gh-131798: Optimize _UNARY_NEGATIVE
#135223
Conversation
Python/optimizer_bytecodes.c
Outdated
| @@ -467,6 +467,15 @@ dummy_func(void) { | |||
| res = sym_new_truthiness(ctx, value, false); | |||
| } | |||
|
|
|||
| op(_UNARY_NEGATIVE, (value -- res)) { | |||
| if (sym_matches_type(value, &PyLong_Type) || sym_matches_type(value, &PyFloat_Type)) { | |||
There was a problem hiding this comment.
I think we can add a check for const, and we can setup the res by using sym_new_const. Not sure this is worth here cc @brandtbucher @Fidget-Spinner
There was a problem hiding this comment.
I'd hold off on that, since our plan is to automatically constant-evaluate stuff like this. Let's stick with hand-writing type-based optimizations like this for now.
There was a problem hiding this comment.
I did not add a const clause for this since it felt more related to #132733.
Misc/NEWS.d/next/Core_and_Builtins/2025-06-06-19-17-22.gh-issue-131798.XoV8Eb.rst
Outdated
Show resolved
Hide resolved
Python/optimizer_bytecodes.c
Outdated
| @@ -467,6 +467,15 @@ dummy_func(void) { | |||
| res = sym_new_truthiness(ctx, value, false); | |||
| } | |||
|
|
|||
| op(_UNARY_NEGATIVE, (value -- res)) { | |||
| if (sym_matches_type(value, &PyLong_Type) || sym_matches_type(value, &PyFloat_Type)) { | |||
There was a problem hiding this comment.
I'd hold off on that, since our plan is to automatically constant-evaluate stuff like this. Let's stick with hand-writing type-based optimizations like this for now.
|
Thanks for the review! @brandtbucher @Zheaoli |
Uh oh!
There was an error while loading. Please reload this page.