Bugzilla

(In reply to Petr Sumbera from comment #7)
> For recorrd. I can still avoid the issue with lastest sources by reverting problematic commit:
> ```
> -  if (AppShutdown::IsInOrBeyond(ShutdownPhase::AppShutdownConfirmed)) {
> -    Cancel(NS_ERROR_ILLEGAL_DURING_SHUTDOWN,
> -           "Aborting OnStartRequest after shutdown started."_ns);
> -    return NS_OK;
> -  }
> ```

I assume only this part matters here. Extensions have [this two-step shutdown blockers](https://searchfox.org/mozilla-central/rev/d0cd6a42de9cbacd6d8f3253e2edc18076386282/toolkit/mozapps/extensions/internal/XPIProvider.jsm#2594-2639) that relies on sending something to the content process running the extension. Olli, could it be that we break that logic here ?

(In reply to Petr Sumbera from comment #7)
> For recorrd. I can still avoid the issue with lastest sources by reverting problematic commit:
> ```
> -  if (AppShutdown::IsInOrBeyond(ShutdownPhase::AppShutdownConfirmed)) {
> -    Cancel(NS_ERROR_ILLEGAL_DURING_SHUTDOWN,
> -           "Aborting OnStartRequest after shutdown started."_ns);
> -    return NS_OK;
> -  }
> ```

I assume only this part matters here. Extensions have [this two-step shutdown blockers](https://searchfox.org/mozilla-central/rev/d0cd6a42de9cbacd6d8f3253e2edc18076386282/toolkit/mozapps/extensions/internal/XPIProvider.jsm#2594-2639) that relies on sending something to the content process running the extension. Olli, could it be that we break that logic here ? Basically we did this to avoid we try to create new processes, but here we probably affect existing ones, too?