Fixed #35252 -- Optimized _route_to_regex().

co-authored-by: Nick Pope <nick@nickpope.me.uk>
This commit is contained in:
Adam Johnson 2024-02-26 10:41:15 +00:00 committed by Mariusz Felisiak
parent 241adf678f
commit eff21d8e7a
4 changed files with 34 additions and 33 deletions

View File

@ -68,11 +68,11 @@ def register_converter(converter, type_name):
REGISTERED_CONVERTERS[type_name] = converter() REGISTERED_CONVERTERS[type_name] = converter()
get_converters.cache_clear() get_converters.cache_clear()
from django.urls.resolvers import _route_to_regex
_route_to_regex.cache_clear()
@functools.cache @functools.cache
def get_converters(): def get_converters():
return {**DEFAULT_CONVERTERS, **REGISTERED_CONVERTERS} return {**DEFAULT_CONVERTERS, **REGISTERED_CONVERTERS}
def get_converter(raw_converter):
return get_converters()[raw_converter]

View File

@ -26,7 +26,7 @@ from django.utils.http import RFC3986_SUBDELIMS, escape_leading_slashes
from django.utils.regex_helper import _lazy_re_compile, normalize from django.utils.regex_helper import _lazy_re_compile, normalize
from django.utils.translation import get_language from django.utils.translation import get_language
from .converters import get_converter from .converters import get_converters
from .exceptions import NoReverseMatch, Resolver404 from .exceptions import NoReverseMatch, Resolver404
from .utils import get_callable from .utils import get_callable
@ -243,7 +243,10 @@ _PATH_PARAMETER_COMPONENT_RE = _lazy_re_compile(
r"<(?:(?P<converter>[^>:]+):)?(?P<parameter>[^>]+)>" r"<(?:(?P<converter>[^>:]+):)?(?P<parameter>[^>]+)>"
) )
whitespace_set = frozenset(string.whitespace)
@functools.lru_cache
def _route_to_regex(route, is_endpoint): def _route_to_regex(route, is_endpoint):
""" """
Convert a path pattern into a regular expression. Return the regular Convert a path pattern into a regular expression. Return the regular
@ -251,40 +254,37 @@ def _route_to_regex(route, is_endpoint):
For example, 'foo/<int:pk>' returns '^foo\\/(?P<pk>[0-9]+)' For example, 'foo/<int:pk>' returns '^foo\\/(?P<pk>[0-9]+)'
and {'pk': <django.urls.converters.IntConverter>}. and {'pk': <django.urls.converters.IntConverter>}.
""" """
original_route = route
parts = ["^"] parts = ["^"]
all_converters = get_converters()
converters = {} converters = {}
while True: previous_end = 0
match = _PATH_PARAMETER_COMPONENT_RE.search(route) for match_ in _PATH_PARAMETER_COMPONENT_RE.finditer(route):
if not match: if not whitespace_set.isdisjoint(match_[0]):
parts.append(re.escape(route))
break
elif not set(match.group()).isdisjoint(string.whitespace):
raise ImproperlyConfigured( raise ImproperlyConfigured(
"URL route '%s' cannot contain whitespace in angle brackets " f"URL route {route!r} cannot contain whitespace in angle brackets <…>."
"<…>." % original_route
) )
parts.append(re.escape(route[: match.start()])) # Default to make converter "str" if unspecified (parameter always
route = route[match.end() :] # matches something).
parameter = match["parameter"] raw_converter, parameter = match_.groups(default="str")
if not parameter.isidentifier(): if not parameter.isidentifier():
raise ImproperlyConfigured( raise ImproperlyConfigured(
"URL route '%s' uses parameter name %r which isn't a valid " f"URL route {route!r} uses parameter name {parameter!r} which "
"Python identifier." % (original_route, parameter) "isn't a valid Python identifier."
) )
raw_converter = match["converter"]
if raw_converter is None:
# If a converter isn't specified, the default is `str`.
raw_converter = "str"
try: try:
converter = get_converter(raw_converter) converter = all_converters[raw_converter]
except KeyError as e: except KeyError as e:
raise ImproperlyConfigured( raise ImproperlyConfigured(
"URL route %r uses invalid converter %r." f"URL route {route!r} uses invalid converter {raw_converter!r}."
% (original_route, raw_converter)
) from e ) from e
converters[parameter] = converter converters[parameter] = converter
parts.append("(?P<" + parameter + ">" + converter.regex + ")")
start, end = match_.span()
parts.append(re.escape(route[previous_end:start]))
previous_end = end
parts.append(f"(?P<{parameter}>{converter.regex})")
parts.append(re.escape(route[previous_end:]))
if is_endpoint: if is_endpoint:
parts.append(r"\Z") parts.append(r"\Z")
return "".join(parts), converters return "".join(parts), converters

View File

@ -393,6 +393,9 @@ Miscellaneous
:py:class:`html.parser.HTMLParser` subclasses. This results in a more robust :py:class:`html.parser.HTMLParser` subclasses. This results in a more robust
and faster operation, but there may be small differences in the output. and faster operation, but there may be small differences in the output.
* The undocumented ``django.urls.converters.get_converter()`` function is
removed.
.. _deprecated-features-5.1: .. _deprecated-features-5.1:
Features deprecated in 5.1 Features deprecated in 5.1

View File

@ -246,14 +246,12 @@ class SimplifiedURLTests(SimpleTestCase):
path("foo", EmptyCBV()) path("foo", EmptyCBV())
def test_whitespace_in_route(self): def test_whitespace_in_route(self):
msg = ( msg = "URL route %r cannot contain whitespace in angle brackets <…>"
"URL route 'space/<int:num>/extra/<str:%stest>' cannot contain "
"whitespace in angle brackets <…>"
)
for whitespace in string.whitespace: for whitespace in string.whitespace:
with self.subTest(repr(whitespace)): with self.subTest(repr(whitespace)):
with self.assertRaisesMessage(ImproperlyConfigured, msg % whitespace): route = "space/<int:num>/extra/<str:%stest>" % whitespace
path("space/<int:num>/extra/<str:%stest>" % whitespace, empty_view) with self.assertRaisesMessage(ImproperlyConfigured, msg % route):
path(route, empty_view)
# Whitespaces are valid in paths. # Whitespaces are valid in paths.
p = path("space%s/<int:num>/" % string.whitespace, empty_view) p = path("space%s/<int:num>/" % string.whitespace, empty_view)
match = p.resolve("space%s/1/" % string.whitespace) match = p.resolve("space%s/1/" % string.whitespace)