I don't think a LOG(WARNING) or LOG(ERROR) is the best approach here. It might be suitable for issues hit in developer-focused feature implementations but many more features will have the potential to hit this and we can't rely on log submissions.

Given that it's not crash-worthy (IMHO) maybe it would make sense to use base::DumpWithoutCrashing() to log these occurrences. This would get us crash reports with full stacks from exactly where a callback was dropped.

SGTM

This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue.

Sorry for the inconvenience if the bug really should have been left as Available.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

DumpWithoutCrashing is actually pretty strongly discouraged now.

I think I'll just make this a CHECK failure, as it should be considered a programmer error in the same vein as a memory leak. I'm terrified of just how many bugs are going to get assigned to me once I land this, but I think it might be better than having random silent failures. :}

Actually, nevermind. It turns out that checking the callback status must be done asynchronously, and generating code to do this on every interface has a non-negligible impact on binary size, so we only generate it with DCHECK on.

I think we have to be OK with this as-is. At least now we have Albatross feeding reports from DCHECK builds in the wild.