https://github.com/python/cpython/commit/21c4b7359d91b3d78acb04afbe339cbea92bffae commit: 21c4b7359d91b3d78acb04afbe339cbea92bffae branch: main author: Pieter Eendebak <[email protected]> committer: serhiy-storchaka <[email protected]> date: 2026-06-24T11:09:50Z summary:
gh-152056: Compile single-category character sets to a bare CATEGORY opcode (GH-152057) A character set containing exactly one category, e.g. [\d] or [^\s], now compiles to a single CATEGORY opcode (like \d or \S) instead of an IN block. The negated form maps to the complementary category. This speeds up matching and reduces the size of the compiled byte code. Co-authored-by: Claude Opus 4.8 (1M context) <[email protected]> files: A Misc/NEWS.d/next/Library/2026-06-24-10-30-00.gh-issue-152056.Qk7mZ2.rst M Doc/whatsnew/3.16.rst M Lib/re/_parser.py diff --git a/Doc/whatsnew/3.16.rst b/Doc/whatsnew/3.16.rst index e3f04739e3b49d..f9e54cde10afe0 100644 --- a/Doc/whatsnew/3.16.rst +++ b/Doc/whatsnew/3.16.rst @@ -288,10 +288,12 @@ re -- * Character class escapes (``\d``, ``\D``, ``\s``, ``\S``, ``\w`` and ``\W``) - outside a character set are now compiled to a single ``CATEGORY`` opcode - instead of being wrapped in an ``IN`` block. This speeds up matching of - patterns such as ``\d+`` and reduces the size of the compiled byte code. - (Contributed by Serhiy Storchaka in :gh:`152033`.) + outside a character set, and character sets containing a single such escape + (such as ``[\d]`` or ``[^\s]``), are now compiled to a single ``CATEGORY`` + opcode instead of being wrapped in an ``IN`` block. This speeds up matching + of patterns such as ``\d+`` and reduces the size of the compiled byte code. + (Contributed by Serhiy Storchaka in :gh:`152033` and Pieter Eendebak in + :gh:`152056`.) module_name ----------- diff --git a/Lib/re/_parser.py b/Lib/re/_parser.py index 3c41c43409534b..b8c19cd3070c4d 100644 --- a/Lib/re/_parser.py +++ b/Lib/re/_parser.py @@ -625,6 +625,12 @@ def _parse(source, state, verbose, nested, first=False): subpatternappend((NOT_LITERAL, set[0][1])) else: subpatternappend(set[0]) + elif _len(set) == 1 and set[0][0] is CATEGORY: + # optimization: a lone category like [\d] or [^\d] + if negate: + subpatternappend((CATEGORY, CH_NEGATE[set[0][1]])) + else: + subpatternappend(set[0]) else: if negate: set.insert(0, (NEGATE, None)) diff --git a/Misc/NEWS.d/next/Library/2026-06-24-10-30-00.gh-issue-152056.Qk7mZ2.rst b/Misc/NEWS.d/next/Library/2026-06-24-10-30-00.gh-issue-152056.Qk7mZ2.rst new file mode 100644 index 00000000000000..6e71d720cd19be --- /dev/null +++ b/Misc/NEWS.d/next/Library/2026-06-24-10-30-00.gh-issue-152056.Qk7mZ2.rst @@ -0,0 +1,5 @@ +Optimize matching of a character set that contains a single character +category, such as ``[\d]`` or ``[^\s]``: it is now compiled to a single +``CATEGORY`` opcode, the same as the corresponding ``\d`` or ``\S`` escape, +instead of being wrapped in an ``IN`` block. This speeds up matching and +reduces the size of the compiled byte code. Patch by Pieter Eendebak. _______________________________________________ Python-checkins mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3//lists/python-checkins.python.org Member address: [email protected]
