Improve leaked thread detector #1435
Merged
Conversation
This file contains 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
marcotc
previously approved these changes
Mar 30, 2021
While I was chasing down a leaked thread, I gathered a few
improvements. These mostly affect tests that finish too fast, my
test case was:
```ruby
RSpec.describe "A short-lived thread" do
it 'lives for a really short while' do
Thread.new { }
end
end
```
Improvements:
* Set the `@caller` before the thread actually starts running. This
works fine (we're using a similar construct in the profiler), and
ensures that the `@caller` is available from the moment the thread
gets created, instead of being a race.
I observed that if a test run too fast, it was possible that
`@caller` was not set because the thread still hadn't run by the
time the detector started working.
* Call `Thread.pass` whenever a backtrace is not available -- this
tries to give newly-created threads a chance to run, so that we
can have a backtrace for them.
* Detect when a backtrace is missing because a thread has died.
ivoanjo
force-pushed
the
ivoanjo/improve-leaked-thread-detector
branch
from
18fbfc3 to
3edae2e
Compare
marcotc
approved these changes
Apr 13, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
While I was chasing down a leaked thread, I gathered a few improvements. These mostly affect tests that finish too fast, my
test case was:
Improvements:
Set the
@callerbefore the thread actually starts running. This works fine (we're using a similar construct in the profiler), and ensures that the@calleris available from the moment the thread gets created, instead of being a race.I observed that if a test run too fast, it was possible that
@callerwas not set because the thread still hadn't run by the time the detector started working.Call
Thread.passwhenever a backtrace is not available -- this tries to give newly-created threads a chance to run, so that we can have a backtrace for them.Detect when a backtrace is missing because a thread has died.