[Deb-scipy-devel] Bug#400158: numpy 1.0 available

[Ed Schofield]

Soeren Sonnenburg wrote:
> On Tue, 2006-12-12 at 13:09 +1100, Ed Schofield wrote:
>   
>> Matthias Klose wrote:
>>     
>>> Marco Presi writes:
>>>   
>>>       
>>>>  || On Fri, 24 Nov 2006 08:48:32 +0100
>>>>  || Soeren Sonnenburg <bugreports at nn7.de> wrote: 
>>>>
>>>> 	ss> Package: python-numpy
>>>> 	ss> Version: 1:1.0rc1-1
>>>> 	ss> Severity: normal
>>>>
>>>> 	ss> we don't want to rc1 to be in sarge if there is 1.0-1 right ?
>>>>
>>>> 	ss> http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=175103
>>>>
>>>>         Not really. Numpy 1.0 is not compatible both with current
>>>>         versions of scipy and matplotlib.
>>>>
>>>>         A newver Scipy version compatible with numpy-1.0 has been
>>>>         released yesterday. I am going to upload it, but I don't think
>>>>         it will be possible to include both of them in Etch.
>>>>     
>>>>         
>>> iff 1.0rc1 is incompatible with 1.0, we should try to get 1.0 into
>>> etch.  is matplotlib released in a version which compatible to 1.0?
>>>   
>>>       
>> Yes, certainly. matplotlib 0.87.7 was released soon after NumPy 1.0, but 
>> primarily to provide compatibility for the Win / Mac crowd, since 
>> NumPy's ABI changed between 1.0rc2 and 1.0rc3. But the API has only 
>> changed in minor ways, and I'm fairly confident that any of the last few 
>> releases of matplotlib (0.87.{5,6,7}) will compile and run fine against 
>> NumPy 1.0. I guess we'd need a version bump for the soname.
>>     
> <snip>
> However travis also mentioned that scipy lagged behind, but as this is now fixed it would really be a good idea to get the so name change going now... and quickly upload scipy/matplotlib afterwards.
>   

Yes, it would be great. But I think, unfortunately, it's too late for Etch:

http://lwn.net/Articles/213566/

Shame on us all for not pushing for this earlier!

-- Ed