Introduce absolute cached property #1100
Conversation
We call is_absolute quite often internally and externally. Curently this requires re-parsing the netloc. To make the API more consistent and perform better, absolute is now preferred over is_absolute. is_absolute is retained for backwards compatibility.
psf-chronographer
bot
added
the
bot:chronographer:provided
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #1100 +/- ##
==========================================
+ Coverage 94.90% 94.91% +0.01%
==========================================
Files 30 30
Lines 4358 4367 +9
Branches 382 383 +1
==========================================
+ Hits 4136 4145 +9
Misses 196 196
Partials 26 26
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
|
Getting closer to the point where |
| @@ -1307,26 +1307,31 @@ def test_with_suffix_replace(): | |||
| def test_is_absolute_for_relative_url(): | |||
| url = URL("/path/to") | |||
| assert not url.is_absolute() | |||
| assert not url.absolute | |||
There was a problem hiding this comment.
@bdraco cached things are usually tested through their __wrapped__ attributes to prevent caching from influencing behavior in tests. Have you considered doing something like this?
There was a problem hiding this comment.
Since the URL object is immutable and the state is already predefined we shouldn’t have much risk of the cached value being calculated incorrectly.
Was there a specific case you are concerned about?
There was a problem hiding this comment.
No, just got surprised seeing this and thought I'd mention it.
What do these changes do?
We call
is_absolutequite often internally and externally. Currently this requires re-parsing thenetloc. To make the API more consistent and perform better,absoluteis now preferred overis_absolute.is_absoluteis retained for backwards compatibility.Are there changes in behavior for the user?
absoluteis now preferred overis_absolute, however there are no plans to removeis_absolute.Probably not the best representation of a microbenchmark since
raw_hostwould previous be a cache miss most of the time and this change will perform much better either way:On previous version when
raw_hostis a cache miss