Looks like the case statements are only going to compile faster, the dream about guard-less programming might finally become reality Fix two bottlenecks in the compiler by bjorng · Pull Request #10252 · erlang/otp · GitHub I’ll keep pushing for the SOTA and I’m greatly appreciating @bjorng abilities to find these issues in the compiler!
Schultzer
rvirding
Creator of Erlang
I think you need to be aware that the Erlang compiler internally converts clauses matching with or without guards to an equivalent case. So whether you use def/defp or case will not make any difference.
4
Schultzer
Thanks, that is an important detail when it comes to general code, but if you’re pushing the compiler to the limit, then you sometimes have to help it on it’s way, or hope there are some extremely skilled people that can identify and fix issues in it.
0
BartOtten
Thanks for the official confirmation there should be no difference in compilation times.
For a time I was under the impression that ‘helping the compiler’ by writing a case statement was beneficial, but my tests in this thread proof otherwise. Even when generating 50.000 heads or cases, no compilation speed difference is measured.
As I posted above there is a difference though: compile time checks.
0
BartOtten
Afaik the compiler fix is beneficial for both def and case and the ‘helping’ applies the same for def and case.
0
josevalim
Creator of Elixir
Correct. And as far as I understand it, the fix seems to apply on “case”s with hundreds of clauses where the bodies are the same. If the bodies were different, it didn’t run into this pathological case, and it was already reasonably fast.
0
Schultzer
Yeah, there are probably more places that can be optimized for even faster compile times.
0
BartOtten
Schultzer
Maybe one day I’ll be able to contribute more than a reproducible example. But as a noob programmer I ain’t currently cut out for this kind of work 
0