanded
to the original isnan macro.
It also probably means that including math.h before cmath will have no
effect, since cmath will #undef the isnan macro. Including math.h
after cmath will also have no effect, since it is guarded with #ifndef
__MATH_H__.
The solution at the URL you mentioned adds a prototype for the isnan
function, which is included in the math library. I think it only
handles doubles, not floats or long doubles (and I have absolutely no
idea whether casting a float NaN to double works), and a function call
causes a minor performance hit. Other than that, I suppose that's
fine.
Somehow your cmath is failing to define isnan. At least in my version,
the definitions in __gnu_cxx (but not the #undef!) are conditioned on
 #if _GLIBCPP_USE_C99
Googling with that macro and isnan shows that a lot of other people
have been having similar problems. Another attempt at a solution is at
http://tolstoy.newcastle.edu.au/~rking/R/devel/05/01/1861.html
Another possibly relevant page is
http://lists.apple.com/archives/Xcode-users/2005/Feb/msg00238.html
Apparently, the upshot is that isnan is a C99 feature and C++ does not
incorporate C99. Perhaps the feature has been added anyway in some
version of gcc 3.3 between yours and mine.
-- 
Jouni

Tom Loredo <lo...@as...> writes:
> I've verified that the new "using std::isnan;" line is in _transforms.cpp
> and is copied to _na_transforms.cpp; it just doesn't seem to satisfy
> the compiler [gcc 3.3 20030304 (Apple Computer, Inc. build 1671) on 10.3.9].
> I'm not a c++ programmer, and I'm stumped as to what further to try.
> Jouni, did this indeed solve your build problems completely? Anyone
> have any other suggestions?
Yes, this solved my build problems, and the resulting library seems to
work. I also tried building with gcc 3.3 (gcc version 3.3 20030304
(Apple Computer, Inc. build 1809)).
Just to be sure, did you delete the "build" directory before trying
again? There could have been something left over that is causing the
error.
-- 
Jouni

OK, it seems trying to pick up "isnan" from the C/C++ stdlib is riddled
with difficulties, at least on Mac OS X.
Following numarray's lead, I've implemented our own test "MPL_isnan64".
This involved the addition of a new header file in src/ and the patch
I'm including. Also included in the patch (and also checked into CVS)
is the unit test which shows why this whole change is necessary.
This all works for me on Mac OS X 10.4 and Debian sarge AMD64. Tom,
maybe you could try it again on Mac OS X 10.3?
Cheers!
Andrew
Tom Loredo wrote:
>Andrew,
>
>Thanks for the suggestions, which proved helpful---transforms now
>builds; the rest of the build is proceeding and my fingers are
>crossed. Here's what I found/did.
>
>Including math.h did not help. Poking around with man and in /usr/include
>revealed that isnan is defined as a macro, not a function; math.h includes 
>/usr/include/architecture/ppc/math.h and the definition is in there. I
>did some googling and found one other case of this causing a problem;
>a workaround is provided here:
>
>http://www.ssl.berkeley.edu/pipermail/boinc_dev/2004-October/000529.html
>
>It involves including math.h and defining a prototype for isnan
>explicitly. I tried this and the build failed, but with a slightly
>different error:
>
>gcc: src/_na_transforms.cpp
>src/_na_transforms.cpp: In member function `Py::Object 
> Bbox::update_numerix(const Py::Tuple&)':
>src/_na_transforms.cpp:452: error: `isnan' not declared
>src/_na_transforms.cpp: In member function `Py::Object 
> Bbox::update_numerix(const Py::Tuple&)':
>src/_na_transforms.cpp:452: error: `isnan' not declared
>
>At this point, I commented out the "using..." line that was just
>added (reasoning that isnan is now in the file's namespace), and
>the build succeeded.
>
>I don't know the build process well enough to know if this can be
>easily automated in a cross-platform manner, or if there is a 
>better solution.
>
>As I write the numarray and scipy builds of _transforms.cpp have
>both succeeded. Hopefully that's the only stumbling block.
>
>-Tom
>
>
>
>-------------------------------------------------
>This mail sent through IMP: http://horde.org/imp/
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems? Stop! Download the new AJAX search engine that makes
>searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Matplotlib-devel mailing list
>Mat...@li...
>https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
> 
>

