Behdad Esfahbod
Behdad Esfahbod
What is IAR?
I'm afraid we can't help. You need someone who knows C++ support in that compiler to debug the issue and ideally propose a solution.
Right. Is there any chance you can send a PR? It's been a while since I've worked on this tool. Thanks.
> Parents: [None, 0, 0, 0, 0, 0, 0, np.int32(1), np.int32(1), np.int32(1), np.int32(1), np.int32(1), np.int32(2), np.int32(2), np.int32(2), np.int32(2), np.int32(3), np.int32(3), np.int32(4), np.int32(4), np.int32(7), np.int32(7), np.int32(7), np.int32(7), np.int32(8), np.int32(8), np.int32(9), np.int32(9),...
> > Parents: [None, 0, 0, 0, 0, 0, 0, np.int32(1), np.int32(1), np.int32(1), np.int32(1), np.int32(1), np.int32(2), np.int32(2), np.int32(2), np.int32(2), np.int32(3), np.int32(3), np.int32(4), np.int32(4), np.int32(7), np.int32(7), np.int32(7), np.int32(7), np.int32(8), np.int32(8), np.int32(9),...
minimum_spanning_tree is not even called when scipy is installed for me.
Sorry, my bad. Ignore me for now.
Can you do some debugging? The different parent order is expected. But that the kink disappears is unfortunate. But for now, the numpy issue is baffling me.
> Size (bytes) Percent > No support 4199928 100.00% > libatures 4222656 100.54% > HarfBuzz 5581968 132.91% See https://github.com/harfbuzz/harfbuzz/blob/main/CONFIG.md HarfBuzz compiles down to less than 500kb for basic shaping.
> you would like the subsetter to automatically check the canonical decomposition of each requested characters, No that would be excessive. The issue here is that some of the shapers...