Simon

Update: first version is working for windows!

Well, as long as things are not set up correctly, a SIGSEV doesn't surprise me;) It doesn't really make sense to investigate it, before the rest works. But there's also...

Sadly, that wasn't in the scope of GSOC. But we're working on it!

This is the isolated version of the osx bug I run into in #146 and will be the place to fix it..

Hm I wonder how this got into the deps. On which Julia version is this?

ahh, that makes more sense! Sorry, i havent even added that install script, so can't really tell what it needs

>Personally, the semantics of zeros(CLArray{Float32},n,n) seems wrong as it looks like it's asking to create a Matrix{CLArray{Float32}} Interesting ambiguity! I wouldn't mind just going with `zero(CLArray{Float32},n,n)`!

Maybe the real answer is defining `CLArrays.zeros(Float32, n, n)::CLArray`... With a alias this could be as short as `cl.zeros(n, n)`

That seems like a good idea :) could be a generator, so it would automatically work for arrays having a constructor for generators!

The macro could be a nice way ;) Note that what we discussed here should have been implemented and works for all sized, isbits iterators!