Hmm, I had tried to use PyArray_SimpleNewFromData with an 'int
dimensions[2]' array, which worked on numpy 1.1 but apparently breaks
in the latest numpy svn. Since I am unsure what is the most portable
thing to do here, I reverted to a pattern that we've used elsewhere
that appears robust across numpy versions (included below). If any
numpy god wants to tell me the right way to create and array from an
existing buffer, let me know.
Fixed in svn 5495
int dimensions[2];
dimensions[0] = get_height(); //numrows
dimensions[1] = get_width(); //numcols
//this is breaking with numpy svn, which wants a npy_intp*
//PyArrayObject *A = (PyArrayObject *) PyArray_SimpleNewFromData(2,
dimensions, PyArray_UBYTE, _buffer);
PS: please keep all replies on list.
PyArrayObject *A = (PyArrayObject *) PyArray_FromDims(2, dimensions,
PyArray_UBYTE);
unsigned char *src = _buffer;
unsigned char *src_end = src + (dimensions[0] * dimensions[1]);
unsigned char *dst = (unsigned char *)A->data;
while (src != src_end) {
*dst++ = *src++;
}
PS: please keep all replies on list
···
On Thu, Jun 12, 2008 at 2:54 PM, Nils Wagner > <nwagner@...425...> wrote:
src/ft2font.cpp: In member function 'Py::Object FT2Image::py_as_array(const
Py::Tuple&)':
src/ft2font.cpp:273: error: cannot convert 'int*' to 'npy_intp*' in argument
passing
error: command 'gcc' failed with exit status 1
I can't explain why this changed in numpy or why PyArray_FromDims works and PyArray_SimplyNewFromData doesn't, but...
In _backend_agg.cpp, I just used a npy_intp array for dimensions everywhere. I think the difference may only matter on 64 platforms where sizeof(int) != sizeof(void*), such as x86_64. npy_intp is a typedef for Py_intptr_t which is defined thusly (in pyport.h):
*/* uintptr_t is the C9X name for an unsigned integral type such that a
* legitimate void* can be cast to uintptr_t and then back to void* again
* without loss of information. Similarly for intptr_t, wrt a signed
* integral type.
*/
···
*
Cheers,
Mike
John Hunter wrote:
On Thu, Jun 12, 2008 at 2:54 PM, Nils Wagner >> <nwagner@...425...> wrote:
src/ft2font.cpp: In member function 'Py::Object FT2Image::py_as_array(const
Py::Tuple&)':
src/ft2font.cpp:273: error: cannot convert 'int*' to 'npy_intp*' in argument
passing
error: command 'gcc' failed with exit status 1
Hmm, I had tried to use PyArray_SimpleNewFromData with an 'int
dimensions[2]' array, which worked on numpy 1.1 but apparently breaks
in the latest numpy svn. Since I am unsure what is the most portable
thing to do here, I reverted to a pattern that we've used elsewhere
that appears robust across numpy versions (included below). If any
numpy god wants to tell me the right way to create and array from an
existing buffer, let me know.
Fixed in svn 5495
int dimensions[2];
dimensions[0] = get_height(); //numrows
dimensions[1] = get_width(); //numcols
//this is breaking with numpy svn, which wants a npy_intp*
//PyArrayObject *A = (PyArrayObject *) PyArray_SimpleNewFromData(2,
dimensions, PyArray_UBYTE, _buffer);
PS: please keep all replies on list.
PyArrayObject *A = (PyArrayObject *) PyArray_FromDims(2, dimensions,
PyArray_UBYTE);
unsigned char *src = _buffer;
unsigned char *src_end = src + (dimensions[0] * dimensions[1]);
unsigned char *dst = (unsigned char *)A->data;
while (src != src_end) {
*dst++ = *src++;
}
PS: please keep all replies on list
-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php
_______________________________________________
Matplotlib-devel mailing list
Matplotlib-devel@lists.sourceforge.net
matplotlib-devel List Signup and Options
--
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA