Just as it says on the tin.
Maybe change indent to be consistent with the other fields?
Maybe simplify to just interruptible?
Maybe simplify to just return the condition?
This is unrelated to your changes, but maybe remove the switch (potentially leaving the comment)?
Seems like this could also be simplified.
Maybe remove == true?
Those should be unnecessary (or am I missing something?)
This is great! I never liked rtsBool :)
I haven't looked through all of it, but there are a couple of concerns inline.
|286 ↗||(On Diff #9393)|
|2111–2120 ↗||(On Diff #9393)|
This is a Win32 API and probably doesn't use bool.
Is this really legit? This is a Win32 API and probably doesn't use bool, so we might be getting implicit promotion.
I think I'll leave this one; if nothing else it's a reminder that cas returns a bool.
Yes, this is actually a separate commit locally since it wasn't entirely obvious what was going on. The definition of GarbageCollect in GC.c disagreed with this declaration. This change brings the two into agreement. I checked the call-sites and they should all still be correct.
Fair point. I think they are semantically equivalent, but perhaps we should play it safe.
Bah. Thanks for catching these.
- Bulk replacement
- rts: Use stdbool
- ghc/hschooks.c: Fix include path of Rts.h
- Fix type of GarbageCollect declaration
- michalt's comments
- Simon's comments
- More Win32 silliness
- rts: Fix bell
Strangely enough I'm seeing the RtsFlags.GcFlags.ringBell get set from defaultsHook in the stage1 compiler. This results in validation failure due to unexpected ^G output on stderr. I'm a bit baffled by this, although have a sneaking suspicion that the problem lies in deriveConstants, which uses the wrong width for bool. However, exactly how this happens isn't clear on reading the code.