Ilhan Polat

I am not sure I understand the objection the test knows the exact result is 0.4999.. that's why it does `norm(a) - 1e4 ?= 0.0` with some tolerance around it....

That switch is for a historical artifact that don't exist today (hopefully that is, see #286 from 13 years ago). It is not meant for the contemporary implementations. Recent BLAS...

It seems much but there is only about 20 lines added. I don't know why git is making a mess showing it as such. But I honestly don't have the...

Not sure if it is worth it, but for the backporting purposes, let me try to separate that part so it can be cherry-picked at least.

I separated the changes into cleaner pieces for review. The first two commits are for bug-fixes that can be backported, if desired. The rest is just shuffling things around. I...

@rgommers Do you still have any capacity for the remaining parts? I now have some time to spare and since works such as @andrewellis55 is slowly coming in, I'd like...

Apologies for making a mess with this one @tylerjereddy. Ralf already sorted it out. I was not able to spare time lately hence it looks disorganized but you already have...

I think the best is to close this one if you agree and do it from scratch. Since the remaining changes are mostly file renaming and move operations, there is...

> Including many copies of the same SciPy license file doesn't seem ideal. Indeed that's not ideal. There is some interest to reuse the C sources outside SciPy project. Not...

I don't think we can support big endian architectures by any means especially if we have no support from infrastructure. In fact if it was working so far I would...