runtime: don't exit the main thread if a coexp attempted to gc it #362
Conversation
There was a problem hiding this comment.
I think the headline comment is misleading. It says we should never GC the main thread, which is obviously wrong. I think the mod is attempting to stop the recovery of the activation block for the main thread (which makes sense, because nobody activated it).
However, the big problem is that the ci fails. I put the patch into my (up to date) master branch on my local machine and it failed there too.
src/runtime/rmemmgt.r
Outdated
| @@ -1535,6 +1535,8 @@ static void cofree() | |||
|
|
|||
| xep = *ep; | |||
| *ep = (*ep)->nextstk; | |||
| if (xep->program == &rootpstate || xep->id == 1) | |||
There was a problem hiding this comment.
This causes build errors for some configurations.
…to gc it. Signed-off-by: Jafar Al-Gharaibeh <[email protected]>
Jafaral
force-pushed
the
fix-program-gc
branch
from
f87e9ee to
30ebc09
Compare
| #endif /* NO_COEXPR_SEMAPHORE_FIX */ | ||
| pthread_exit(0); | ||
| } else if (cp->id == 1 && cp->id == curtstate->c->id) { | ||
| /* FIXME: Main thread, should just return? */ |
There was a problem hiding this comment.
Should this FIXME be here? If yes, it's new code; why aren't we fixing it now?
There was a problem hiding this comment.
Left it there intentionally because I wasn't there if this has any side effects.
Jafaral
changed the title
runtime: don't exit the main thread if a coexp attempted to gc it
No description provided.