[cpp-threads] Belated comments on N2052

[Nelson, Clark]

> From: cpp-threads-bounces at decadentplace.org.uk 
> [mailto:cpp-threads-bounces at decadentplace.org.uk] On Behalf 
> Of Peter Dimov
> Sent: Monday, October 23, 2006 15:41
> 
> * 1.9p17: the example seems incorrect; increment_x doesn't 
> return a value, 

Thanks; I've fixed the definition of increment_x (and renamed it inc_x).

> and the comment "x may be incremented only once" looks wrong.

I'm not sure what you mean by "looks wrong". In response to other
feedback I've received, the comment now reads, "Effect of inc_x may be
lost: call may occur between load and store of x". From your
perspective, is that better or worse?

> * 1.10p8: "for any evaluation X" is ambiguous, can mean 
> "there exists an 
> evaluation X such that..." or "for all evaluations X..."

That now reads "for some evaluation X".

> * 17.4.4.8+ (library):
> 
> "Library calls do not introduce synchronizes-with relationships"
> 
> ... unless specified otherwise?

No doubt. But I'm expecting (hoping) that someone from the library group
will step up and take over drafting of what needs to be said about the
library. I'm really not the guy for that job.

Clark