List.traverseResultA and friends have x @ y on the error branch, this could potentially blow up about fstoolkit.errorhandling HOT 5 CLOSED

Ouch, yeah that's bad. I mentioned the error path needs to be considered as well in #167 when doing a performance pass at it.

from fstoolkit.errorhandling.

Change looks good, thanks!

from fstoolkit.errorhandling.

I looked into this briefly and errs @ e are two lists so your recommended fix wouldn't work exactly. So I think this would be better changed to be some type of while loop semantics to get performance.

from fstoolkit.errorhandling.

Oh, yeah, you're right. Though, arguably, the new lists will be shorter than the tally, so a quick rec with e::errs would be nice. But that would, perhaps, invert the order of errors? Hmm, tricky. Yet, such loops are quite common, and with a List.rev at the end, should work.

from fstoolkit.errorhandling.

I did benchmark this a bit ago: #167, various tradeoffs

Yeah idk why it's currently implemented with the @ after looking at it now.

from fstoolkit.errorhandling.