| Age | Commit message (Collapse) | Author |
|---|
|
The current implementation of LimitedStream is slow because .read()
performs an extra copy into a buffer and .readline() performs two
extra copies. The stream being wrapped is already typically a BytesIO
object so this is unnecessary.
This implementation has largely been untouched for 12 years and,
inspired by a simpler implementation in werkzeug, it was possible to
achieve the following performance improvement:
LimitedStream.read() (single line):
Mean +- std dev: [bench_limitedstream-main] 286 ns +- 6 ns
-> [bench_limitedstream-patch] 227 ns +- 6 ns: 1.26x faster
LimitedStream.readline() (single line):
Mean +- std dev: [bench_limitedstream-main] 507 ns +- 11 ns
-> [bench_limitedstream-patch] 232 ns +- 8 ns: 2.18x faster
LimitedStream.read(8192) (single line):
Mean +- std dev: [bench_limitedstream-main] 360 ns +- 8 ns
-> [bench_limitedstream-patch] 297 ns +- 6 ns: 1.21x faster
LimitedStream.readline(8192) (single line):
Mean +- std dev: [bench_limitedstream-main] 602 ns +- 10 ns
-> [bench_limitedstream-patch] 305 ns +- 10 ns: 1.98x faster
LimitedStream.read() (multiple lines):
Mean +- std dev: [bench_limitedstream-main] 290 ns +- 5 ns
-> [bench_limitedstream-patch] 236 ns +- 6 ns: 1.23x faster
LimitedStream.readline() (multiple lines):
Mean +- std dev: [bench_limitedstream-main] 517 ns +- 19 ns
-> [bench_limitedstream-patch] 239 ns +- 7 ns: 2.16x faster
LimitedStream.read(8192) (multiple lines):
Mean +- std dev: [bench_limitedstream-main] 363 ns +- 8 ns
-> [bench_limitedstream-patch] 311 ns +- 11 ns: 1.17x faster
LimitedStream.readline(8192) (multiple lines):
Mean +- std dev: [bench_limitedstream-main] 601 ns +- 12 ns
-> [bench_limitedstream-patch] 308 ns +- 7 ns: 1.95x faster
Geometric mean: 1.59x faster
|
|
FakePayload is a wrapper around io.BytesIO and is expected to
masquerade as though it is a file-like object. For that reason it makes
sense that it should inherit the correct signatures from io.BytesIO
methods.
Crucially an implementation of .readline() is added which will be
necessary for this to behave more like the expected file-like objects as
LimitedStream will be changed to defer to the wrapped stream object
rather than rolling its own implementation for improved performance.
It should be safe to adjust these signatures because FakePayload is
only used internally within test client helpers, is undocumented, and
thus private.
|
|
interval longer than 1 month.
Regression in 8d67e16493c903adc9d049141028bc0fff43f8c8.
|
|
mysqlclient 1.4.3 is the first release to support Python 3.8.
|
|
Instead of the separate property, we can just not set self.timezone if
the datetime is ambiguous or imaginary. This ensures that this check
will only ever happen once as it's dependant on the datetime object and
not the format string characters.
|
|
|
|
O() - we should try to avoid calling specifier methods from each other
to avoid extra function call overhead. In addition we end up, in this
case, duplicating the ambiguous/imaginary datetime checks. We're also
going to be looking at simplifying things by having all of these
specifier methods return strings and not an random mix of types.
t() - the value can only be one of 28, 29, 30, or 31. As such, there is
no need to zero-pad to a width of two.
|
|
Avoid extra attribute lookup in specifier methods, etc. by importing
classes from datetime directly.
|
|
This removes redundant get_format() calls and passing a default value
for the format argument.
|
|
|
|
|
|
|
|
|
|
PostGIS.
|
|
This adds ManifestFilesMixin.manifest_hash attribute exposing a "hash"
of the full manifest. This allows applications to determine when their
static files have changed.
|
|
|
|
UserCreationForm.
Co-Authored-By: Neven Mundar <nmundar@gmail.com>
|
|
|
|
Co-authored-by: Mariusz Felisiak <felisiak.mariusz@gmail.com>
|
|
Thanks Jared Chung, Tom Carrick, David Smith, Nick Pope, and Mariusz
Felisiak for reviews.
Co-authored-by: Mariusz Felisiak <felisiak.mariusz@gmail.com>
Co-authored-by: Nick Pope <nick@nickpope.me.uk>
|
|
MySQL < 8.0.16.
|
|
FilteredRelations to the OneToOneField.
|
|
|
|
SQLCompiler.get_related_selections() to avoid redefinition.
|
|
|
|
Thanks to Florian Vazelle for initial exploratory work, and to Nick
Pope and Mariusz Felisiak for review.
|
|
This moves setting a database collation to the column type alteration
as both must be set at the same time.
This should also avoid another layer of the column type alteration when
adding database comments support (#18468).
|
|
Updated Bulgarian, Esperanto, Hungarian, Japanese, Macedonian, Persian,
Portuguese (Brazil), Russian, Spanish, and Turkmen translations.
Forwardport of 46b28bbe151282e9198b719d00f84b48c92606ba from stable/4.1.x.
|
|
Thanks to Mariusz Felisiak for review.
|
|
|
|
|
|
Co-authored-by: Sin-Woo Bang <sinwoobang@gmail.com>
|
|
Thanks Simon Charette, Tim Graham, and Adam Johnson for reviews.
Co-authored-by: Florian Apolloner <florian@apolloner.eu>
Co-authored-by: Mariusz Felisiak <felisiak.mariusz@gmail.com>
|
|
condition.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
__len lookup.
Regression in 88fc9e2826044110b7b22577a227f122fe9c1fb5 that began
manifesting in Django 4.1.
|
|
Signer/TimestampSigner.
Thanks Jacob Walls for the implementation idea.
|
|
internal_size is size for fixed-size types not for char types.
|
|
|
|
Co-authored-by: James Bligh <james.bligh@silvercloudhealth.com>
|
|
- Used connection.info instead of connection.get_parameter_status() and
connection.server_info which don't exist in psycopg 3.
- Set encoding using the client_encoding connection parameter instead
of connection.set_client_encoding() that doesn't exist in psycopg 3.
- Used the dbname connection parameter instead of deprecated
alias - database.
|
|
|
|
MySQL.
|
|
|
|
to JSONField & co.
JSON should be provided as literal Python objects an not in their
encoded string literal forms.
|
