Sebastian Berg

Is it worth it? I had done this intentionally to not change behavior. I don't _mind_ changing behavior here (who uses `-inf` as priority...). IIRC, we tried relying on this...

> but the method showed up in profiling Ah, interesting. I would have thought it doesn't matter much (besides the attribute lookups).

My feeling is that yes, lazy/proxy return can work. I.e. it works fine in practice, except some niche users that really check for `isinstance(arr, np.ndarray)` and use that directly in...

Thanks for taking a look! I am pretty confident it not returning bad results and I don't think it's a giant user base yet either way. Suppose that is to...

Just to keep track, saw a similar (but different error code) error here: https://ci.preferred.jp/cupy.win.cuda120/202816/ ``` E subprocess.CalledProcessError: Command '('C:\\Development\\Python\\Python310\\python.exe', 'C:\\Windows\\TEMP\\flexci\\run-00416755\\work\\src\\tests\\example_tests\\..\\..\\examples\\gemm/sgemm.py')' returned non-zero exit status 3221226505. ```

gh-226 is of no use for pytables, which uses the new iterators... MapIter should only use MaxDims I think, not maxargs, so it should be fine (if it uses maxargs...

No, there was no change in ndite regarding this.

Good plan (I see you already did, although I guess it isn't technically backward incompatible). I wonder if now with C99 which makes not-fixed width stack arrays a bit more...

I am removing the "major release" again. There is no real ABI issue with this (unlike `NPY_MAXDIMS`). We could bump it, but honestly, in most places its used it probably...

I had left it open, because there is an argument to remove the limit entirely or increase it more (at least for the main iterator). But we could also close...