Expressions referring to different bound fields should not be
considered equal.
Thanks Julien Enselme for the detailed report.
Regression in bc7e288ca9554ac1a0a19941302dea19df1acd21.
An HTTP request would not be redirected to HTTPS when the
SECURE_PROXY_SSL_HEADER and SECURE_SSL_REDIRECT settings were used if
the proxy connected to Django via HTTPS.
HttpRequest.scheme will now always trust the SECURE_PROXY_SSL_HEADER if
set, rather than falling back to the request scheme when the
SECURE_PROXY_SSL_HEADER did not have the secure value.
Thanks to Gavin Wahl for the report and initial patch suggestion, and
Shai Berger for review.
Previous behavior was inconsistent with django.urls.reverse() and
caused that translate_url() created an incorrect URL when an optional
parameter was missing.
Previously, when running the Django test suite with warnings enabled,
the following was emitted:
/usr/lib64/python3.7/urllib/parse.py:915: BytesWarning: str() on a bytearray instance
v = quote_via(str(v), safe, encoding, errors)
This occurred due to the bytearray() being passed to
urllib.parse.urlencode() which eventually calls str() on it. The test
does not represent desired real world behavior. Rather than test for and
assert strange unspecified behavior that emits a warning, remove it.
This was also discussed in PR #11374.
email.headerregistry.parser.get_mailbox() returns a token with a `token_type` attribute.
If `token_type` is `’invalid-mailbox’` then RFC violations have been detected. Emails with only the local part, and no domain, are correctly parsed but are marked as `’invalid-mailbox’`.
As per #15042, local-only are supported, to enable sending to addresses on localhost.
sanitize_email() does not currently check `token_type`. This test is added to avoid a regression in case this is revisited in the future.
intword and filesizeformat passed floats to ngettext() which is
deprecated in Python 3.7. The rationale for this warning is documented
in BPO-28692: https://bugs.python.org/issue28692.
For filesizeformat, the filesize value is expected to be an int -- it
fills %d string formatting placeholders. It was likely coerced to a
float to ensure floating point division on Python 2. Python 3 always
does floating point division, so coerce to an int instead of a float to
fix the warning.
For intword, the number may contain a decimal component. In English, a
decimal component makes the noun plural. A helper function,
round_away_from_one(), was added to convert the float to an integer that
is appropriate for ngettext().
