On Jun 20, 8:24 am, Paul Pluzhnikov <[EMAIL PROTECTED]> wrote: > mathieu <[EMAIL PROTECTED]> writes: > > On Jun 18, 9:15 pm, Francis Litterio <[EMAIL PROTECTED]> wrote: > >> mathieu wrote: > ... > >> > std::map. Those accesses are concurrent (SMP machine). Is there > >> > something I need to do even for read only access ? > > ^^^^^^^^^^^^^^^^ > ... > > > #4 0x00002b78ba63bc3f in std::map<...>::insert () > > What made you believe that map::insert() is a read-only access?
I wrote my comment backward... I finally understood what went wrong when I hit to particular seg fault. the previous seg fault would appear much later during a read access to the std::map (I did not understood it had been trashed by incorrect write access). Sorry for the confusion -Mathieu Ps: I was trying to 'be smart' and delay the loading (=write) of the std::map only at the first user request but that was just stupid... _______________________________________________ help-gplusplus mailing list help-gplusplus@gnu.org http://lists.gnu.org/mailman/listinfo/help-gplusplus