Andrew,
Thanks for the suggestions, which proved helpful---transforms now
builds; the rest of the build is proceeding and my fingers are
crossed. Here's what I found/did.
Including math.h did not help. Poking around with man and in /usr/include
revealed that isnan is defined as a macro, not a function; math.h includes 
/usr/include/architecture/ppc/math.h and the definition is in there. I
did some googling and found one other case of this causing a problem;
a workaround is provided here:
http://www.ssl.berkeley.edu/pipermail/boinc_dev/2004-October/000529.html
It involves including math.h and defining a prototype for isnan
explicitly. I tried this and the build failed, but with a slightly
different error:
gcc: src/_na_transforms.cpp
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:452: error: `isnan' not declared
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:452: error: `isnan' not declared
At this point, I commented out the "using..." line that was just
added (reasoning that isnan is now in the file's namespace), and
the build succeeded.
I don't know the build process well enough to know if this can be
easily automated in a cross-platform manner, or if there is a 
better solution.
As I write the numarray and scipy builds of _transforms.cpp have
both succeeded. Hopefully that's the only stumbling block.
-Tom
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/

Hmm... I've tested with gcc 3.3 on linux, so I'm not sure why it's breaking.
First, check if adding "#include <math.h>" will help.
Some other ideas:
 * Do "man isnan" from the command line and see if it specifies another
header file.
 * Search inside /usr/include for isnan and see what comes up.
Good luck,
Andrew
Tom Loredo wrote:
>Hi folks,
>
>Thanks for the quick help. I do also have numarray installed, so
>I don't think the NUMARRAY setting is the problem.
>
>Jouni, thanks for the attempted fix, but it continues to fail
>for me, at the same place:
>
>gcc: src/_na_transforms.cpp
>src/_na_transforms.cpp: In member function `Py::Object 
> Bbox::update_numerix(const Py::Tuple&)':
>src/_na_transforms.cpp:447: error: `isnan' not declared
>src/_na_transforms.cpp:494: error: `isnan' undeclared (first use this function)
>src/_na_transforms.cpp:494: error: (Each undeclared identifier is reported only 
> once for each function it appears in.)
>src/_na_transforms.cpp: In member function `Py::Object 
> Bbox::update_numerix(const Py::Tuple&)':
>src/_na_transforms.cpp:447: error: `isnan' not declared
>src/_na_transforms.cpp:494: error: `isnan' undeclared (first use this function)
>src/_na_transforms.cpp:494: error: (Each undeclared identifier is reported only 
> once for each function it appears in.)
>error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic 
>-DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Isrc -I. -I/usr/local/include -I/usr/include -I/sw/include -I. -I/Library/
>Frameworks/Python.framework/Versions/2.4/include/python2.4 -c src/_na_transforms.cpp -o build/temp.darwin-7.9.0-
>Power_Macintosh-2.4/src/_na_transforms.o -DNUMARRAY=1" failed with exit status 1
>
>I've verified that the new "using std::isnan;" line is in _transforms.cpp
>and is copied to _na_transforms.cpp; it just doesn't seem to satisfy
>the compiler [gcc 3.3 20030304 (Apple Computer, Inc. build 1671) on 10.3.9].
>I'm not a c++ programmer, and I'm stumped as to what further to try.
>Jouni, did this indeed solve your build problems completely? Anyone
>have any other suggestions?
>
>Thanks,
>Tom
>
>
>-------------------------------------------------
>This mail sent through IMP: http://horde.org/imp/
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems? Stop! Download the new AJAX search engine that makes
>searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Matplotlib-devel mailing list
>Mat...@li...
>https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
> 
>

