[Numpy-discussion] 64bit infrastructure

Travis Oliphant travis at continuum.io
Wed Aug 22 10:25:09 EDT 2012


On Aug 22, 2012, at 3:59 AM, Ralf Gommers wrote:

> 
> 
> On Tue, Aug 21, 2012 at 12:51 AM, Travis Oliphant <travis at continuum.io> wrote:
> I'm actually not sure, why.   I think the issue is making sure that the release manager can actually "build" NumPy without having to buy a particular compiler.
> 
> That would help, yes. MS Express doesn't work under Wine last time I checked by the way. 
> 
> However, the issue is more than just one license. There's a large number of packages that depend on numpy and provide binaries. If they can't make those compatible with numpy ones, that's a problem. Users will first install numpy 64-bit, and then later find out that part of the scientific Python stack isn't available to them anymore.
>  

As far as I understand, you don't *have* to build all downstream dependencies with the same compiler that NumPy was built with unless your extension relies on the way C-functions pass structures on the stack (not pointers to them, but structures as a whole) or if it relies on the representation of FILE*.      At one time all structures were passed as pointers specifically for this reason.   The FILE* situation is a problem, but most extensions don't use NumPy C-API calls that have a FILE* argument. 

> 
> But, I would rather have official builds of NumPy for all platforms with a compiler paid for by a NumPy-sponsor than not have them.
> 
> I would only want to have those if I can have the whole stack. If that's possible, then (if Christoph is okay with it) why not upload Christoph's binaries also to SF and call them "official" then? I don't see the point in duplicating his efforts.
> 

Yes, I agree.   I would really like there to be several build machines, maintained by the community (NumFOCUS could sponsor this) which can be used to build windows binaries from a build specification.   I think bento is a good tool.   We could accelerate its use by offering machines that automatically build binaries for packages with bento scripts.    

If someone is interested in doing a project like this for the community, let me know as there are funds available to sponsor it --- just the lack of someone to see it through.  

-Travis


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20120822/48dcaaf2/attachment.html>


More information about the NumPy-Discussion mailing list