| Age | Commit message (Collapse) | Author |
|
MultiPartParser.
The mitigation for CVE-2026-33033 aligned base64 data to a multiple of
4 by reading whole chunks. As the chunk size is itself a multiple of 4,
those reads never restored the alignment, so a misaligned upload was
read and decoded in full before reaching upload handlers.
Trailing undecoded bytes are now carried over to the next chunk, so no
extra reads are needed and memory use stays bounded by the chunk size.
Thanks to Ignacy Kujawa for the report, and Jake Howard for the review.
|
|
prefixes.
|
|
functools.cached_property.
Since Python 3.12, `functools.cached_property` no longer takes a lock
(https://github.com/python/cpython/issues/87634), which was the main
objection to using it instead of Django's own implementation.
`Prefetch(to_attr=...)` and the admindocs model detail page now accept
`functools.cached_property` as well as Django's own `cached_property`.
Co-authored-by: Natalia <124304+nessita@users.noreply.github.com>
|
|
Replaced bare `async for` over a not-locally-defined async iterable with
`async with (maybe_)aclosing(aiter(...)) as it: async for item in it:` at
every consumption site, so the iterator's aclose() runs deterministically
when the consumer exits — instead of being deferred to asyncio's asyncgen
finalizer hook.
Added django.utils.asyncio.maybe_aclosing as a small wrapper that uses
contextlib.aclosing when the iterator has aclose() and contextlib.nullcontext
otherwise (for caller-supplied iterators that may only implement __aiter__
and __anext__).
Sites:
- contrib/auth/backends.py:_aget_permissions (aclosing)
- core/paginator.py:AsyncPage.__aiter__ (maybe_aclosing)
- core/paginator.py:AsyncPage.aget_object_list (maybe_aclosing)
- db/models/query.py:QuerySet.aiterator (aclosing, hoisted across both
prefetch and no-prefetch branches)
- http/response.py:StreamingHttpResponse.streaming_content awrapper
(maybe_aclosing)
- http/response.py:StreamingHttpResponse.__iter__ to_list (maybe_aclosing)
- http/response.py:StreamingHttpResponse.__aiter__ (aclosing)
- shortcuts.py:aget_list_or_404 (aclosing)
- test/client.py:aclosing_iterator_wrapper (maybe_aclosing)
- utils/text.py:acompress_sequence (maybe_aclosing)
|
|
Per DEP 20, Django's version numbers are calendar-based from 2028, so
deprecation warning classes are now named after the calendar version
that removes the feature they mark, and the deprecation timeline is
headed by the same calendar version.
Updated references to RemovedInDjango71Warning to use the new warning,
and adjusted references to "Django 7.1" to "Django 2029".
|
|
The docstring and reference documentation still described the safe
parameter as defaulting to True after its deprecation.
Follow-up to 6e15ac8066312328de279e3e072667416c205bfc.
|
|
Thanks to these people for pointing these out:
* Hojeong Park
* Natalia Bidart
* Sulav Raj Bista
|
|
|
|
Follow-up to 6e15ac8066312328de279e3e072667416c205bfc.
|
|
Made signed cookies derive their signer namespace from an injective
encoding of `(name, salt)` while preserving compatibility with legacy
`name + salt` cookies behind SIGNED_COOKIE_LEGACY_SALT_FALLBACK.
Thanks Peng Zhou for the report, and Shai Berger, Markus Holterman,
Jake Howard, and Paul McMillan for reviews.
Co-authored-by: Jacob Walls <jacobtylerwalls@gmail.com>
Co-authored-by: Natalia <124304+nessita@users.noreply.github.com>
|
|
|
|
Peer frameworks have long since dropped their analogous checks for this
vulnerability that was fixed in ES5.
|
|
|
|
|
|
|
|
HttpResponseRedirect.
|
|
|
|
ASGI requests.
The `body` property in `HttpRequest` checks DATA_UPLOAD_MAX_MEMORY_SIZE
against the declared `Content-Length` header before reading. On the ASGI
path, chunked requests carry no `Content-Length`, so the check evaluated
to 0 and always passed regardless of the actual body size.
This work adds a new check on the actual number of bytes consumed.
Thanks to Superior for the report, and to Jake Howard and Jacob Walls
for reviews.
|
|
When a multipart file part used `Content-Transfer-Encoding: base64` and
the non-whitespace base64 bytes did not align to a multiple of 4 within
a chunk, the parser entered a loop calling `field_stream.read(1-3)` once
per whitespace byte. Each such call fetched the entire internal buffer,
sliced off 1-3 bytes, and pushed the remainder back via unget(), doing
an O(n) memory copy per call. A 2.5 MB payload of mostly whitespace
produced CPU amplification relative to a normal upload of the same size.
