diff options
| author | David Baker <dave@matrix.org> | 2017-10-05 11:57:43 +0100 |
|---|---|---|
| committer | David Baker <dave@matrix.org> | 2017-10-05 11:57:43 +0100 |
| commit | 0c8da8b519fbd8bca984117e354fe57c3a76e154 (patch) | |
| tree | 11ba51e0f9f9464e9522ed0d6dff943383efd57d | |
| parent | pep8 (diff) | |
| download | synapse-0c8da8b519fbd8bca984117e354fe57c3a76e154.tar.xz | |
Use better method for word boundary searching
From https://github.com/matrix-org/matrix-js-sdk/commit/ebc95667b8a5777d13e5d3c679972bedae022fd5
| -rw-r--r-- | synapse/push/push_rule_evaluator.py | 14 |
1 files changed, 3 insertions, 11 deletions
diff --git a/synapse/push/push_rule_evaluator.py b/synapse/push/push_rule_evaluator.py index b78f2d90d7..65f9a63fd8 100644 --- a/synapse/push/push_rule_evaluator.py +++ b/synapse/push/push_rule_evaluator.py @@ -26,8 +26,6 @@ logger = logging.getLogger(__name__) GLOB_REGEX = re.compile(r'\\\[(\\\!|)(.*)\\\]') IS_GLOB = re.compile(r'[\?\*\[\]]') INEQUALITY_EXPR = re.compile("^([=<>]*)([0-9]*)$") -STARTS_WITH_WORD_CHAR_REGEX = re.compile(r"^\w") -ENDS_WITH_WORD_CHAR_REGEX = re.compile(r"\w$") def _room_member_count(ev, condition, room_member_count): @@ -209,15 +207,9 @@ def _re_word_boundary(r): but do so respecting the fact that strings starting or ending with non-word characters will change word boundaries. """ - # Matching a regex string aginst a regex, since by definition - # \b is the boundary between a \w and a \W, so match \w at the - # start or end of the expression (although this will miss, eg. - # "[dl]og") - if STARTS_WITH_WORD_CHAR_REGEX.search(r): - r = r"\b%s" % (r,) - if ENDS_WITH_WORD_CHAR_REGEX.search(r): - r = r"%s\b" % (r,) - return r + # we can't use \b as it chokes on unicode. however \W seems to be okay + # as shorthand for [^0-9A-Za-z_]. + return r"(^|\W)%s(\W|$)" % (r,) def _flatten_dict(d, prefix=[], result=None): |