Open fabianhjr opened 4 years ago
Marceline discovered that this PR causes an explosion of memory usage while compiling Idris2 in Core/Binary (Extra ~ 5GB or so of memory usage)
This does make the nat hack a bit less hacky, thanks! Is the compile time performance still an issue? Sometimes that can be due to ambiguity resolution going a bit out of control.
Yes, there is still an issue with Idris1 compilation of Idris2.
Idris2 runs fine though.
Helped me improve the performance of an alternative idris2 prelude I am playing with, tested with mock of ProjectEuler problem #1: