------- Additional Comments From gniccolai at yahoo dot com 2005-02-22 10:27 ------- > > You say a hypothetical multithreaded C++ should state this as the semantics > for > static locals, and I don't disagree with you, provided we are not talking > about > one based on POSIX threads. POSIX specifically made the design decision to > always require explicit locks and to always require the programmer to find, > document, and lock cases where concurrent accesses might occur.
I think I wrote a rationale that explains and justifies this choice at http://www.niccolai.ws/works/articoli/art-multithreading-en-1a.html I had not been able to find an official rationale from open group; the nearest thing seems to be: http://www.unix.org/whitepapers/reentrant.html If more precise and authoritative ones are available, please point them out. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20099
