fix(memory): reject a cleanup_interval that is not positive - #198
Merged
Merged
Conversation
With 0 or a negative value asyncio.sleep() returned at once, so the cleanup loop spun on a full CPU core and took the cache lock on every pass. MemoryBackend now raises ValueError for it. Closes #180
allen0099
added
bug
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #180.
Problem
MemoryBackend(cleanup_interval=0), or any negative value, was accepted.asyncio.sleep()returns at once for these values, so the cleanup loop never yielded for long. Measured over 0.5 s:cleanup_interval0-560Each sweep takes the cache lock and scans every entry, so requests also contend with it.
Change
MemoryBackend.__init__raisesValueErrorwhencleanup_interval <= 0.docs/BACKENDS.md(en and zh-TW) and the CHANGELOG are updated.Tests
test_cleanup_interval_must_be_positive, parametrized over0and-5.memory.pycoverage stays at 100%. No live servers were run, since the change is memory-only.