Hi folks,
Thanks for the quick help. I do also have numarray installed, so
I don't think the NUMARRAY setting is the problem.
Jouni, thanks for the attempted fix, but it continues to fail
for me, at the same place:
gcc: src/_na_transforms.cpp
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:447: error: `isnan' not declared
src/_na_transforms.cpp:494: error: `isnan' undeclared (first use this function)
src/_na_transforms.cpp:494: error: (Each undeclared identifier is reported only 
 once for each function it appears in.)
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:447: error: `isnan' not declared
src/_na_transforms.cpp:494: error: `isnan' undeclared (first use this function)
src/_na_transforms.cpp:494: error: (Each undeclared identifier is reported only 
 once for each function it appears in.)
error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic 
-DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Isrc -I. -I/usr/local/include -I/usr/include -I/sw/include -I. -I/Library/
Frameworks/Python.framework/Versions/2.4/include/python2.4 -c src/_na_transforms.cpp -o build/temp.darwin-7.9.0-
Power_Macintosh-2.4/src/_na_transforms.o -DNUMARRAY=1" failed with exit status 1
I've verified that the new "using std::isnan;" line is in _transforms.cpp
and is copied to _na_transforms.cpp; it just doesn't seem to satisfy
the compiler [gcc 3.3 20030304 (Apple Computer, Inc. build 1671) on 10.3.9].
I'm not a c++ programmer, and I'm stumped as to what further to try.
Jouni, did this indeed solve your build problems completely? Anyone
have any other suggestions?
Thanks,
Tom
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/

Tom, thanks for the bug report and Jouni thanks for the fix, which I've
just checked in.
Cheers!
Andrew
Jouni K Seppanen wrote:
>Tom Loredo <lo...@as...> writes:
>
> 
>
>>I'm having trouble building the current CVS version of mpl
>>on OS X (10.3.9, Python 2.4). [...] numerix=scipy [...]
>>src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this function)
>> 
>>
>
>I'm having the same problem in _nc_transforms.cpp, and I'm building on
>OS X 10.4.3, Python 2.4, gcc 4.0, and with -DNUMERIC=1 (no numarray,
>no scipy). It seems that the relevant lines in _transforms.cpp were
>last modified by astraw on 15-Dec-05.
>
>The following one-line change seems to fix that problem (but the rest
>of the build hasn't finished yet). I don't know what this does in
>other compilers, but comments in gcc's cmath suggest that the standard
>requires isnan to be in the std namespace.
>
> 
>
>------------------------------------------------------------------------
>
>Index: src/_transforms.cpp
>===================================================================
>RCS file: /cvsroot/matplotlib/matplotlib/src/_transforms.cpp,v
>retrieving revision 1.39
>diff -u -r1.39 _transforms.cpp
>--- src/_transforms.cpp	15 Dec 2005 04:54:52 -0000	1.39
>+++ src/_transforms.cpp	19 Dec 2005 06:14:42 -0000
>@@ -444,6 +444,7 @@
> Py::Object
> Bbox::update_numerix(const Py::Tuple &args) {
> //update the boox from the numerix arrays x and y
>+ using std::isnan;
> _VERBOSE("Bbox::update_numerix");
> 
> args.verify_length(3);
> 
>

Tom Loredo <lo...@as...> writes:
> I'm having trouble building the current CVS version of mpl
> on OS X (10.3.9, Python 2.4). [...] numerix=scipy [...]
> src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this function)
I'm having the same problem in _nc_transforms.cpp, and I'm building on
OS X 10.4.3, Python 2.4, gcc 4.0, and with -DNUMERIC=1 (no numarray,
no scipy). It seems that the relevant lines in _transforms.cpp were
last modified by astraw on 15-Dec-05.
The following one-line change seems to fix that problem (but the rest
of the build hasn't finished yet). I don't know what this does in
other compilers, but comments in gcc's cmath suggest that the standard
requires isnan to be in the std namespace.
-- 
Jouni K Seppänen