The alignment loop now reads `self._chunk_size` bytes at a time, and
accumulates stripped parts in a list joined once at the end.
Thanks to Seokchan Yoon for the report and the fixing patch.
|
|
encoding.
Added LookupError to the except clause so invalid headers are silently
skipped, consistent with other malformed header handling.
|
|
|
|
Refs CVE-2025-64458.
The previous limit of 2048 characters reused the URLValidator constant
and proved too restrictive for legitimate redirects to some third-party
services. This change introduces a separate `MAX_URL_REDIRECT_LENGTH`
constant (defaulting to 16384) and uses it in HttpResponseRedirectBase.
Thanks Jacob Walls for report and review.
|
|
HttpResponseRedirect/HttpResponsePermanentRedirect on Windows.
Thanks Seokchan Yoon for the report, Markus Holtermann for the
triage, and Jake Howard for the review.
Follow-up to CVE-2025-27556 and 39e2297210d9d2938c75fc911d45f0e863dc4821.
|
|
parse_header_parameters for multipart requests.
Header parsing should apply only to the header value. The previous
implementation happened to work but relied on unintended behavior.
|
|
Rewrapped long docstrings and block comments to 79 characters + newline
using script from https://github.com/medmunds/autofix-w505.
|
|
Manually reformatted some long docstrings and comments that would be
damaged by the to-be-applied autofixer script, in cases where editorial
judgment seemed necessary for style or wording changes.
|
|
|
|
When matching which entry in the `Accept` header should be used for
a given media type, the specificity matters. However once those are
resolved, only the quality matters when selecting preference.
Regression in c075508b4de8edf9db553b409f8a8ed2f26ecead.
Thank you to Anders Kaseorg for the report.
|
|
The "q" key was removed while addressing ticket #36411. Despite
`MediaType.params` is undocumented and considered internal, it was used
in third-party projects (Zulip reported breakage), so this work restored
the `q` key in `params`.
Thanks Anders Kaseorg for the report.
Regression in c075508b4de8edf9db553b409f8a8ed2f26ecead.
|
|
parameters.
HttpRequest.get_preferred_type() did not account for parameters in
Accept header media types (e.g., "text/vcard; version=3.0"). This caused
incorrect content negotiation when multiple types differed only by
parameters, reducing specificity as per RFC 7231 section 5.3.2
(https://datatracker.ietf.org/doc/html/rfc7231.html#section-5.3.2).
This fix updates get_preferred_type() to treat media types with
parameters as distinct, allowing more precise and standards-compliant
matching.
Thanks to magicfelix for the report, and to David Sanders and Sarah
Boyce for the reviews.
|
|
|
|
datetime.UTC was added in Python 3.11.
|
|
HttpResponseRedirectBase.
Co-authored-by: Natalia <124304+nessita@users.noreply.github.com>
|
|
Signed-off-by: SaJH <wogur981208@gmail.com>
|
|
|
|
StreamingHttpResponse.__iter__()/__aiter__().
|
|
This work should not generate any change of functionality, and
`urlsplit` is approximately 6x faster.
Most use cases of `urlparse` didn't touch the path, so they can be
converted to `urlsplit` without any issue. Most of those which do use
`.path`, simply parse the URL, mutate the querystring, then put them
back together, which is also fine (so long as urlunsplit is used).
|
|
https://github.com/psf/black/releases/tag/24.1.0
|
|
headers.
This also allow customizing the maximum size of headers via
MAX_TOTAL_HEADER_SIZE.
|
|
application/x-www-form-urlencoded content type.
Thanks Eki Xu for the report.
|
|
Use the capture groups from the regular expression that has already been
matched to avoid resplitting and the need to special case for IPv6.
|
|
subclasses."
This reverts commit d7f5bfd241666c0a76e90208da1e9ef81aec44db.
Thanks Márton Salomváry for the report.
|
|
HttpRequest and subclasses."
This reverts commit 6220c445c40a6a7f4d442de8bde2628346153963.
Thanks Adam Johnson and Márton Salomváry for reports.
|
|
compress and brotli.
Thanks Chamal De Silva for the report.
|
|
Thanks to Jakob Ackermann for the report.
|
|
|
|
Available since Python 3.10.
|
|
Thanks to Florian Vazelle for initial exploratory work, and to Nick
Pope and Mariusz Felisiak for review.
|
|
|
|
|