Uh oh!
There was an error while loading. Please reload this page.
bpo-43683: Handle generator entry in bytecode - #25138
Conversation
4dff1d0 to
46c71f5CompareUh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
…o bytecode. Document new bytecode and make it fail gracefully if mis-compiled. Add NEWS item Bump magic number.
46c71f5 to
1a29ec3CompareUh oh!
There was an error while loading. Please reload this page.
| entryblock = b; | ||
| } | ||
| if (insert_generator_prefix(c, entryblock)) { |
There was a problem hiding this comment.
If we're compiling a generator (or coroutine etc.) the new function assumes (without asserting) that entryblock is not NULL. But evidently it could be, since below on line 6690 there's a check whether it could be. I don't know how it could be NULL - that would mean there's no code in the function at all, but it seems there's code above that always adds at least one instruction (RETURN_VALUE on line 6668). So maybe the check for entryblock == NULL below can be replaced by an assert? (I'd place that assert right underneath the for-loop above on lines 6683-6686.
There was a problem hiding this comment.
entryblock cannot be NULL. I'll add the assert.
markshannon
commented
Apr 6, 2021
Fixes are in #25224 |
Adds GEN_START bytecode to handle checking for None when starting a generator or coroutine.
This means that:
send()doesn't need to check the state of the generator.https://bugs.python.org/issue43683