[Issue 9163] std.parallelism broken with extensive optimizations (gdc)

[d-bugmail at puremagic.com]

http://d.puremagic.com/issues/show_bug.cgi?id=9163


David Simcha <dsimcha at yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dsimcha at yahoo.com


--- Comment #14 from David Simcha <dsimcha at yahoo.com> 2013-01-06 05:56:19 PST ---
Sorry, I've been incredibly busy lately and just saw this post now.  IIUC the
**whole point** of atomic operations in core.atomic is that they're supposed to
act as fences.  No code motion should take place across ASM statements at the
compiler level, and none should take place across lock; instructions at the
lower levels.  Obviously either my understanding of the spec is wrong or GDC
isn't optimizing them properly if that's the case.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------