8000 gh-114058: More robust method handling in redundancy eliminator by Fidget-Spinner · Pull Request #115779 · python/cpython · GitHub
[go: up one dir, main page]

Skip to content

Conversation

Fidget-Spinner
Copy link
Member
@Fidget-Spinner Fidget-Spinner commented Feb 21, 2024

Method handling was actually previously buggy.

Copy link
Member
@gvanrossum gvanrossum left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you explain what was buggy in the previous version? Neither the issue you linked to nor the diff helps me understand the problem you're addressing here. Was it a correctness bug or was it simply not optimizing as much as it could around the four newly special-cased opcodes?

@Fidget-Spinner
Copy link
Member Author

Can you explain what was buggy in the previous version? Neither the issue you linked to nor the diff helps me understand the problem you're addressing here. Was it a correctness bug or was it simply not optimizing as much as it could around the four newly special-cased opcodes?

The bug is that we did not mark notnull as KNOWN. I added the other method opcodes to test out the frame locals copying.

@Fidget-Spinner Fidget-Spinner merged commit a33ffe4 into python:main Feb 23, 2024
@Fidget-Spinner Fidget-Spinner deleted the buggy_methods branch February 23, 2024 07:42
@markshannon
Copy link
Member

I think this approach is flawed.
We only push NULL to the stack in a few places. It would be safer to explicitly mark those places as pushing NULL and make the default "not NULL".

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants
0