You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@stdcxx.apache.org by Liviu Nicoara <ni...@hates.ms> on 2012/09/23 17:34:26 UTC
[PATCH] STDCXX-853
Umm, I didn't think to search for a corresponding incident and I considered the defect to be so minor as to not warrant an issue. The following patch has been applied already on 4.2.x:
Index: tests/support/atomic_xchg.cpp
===================================================================
--- tests/support/atomic_xchg.cpp (revision 1388732)
+++ tests/support/atomic_xchg.cpp (revision 1388733)
@@ -297,7 +297,7 @@ void run_test (intT, thr_args_base::tag_
// compute the expected result, "skipping" zeros by incrementing
// expect twice when it overflows and wraps around to 0 (zero is
// used as the lock variable in thread_routine() above)
- intT expect = intT (1);
+ intT volatile expect = intT (1);
const unsigned long nincr = (Args::nthreads_ * Args::nincr_) / 2U;
Re: [PATCH] STDCXX-853
Posted by Liviu Nicoara <ni...@hates.ms>.
On 09/24/12 01:18, Travis Vitek wrote:
>
> Liviu,
>
> Should the volatile be to the left of the intT typename here? I know it is equivalent, but it is weird to look at the line of code below and see that we're following two different conventions.
>
Thanks, will do.
> Travis
> _______________________________________
> From: Liviu Nicoara
> Sent: Sunday, September 23, 2012 8:34 AM
> To: dev@stdcxx.apache.org
> Subject: [PATCH] STDCXX-853
>
> Umm, I didn't think to search for a corresponding incident and I considered the defect to be so minor as to not warrant an issue. The following patch has been applied already on 4.2.x:
>
> Index: tests/support/atomic_xchg.cpp
> ===================================================================
> --- tests/support/atomic_xchg.cpp (revision 1388732)
> +++ tests/support/atomic_xchg.cpp (revision 1388733)
> @@ -297,7 +297,7 @@ void run_test (intT, thr_args_base::tag_
> // compute the expected result, "skipping" zeros by incrementing
> // expect twice when it overflows and wraps around to 0 (zero is
> // used as the lock variable in thread_routine() above)
> - intT expect = intT (1);
> + intT volatile expect = intT (1);
>
> const unsigned long nincr = (Args::nthreads_ * Args::nincr_) / 2U;
>
--
And now I see with eye serene
The very pulse of the machine.
RE: [PATCH] STDCXX-853
Posted by Travis Vitek <Tr...@roguewave.com>.
Liviu,
Should the volatile be to the left of the intT typename here? I know it is equivalent, but it is weird to look at the line of code below and see that we're following two different conventions.
Travis
_______________________________________
From: Liviu Nicoara
Sent: Sunday, September 23, 2012 8:34 AM
To: dev@stdcxx.apache.org
Subject: [PATCH] STDCXX-853
Umm, I didn't think to search for a corresponding incident and I considered the defect to be so minor as to not warrant an issue. The following patch has been applied already on 4.2.x:
Index: tests/support/atomic_xchg.cpp
===================================================================
--- tests/support/atomic_xchg.cpp (revision 1388732)
+++ tests/support/atomic_xchg.cpp (revision 1388733)
@@ -297,7 +297,7 @@ void run_test (intT, thr_args_base::tag_
// compute the expected result, "skipping" zeros by incrementing
// expect twice when it overflows and wraps around to 0 (zero is
// used as the lock variable in thread_routine() above)
- intT expect = intT (1);
+ intT volatile expect = intT (1);
const unsigned long nincr = (Args::nthreads_ * Args::nincr_) / 2U;