Tom,
The output to gcc shows that NUMARRAY is defined. Try undefining NUMARRAY
and NUMERIC.
 -- Paul
On 12/18/05, Tom Loredo <lo...@as...> wrote:
>
>
> Hi folks,
>
> I'm having trouble building the current CVS version of mpl
> on OS X (10.3.9, Python 2.4). I am not using the 0.85 release
> because I want to use numerix=3Dscipy. I previously installed
> a CVS checkout from Dec 2, and it built fine. It seemed to
> plot fine, except that the TeX example produced blank or
> flat-line labels. I just installed the latest scipy_core
> release (and the latest scipy from svn), and also thought
> I'd try the latest mpl from CVS to see if the TeX problem was
> fixed. Now it won't build; about 6min into the build it exits:
>
> gcc: src/mplutils.cpp
> gcc: CXX/cxx_extensions.cxx
> gcc: src/_na_transforms.cpp
> src/_na_transforms.cpp: In member function `Py::Object
> Bbox::update_numerix(const Py::Tuple&)':
> src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this
> function)
> src/_na_transforms.cpp:493: error: (Each undeclared identifier is reporte=
d
> only
> once for each function it appears in.)
> src/_na_transforms.cpp: In member function `Py::Object
> Bbox::update_numerix(const Py::Tuple&)':
> src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this
> function)
> src/_na_transforms.cpp:493: error: (Each undeclared identifier is reporte=
d
> only
> once for each function it appears in.)
> error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp
> -mno-fused-madd -fno-common -dynamic
> -DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Isrc -I. -I/usr/local/include
> -I/usr/include -I/sw/include -I. -I/Library/
> Frameworks/Python.framework/Versions/2.4/include/python2.4 -c
> src/_na_transforms.cpp -o build/temp.darwin-7.9.0-
> Power_Macintosh-2.4/src/_na_transforms.o -DNUMARRAY=3D1" failed with exit
> status 1
>
> I'd appreciate any help on how to fix this. I'm teaching a class
> of students using Python in about 4 weeks, and I need to provide
> them a version of things to set up on their laptops that I know
> will work with my examples.
>
> Thanks,
> Tom
>
> PS: If anyone has any tips on getting scipy to build properly on
> OS X, pass them along. The current version appears unable to find
> Apple's atlas, nor does it find the fftw libraries I installed in
> /usr/lib.
>
>
>
>
>
> -------------------------------------------------
> This mail sent through IMP: http://horde.org/imp/
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log
> files
> for problems? Stop! Download the new AJAX search engine that makes
> searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=3D7637&alloc_id=3D16865&op=3Dclick
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
--
Paul Barrett, PhD Johns Hopkins University
Assoc. Research Scientist Dept of Physics and Astronomy
Phone: 410-516-5190 Baltimore, MD 21218

Hi folks,
I'm having trouble building the current CVS version of mpl
on OS X (10.3.9, Python 2.4). I am not using the 0.85 release
because I want to use numerix=scipy. I previously installed
a CVS checkout from Dec 2, and it built fine. It seemed to
plot fine, except that the TeX example produced blank or
flat-line labels. I just installed the latest scipy_core
release (and the latest scipy from svn), and also thought
I'd try the latest mpl from CVS to see if the TeX problem was
fixed. Now it won't build; about 6min into the build it exits:
gcc: src/mplutils.cpp
gcc: CXX/cxx_extensions.cxx
gcc: src/_na_transforms.cpp
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this function)
src/_na_transforms.cpp:493: error: (Each undeclared identifier is reported only 
 once for each function it appears in.)
src/_na_transforms.cpp: In member function `Py::Object 
 Bbox::update_numerix(const Py::Tuple&)':
src/_na_transforms.cpp:493: error: `isnan' undeclared (first use this function)
src/_na_transforms.cpp:493: error: (Each undeclared identifier is reported only 
 once for each function it appears in.)
error: Command "gcc -fno-strict-aliasing -Wno-long-double -no-cpp-precomp -mno-fused-madd -fno-common -dynamic 
-DNDEBUG -g -O3 -Wall -Wstrict-prototypes -Isrc -I. -I/usr/local/include -I/usr/include -I/sw/include -I. -I/Library/
Frameworks/Python.framework/Versions/2.4/include/python2.4 -c src/_na_transforms.cpp -o build/temp.darwin-7.9.0-
Power_Macintosh-2.4/src/_na_transforms.o -DNUMARRAY=1" failed with exit status 1
I'd appreciate any help on how to fix this. I'm teaching a class
of students using Python in about 4 weeks, and I need to provide
them a version of things to set up on their laptops that I know
will work with my examples.
Thanks,
Tom
PS: If anyone has any tips on getting scipy to build properly on
OS X, pass them along. The current version appears unable to find
Apple's atlas, nor does it find the fftw libraries I installed in
/usr/lib.
 
-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/