NumPy, SciPy, universal builds, and default fortran compiler for scientific ports

[Joshua Root]

On 2010-6-17 02:24 , Adam Mercer wrote:
> Hi
> 
> There are several tickets open regarding build issues that can
> essentially be stemmed to the problem that the NumPy and SciPy, at the
> moment, do not support universal variants. The first of which #19397
> [1] provides a solution to the universal build problem by using a
> wrapper script for the compiler.
> 
> As far as i can tell tickets #22165, #22360, #22731, #23244, and
> #24942 [2-6] are all more or less due to NumPy not building
> universally. I never build universally myself so cannot reproduce
> these but it seems like a lot of people do so this issue needs to be
> fixed.

I don't know why anyone would need a universal scipy or numpy. Do they
have any dependents that only build for a particular arch? It might be
best to just mark all the dependents as non-universal too.

> Another problem is that the default compiler for NumPy and SciPy is
> gcc43, whereas ATLAS is built gcc44 as it doesn't build correctly with
> gcc43, ticket #25206 [7] just switches over to using gcc44 as the
> default compiler so this will not be a solution for the universal
> build issue.
> 
> The solution proposed in #19397 also switches the default compiler to
> gcc44 so this may be the correct approach. What do people this  about
> bumping the default compiler to gcc44 for scientific ports (AFAICT
> there are no open tickets regarding gcc44 build issues)?

It's high time this happened. There has been no movement on these
tickets, so maintainer timeout is going to be invoked.

- Josh