GH-98831: Update generate_cases.py: register inst, opcode_metadata.h#100735
Uh oh!
There was an error while loading. Please reload this page.
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters. Learn more about bidirectional Unicode characters
(These aren't used yet, but may be coming soon, and it's easier to keep this tool the same between branches.)
gvanrossum changed the title Python/opcode_metadata.h Outdated
| [BINARY_OP] ={2, 1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [SWAP] ={-1, -1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [EXTENDED_ARG] ={-1, -1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [CACHE] ={-1, -1, DIR_NONE, DIR_NONE, DIR_NONE }, |
Member
There was a problem hiding this comment.
END_FOR is missing, I think we need something to handle macros.
Also, note that missing entries default to 0. Would be safer if they defaulted to -1.
MemberAuthor
There was a problem hiding this comment.
Also, note that missing entries default to 0. Would be safer if they defaulted to -1.
The code generator doesn't know the numeric values of the opcodes (yet) so it doesn't have enough info to generate explicit entries filled with -1 for missing values. The best we can do is add an additional flag valid_entry.
Python/opcode_metadata.h Outdated
| [GET_AWAITABLE] ={1, 1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [SEND] ={-1, -1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [ASYNC_GEN_WRAP] ={1, 1, DIR_NONE, DIR_NONE, DIR_NONE }, | ||
| [YIELD_VALUE] ={1, 0, DIR_NONE, DIR_NONE, DIR_NONE }, |
Member
There was a problem hiding this comment.
stack_effect of YIELD_VALUE is 0. Why is it here 1 popped, 0 pushed?
MemberAuthor
There was a problem hiding this comment.
That seems to be a bug in bytecodes.c. The code ends with goto resume_frame so the push sequence at the end is never executed. I can fix it by making the IO effect (retval -- retval) (indicating that it doesn't change) or perhaps (retval -- unused), indicating that we don't write it here. I presume it's written in some weird way when the frame is resumed.
Member
This is the sanity check I used to find those two issues: --- a/Python/compile.c+++ b/Python/compile.c@@ -8854,6 +8854,26 @@ assemble(struct compiler *c, int addNone) assert(no_redundant_jumps(g)); + for (basicblock *b = g->g_entryblock; b != NULL; b = b->b_next){+ for (int j = 0; j < b->b_iused; j++){+ struct instr *instr = &b->b_instr[j];+ int opcode = instr->i_opcode;+ assert(opcode < 256);+ int pushed = _PyOpcode_opcode_metadata[opcode].n_pushed;+ int popped = _PyOpcode_opcode_metadata[opcode].n_popped;+ if (pushed >= 0 && popped >= 0){+ int oparg = instr->i_oparg;+ int effect = stack_effect(opcode, oparg, -1);+ if (effect != pushed-popped){+ fprintf(stderr, "opcode = %d oparg = %d effect = %d pushed = %d popped = %d\n",+ opcode, oparg, effect, pushed, popped);+ }+ assert(effect == pushed-popped);+ }+ }+ }++ |
MemberAuthor
gvanrossum commented Jan 4, 2023
Maybe we should include that in the code as a helper function and call it in an |
iritkatriel approved these changes Jan 5, 2023
gvanrossum changed the title 
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
MemberAuthor
gvanrossum commented Jan 5, 2023
Have to revert, see #100777 |
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
bedevere-bot commented Jan 5, 2023
|
MemberAuthor
gvanrossum commented Jan 5, 2023
Bot failures fixed by #100778 |
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
(These aren't used yet, but may be coming soon,
and it's easier to keep this tool the same between branches.)