Conversation
Domiii
force-pushed
the
dominik/tt-1480-assert-more-counter-code
branch
from
5310396 to
59ad771
Compare
…-1480-assert-more-counter-code
…-1480-assert-more-counter-code
…-1480-assert-more-counter-code
…-1480-assert-more-counter-code
…-1480-assert-more-counter-code
| HeapVector<Member<ContextLifecycleObserver>> observers; | ||
| observers_.ForEachObserver([&](ContextLifecycleObserver* observer) { | ||
| observers.push_back(observer); | ||
| observers_.ForEachObserver([](ContextLifecycleObserver* observer) { |
There was a problem hiding this comment.
NOTE: Undid outdated changes. This looks like a very old version of assuring determinism of a HeapObserverSet. I already fixed most of that inside HeapObserverSet itself a long time ago. I added the additional "leaking on Remove" part down below.
|
I closed the wrong thing... |
…-1480-assert-more-counter-code
…-1480-assert-more-counter-code
toshok
left a comment
toshok
left a comment
There was a problem hiding this comment.
either a naming problem or a logic problem (guessing the former) in the heap_observer_set.h change. either way, a little confused there.
| feature.type(), | ||
| feature.value()); | ||
| REPLAY_ASSERT_IF( | ||
| // [Rebase-Check] This might change with a rebase. |
There was a problem hiding this comment.
what exactly might change? can the answer go in the comment?
| CHECK(iteration_state_ & kAllowingRemoval); | ||
| if (recordreplay::AreEventsDisallowed() && | ||
| recordreplay::IsRecordingOrReplaying("leak-references", "HeapObserverSet")) { | ||
| return; |
There was a problem hiding this comment.
🤔 at "leak-references". this does more than leak/keep-alive: the observer stays around in the list of observers to notify.
2 participants
Uh oh!
There was an error while loading. Please reload this page.