[SciPy-user] Matlab IO -- can others with matlab test this bug?
Wed Jul 30 09:45:46 CDT 2008
Thanks! It looks like it's a limited matlab bug...
I'll still be making a patch to fix this, but that will definitely
need some testing to make sure I didn't inadvertently break other
stuff. I'll post more information after I have a patch.
On Jul 30, 2008, at 10:16 AM, Jonathan Hunt wrote:
> Hi Zachary,
> I just tested your test with yesterdays SVN SciPy and both .mat files
> load with no problems in MATLAB 18.104.22.1684 (R2008a) (with Python 2.5,
> OS X 10.5.4). Perhaps Mathworks fixed the bug in there end if that's
> where it is.
> Hope that helps.
> On Thu, Jul 31, 2008 at 12:00 AM, Zachary Pincus
> <email@example.com> wrote:
>>> This is no doubt a dumb question. But why it is not possible
>>> simply to
>>> import the library in place (i.e no install necessary, just use the
>>> python files as they are laid out in the SVN co directory). Is it
>>> because of the C/Fortran files that need to be compiled?
>> Yes, basically. Pure-python libraries can usually be imported in-
>> -- though sometimes, too-clever setup.py files can actually make it
>> the installed folder structure is different from the one in the
>> directory, and thus an in-place import will fail. (I think this is
>> probably considered bad form though.)
>> In the specific, limited case of the matlab IO sub-sub-library, it is
>> indeed pure python and can be imported in-place -- this was one of my
>> hackish suggestions for testing it.
>> SciPy-user mailing list
> Jonathan J Hunt <firstname.lastname@example.org>
> Homepage: http://www.42quarks.net.nz/wiki/JJH
> (Further contact details there)
> "Physics isn't the most important thing. Love is." Richard Feynman
> SciPy-user mailing list
More information about the SciPy-user