mirror of
https://github.com/python/cpython.git
synced 2026-10-28 10:25:05 +08:00
Pablo Galindo Salgado
0b8aa89a05
* Use diagnostic line for missing-brace errors * Keep implicit prepared newline metadata in the reader * Reuse parsed formatted string kind in lexer * Transfer owned tokenizer encoding strings * Remove unused tokenizer state and redundant branches * Avoid rescanning partial decoded tokenizer lines * Reuse prepared tokenizer line terminators for CR detection * Reuse interactive tokenizer chunks without carriage returns * Release tokenizer constructor resources on allocation failure * Narrow interactive newline normalization wrapper * Close duplicated tokenizer descriptors when fdopen fails * Preserve cached tokenizer lines when newline allocation fails * Collect cycles involving tokenizer readline callbacks * Preserve buffered text when decoding source files * Clear the transferred syntax error line before cleanup * Propagate tokenizer column conversion failures * Allocate parser repetition buffers on first match * Allocate type comment storage only when needed * Reuse decoded multiline text when counting token columns * Skip empty decoded tokenizer chunks * Check parser array growth before allocating * Name tokenizer bounds and derive byte lengths from their data * Set the tokenizer exception when file input size overflows * gh-153569: Send exact line endings in the REPL test * Share parser repetition buffer growth * Fold Unicode input coverage into the existing REPL test * Keep parser array growth policy private * Drop obsolete bitmap stress from source discard test * Preserve input failures while scanning string literals * fixup! Preserve input failures while scanning string literals * fixup! fixup! Preserve input failures while scanning string literals * gh-153569: Remove assertion-only lexer bookkeeping