Bytecode parity - #7536
Merged
youknowone merged 4 commits intoMar 30, 2026
Merged
Conversation
Copy link
Copy Markdown
Member
Sorry, something went wrong.
… genexpr StopIteration wrapper - split_blocks_at_jumps: split blocks at branch points so each has one exit - jump_threading: thread jumps through single-jump blocks (flowgraph.c jump_thread) - Backward conditional jump normalization: invert and create NOT_TAKEN+JUMP block - Follow empty blocks in jump-to-return optimization (next_nonempty_block) - Add PEP 479 StopIteration handler to compile_comprehension for generators
- Add Slice variant to ConstantData and BorrowedConstant - Fold constant slices (x[:3], x[1:4]) into LOAD_CONST(slice(...)) - Marshal serialization/deserialization for Slice type - Box::leak in borrow_obj_constant for PySlice roundtrip
Prepare infrastructure for exit block duplication optimization. Currently disabled pending stackdepth integration.
Copy link
Copy Markdown
Contributor
📝 Walkthrough
WalkthroughThis PR refactors bytecode compilation in the code generator, optimizing assertion message handling, rewriting f-string compilation to merge adjacent literals, improving boolean jump code generation for chained comparisons, strengthening constant folding with error propagation, guarding super-call optimizations for extended-call syntax, and revising IR-level control flow normalization for line-number resolution and small-block inlining. Changes
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Suggested reviewers
Poem🚥 Pre-merge checks | ✅ 1 | ❌ 2 ❌ Failed checks (1 warning, 1 inconclusive)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches 🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. ❤️ ShareComment @coderabbitai help to get the list of available commands and usage tips. |
Sorry, something went wrong.
youknowone
force-pushed
the
bytecode-parity
branch
from
688d815 to
c717101
Compare
March 30, 2026 06:06
Copy link
Copy Markdown
Contributor
📦 Library DependenciesThe following Lib/ modules were modified. Here are their dependencies: [x] test: cpython/Lib/test/test_compile.py (TODO: 27) dependencies: dependent tests: (no tests depend on compile) Legend:
|
Sorry, something went wrong.
youknowone
marked this pull request as ready for review
March 30, 2026 07:09
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)crates/codegen/src/compile.rs (1)🤖 Prompt for all review comments with AI agents6786-6867: Extract the condition-dependent jump emission.
Lines 6804-6818 and 6837-6851 choose the same PopJumpIf* opcode twice. A tiny helper keeps the bytecode shape in one place and makes future parity tweaks less error-prone.
As per coding guidelines, "When branches differ only in a value but share common logic, extract the differing value first, then call the common logic once to avoid duplicate code".
🤖 Prompt for AI AgentsVerify each finding against the current code and only fix it if needed. In `@crates/codegen/src/compile.rs` around lines 6786 - 6867, The compile_jump_if_compare function duplicates emission of PopJumpIfTrue vs PopJumpIfFalse in two places; refactor by computing the appropriate instruction once based on the boolean condition (e.g., determine an opcode or small helper like emit_pop_jump_if(condition, delta)) and then call emit! with that chosen instruction and target_block in both spots (replace the two conditional emit! blocks around the final compare and the single-mid-cleanup branch). Ensure the helper/variable is used in both locations inside compile_jump_if_compare so the bytecode shape is controlled from one place.
Verify each finding against the current code and only fix it if needed. Inline comments: In `@crates/codegen/src/compile.rs`: - Around line 9052-9059: The current compile_fstring_part_literal_value reparses the entire ast::StringLiteral range (via source_file.slice(string.range)) which can span multiple adjacent literal tokens and thus misdecode in the surrogate-preservation path; change this to reparse each constituent literal token individually and concatenate their parsed results: when string.value.contains(char::REPLACEMENT_CHARACTER) iterate the literal tokens that form this ast::StringLiteral (use each token's range / slice from source_file.slice(token.range)), call crate::string_parser::parse_string_literal on each token slice with string.flags.into(), collect/concatenate the resulting buffers, and return the combined Wtf8Buf instead of reparsing the whole range at once. --- Nitpick comments: In `@crates/codegen/src/compile.rs`: - Around line 6786-6867: The compile_jump_if_compare function duplicates emission of PopJumpIfTrue vs PopJumpIfFalse in two places; refactor by computing the appropriate instruction once based on the boolean condition (e.g., determine an opcode or small helper like emit_pop_jump_if(condition, delta)) and then call emit! with that chosen instruction and target_block in both spots (replace the two conditional emit! blocks around the final compare and the single-mid-cleanup branch). Ensure the helper/variable is used in both locations inside compile_jump_if_compare so the bytecode shape is controlled from one place.
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info ⚙️ Run configuration
Configuration used: Path: .coderabbit.yml
Review profile: CHILL
Plan: Pro
Run ID: 9842e9fe-d3c7-43f6-91ff-60550c5cff91
📥 CommitsReviewing files that changed from the base of the PR and between 3706c53 and c717101.
⛔ Files ignored due to path filters (1)- Lib/test/test_peepholer.py is excluded by !Lib/**
- crates/codegen/src/compile.rs
- crates/codegen/src/ir.rs
- crates/jit/tests/common.rs
Sorry, something went wrong.
Comment thread
crates/codegen/src/compile.rs
Comment on lines
+9052
to
+9059
| fn compile_fstring_part_literal_value(&self, string: &ast::StringLiteral) -> Wtf8Buf { | ||
| if string.value.contains(char::REPLACEMENT_CHARACTER) { | ||
| let source = self.source_file.slice(string.range); | ||
| crate::string_parser::parse_string_literal(source, string.flags.into()).into() | ||
| } else { | ||
| string.value.to_string().into() | ||
| } | ||
| } |
Copy link
Copy Markdown
Contributor
There was a problem hiding this comment.
⚠️ Potential issue | 🟡 Minor
Reparse each constituent literal token here.
Line 9054 reparses the combined ast::StringLiteral range with one flag set, but this node can span multiple adjacent literals. In the surrogate-preservation path that can misdecode or drop later pieces, so split plain-string parts next to an f-string can produce the wrong merged literal.
💡 Proposed fix fn compile_fstring_part_literal_value(&self, string: &ast::StringLiteral) -> Wtf8Buf {
if string.value.contains(char::REPLACEMENT_CHARACTER) {
- let source = self.source_file.slice(string.range);
- crate::string_parser::parse_string_literal(source, string.flags.into()).into()
+ string
+ .value
+ .iter()
+ .map(|lit| {
+ let source = self.source_file.slice(lit.range);
+ crate::string_parser::parse_string_literal(source, lit.flags.into())
+ })
+ .collect()
} else {
string.value.to_string().into()
}
}Verify each finding against the current code and only fix it if needed. In `@crates/codegen/src/compile.rs` around lines 9052 - 9059, The current compile_fstring_part_literal_value reparses the entire ast::StringLiteral range (via source_file.slice(string.range)) which can span multiple adjacent literal tokens and thus misdecode in the surrogate-preservation path; change this to reparse each constituent literal token individually and concatenate their parsed results: when string.value.contains(char::REPLACEMENT_CHARACTER) iterate the literal tokens that form this ast::StringLiteral (use each token's range / slice from source_file.slice(token.range)), call crate::string_parser::parse_string_literal on each token slice with string.flags.into(), collect/concatenate the resulting buffers, and return the combined Wtf8Buf instead of reparsing the whole range at once.
Sorry, something went wrong.
Merged
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Summary by CodeRabbit
Bug Fixes
Refactor
Tests