| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
The previous `_parse_args()` method pulled the `_parts` out of any
supplied `PurePath` objects; these were subsequently joined in
`_from_parts()` using `os.path.join()`. This is actually a slower form of
joining than calling `fspath()` on the path object, because it doesn't take
advantage of the fact that the contents of `_parts` is normalized!
This reduces the time taken to run `PurePath("foo", "bar") by ~20%, and
the time taken to run `PurePath(p, "cheese")`, where
`p = PurePath("/foo", "bar", "baz")`, by ~40%.
|
@zooba per our discussion in #101362, I've tweaked the docs in this PR. I'm trying to convey that we treat PurePath objects like any other kind of os.PathLike object, which implies we call os.fspath(). Let me know if you'd like us to be more explicit? But I don't want to consume too much docs space on this, as it's a bit off the trodden path, and it stops making sense when paths have drives IMO. |
Sorry, something went wrong.
|
I guess my only question now is whether join handles differing types the same way as parse_args used to? Specifically, each element is converted to a string before being joined, rather than failing with TypeError. If it works the same, then I have no concerns. |
Sorry, something went wrong.
|
os.path.join() requires that you don't mix str and bytes arguments (after fspath()) - it raises TypeError if you do. But it supports subclasses of str and bytes and won't convert them to true strings/bytes. pathlib requires that every argument is a str (after fspath()). It also supports subclasses of str, and will convert them to true strings. I think this patch maintains all previous behaviour. You still get a TypeError if any argument you supply, after fspath(), isn't a subclass of str. |
Sorry, something went wrong.
|
Actually, there don't seem to be good tests for this. I'll write some! |
Sorry, something went wrong.
|
LGTM! |
Sorry, something went wrong.
There was a problem hiding this comment.
Looks good! A few minor suggestions
Sorry, something went wrong.
Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
…lk6mt.rst Co-authored-by: Alex Waygood <Alex.Waygood@Gmail.com>
There was a problem hiding this comment.
Thank you!
Sorry, something went wrong.
|
Thanks so much, Alex! \o/ |
Sorry, something went wrong.
|
Status check is done, and it's a success ✅. |
Sorry, something went wrong.
The previous `_parse_args()` method pulled the `_parts` out of any supplied `PurePath` objects; these were subsequently joined in `_from_parts()` using `os.path.join()`. This is actually a slower form of joining than calling `fspath()` on the path object, because it doesn't take advantage of the fact that the contents of `_parts` is normalized!
This reduces the time taken to run `PurePath("foo", "bar")` by ~20%, and the time taken to run `PurePath(p, "cheese")`, where `p = PurePath("/foo", "bar", "baz")`, by ~40%.
Automerge-Triggered-By: GH:AlexWaygood
* main: (21 commits) pythongh-102192: Replace PyErr_Fetch/Restore etc by more efficient alternatives in sub interpreters module (python#102472) pythongh-95672: Fix versionadded indentation of get_pagesize in test.rst (pythongh-102455) pythongh-102416: Do not memoize incorrectly loop rules in the parser (python#102467) pythonGH-101362: Optimise PurePath(PurePath(...)) (pythonGH-101667) pythonGH-101362: Check pathlib.Path flavour compatibility at import time (pythonGH-101664) pythonGH-101362: Call join() only when >1 argument supplied to pathlib.PurePath() (python#101665) pythongh-102444: Fix minor bugs in `test_typing` highlighted by pyflakes (python#102445) pythonGH-102341: Improve the test function for pow (python#102342) Fix unused classes in a typing test (pythonGH-102437) pythongh-101979: argparse: fix a bug where parentheses in metavar argument of add_argument() were dropped (python#102318) pythongh-102356: Add thrashcan macros to filter object dealloc (python#102426) Move around example in to_bytes() to avoid confusion (python#101595) pythonGH-97546: fix flaky asyncio `test_wait_for_race_condition` test (python#102421) pythongh-96821: Add config option `--with-strict-overflow` (python#96823) pythongh-101992: update pstlib module documentation (python#102133) pythongh-63301: Set exit code when tabnanny CLI exits on error (python#7699) pythongh-101863: Fix wrong comments in EUC-KR codec (pythongh-102417) pythongh-102302 Micro-optimize `inspect.Parameter.__hash__` (python#102303) pythongh-102179: Fix `os.dup2` error reporting for negative fds (python#102180) pythongh-101892: Fix `SystemError` when a callable iterator call exhausts the iterator (python#101896) ...
* main: (37 commits) pythongh-102192: Replace PyErr_Fetch/Restore etc by more efficient alternatives in sub interpreters module (python#102472) pythongh-95672: Fix versionadded indentation of get_pagesize in test.rst (pythongh-102455) pythongh-102416: Do not memoize incorrectly loop rules in the parser (python#102467) pythonGH-101362: Optimise PurePath(PurePath(...)) (pythonGH-101667) pythonGH-101362: Check pathlib.Path flavour compatibility at import time (pythonGH-101664) pythonGH-101362: Call join() only when >1 argument supplied to pathlib.PurePath() (python#101665) pythongh-102444: Fix minor bugs in `test_typing` highlighted by pyflakes (python#102445) pythonGH-102341: Improve the test function for pow (python#102342) Fix unused classes in a typing test (pythonGH-102437) pythongh-101979: argparse: fix a bug where parentheses in metavar argument of add_argument() were dropped (python#102318) pythongh-102356: Add thrashcan macros to filter object dealloc (python#102426) Move around example in to_bytes() to avoid confusion (python#101595) pythonGH-97546: fix flaky asyncio `test_wait_for_race_condition` test (python#102421) pythongh-96821: Add config option `--with-strict-overflow` (python#96823) pythongh-101992: update pstlib module documentation (python#102133) pythongh-63301: Set exit code when tabnanny CLI exits on error (python#7699) pythongh-101863: Fix wrong comments in EUC-KR codec (pythongh-102417) pythongh-102302 Micro-optimize `inspect.Parameter.__hash__` (python#102303) pythongh-102179: Fix `os.dup2` error reporting for negative fds (python#102180) pythongh-101892: Fix `SystemError` when a callable iterator call exhausts the iterator (python#101896) ...
| Back | FazBrowse Home | New Git URL |
The previous _parse_args() method pulled the _parts out of any supplied PurePath objects; these were subsequently joined in _from_parts() using os.path.join(). This is actually a slower form of joining than calling fspath() on the path object, because it doesn't take advantage of the fact that the contents of _parts is normalized!
This reduces the time taken to run PurePath("foo", "bar") by ~20%, and the time taken to run PurePath(p, "cheese"), where p = PurePath("/foo", "bar", "baz"), by ~40%.
Automerge-Triggered-By: GH:AlexWaygood