Fix one more lifetime bug in ensure_started for cases when the sender is dropped without connecting it
#797
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
msimberg
force-pushed
the
more-ensure-started-fixes
branch
from
72b1197 to
332e62a
Compare
msimberg
force-pushed
the
more-ensure-started-fixes
branch
from
44a44a4 to
12d0086
Compare
|
The failure doesn't reproduce every run in CI, but has now reproduced 2/3 times. I'm going to push the fix next. |
msimberg
force-pushed
the
more-ensure-started-fixes
branch
from
12d0086 to
76e6b0d
Compare
…ed state early with dropped ensure_started senders
aurianer
approved these changes
Oct 4, 2023
2 tasks
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Outdated suggestions cannot be applied.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.
If the
ensure_startedsender is not connected (i.e. there are no continuations) theensure_started_receiverholds the last reference to the shared state. This will be released when releasing the operation state until theensure_startedoperation state, since theensure_started_receiverwill be released with the operation state. This happens because theensure_started_receiveris taken by rvalue reference inset_*. This PR changes them to take the receiver by value, ensuring that the reference count stays non-zero also after the operation state has explicitly been reset.This is another way to fix #795 but I'm keeping that change as it doesn't hurt. Importantly, this fixes a distinct bug that is reported by valgrind in the
std_thread_schedulertest (the test that callsensure_startedbut drops the sender immediately). I've enabled valgrind testing with thestd_thread_scheduler_testin this PR.