On May 12 15:30, Ryan Johnson wrote: > On 12/05/2011 2:42 PM, Corinna Vinschen wrote: > >On May 12 13:53, Ryan Johnson wrote: > >>On 12/05/2011 12:55 PM, Corinna Vinschen wrote: > >>>>> heap *heaps; > >>>>This is a misnomer now -- it's really a list of heap regions/blocks. > >>>I don't think so. The loop stores only the blocks which constitute > >>>the original VirtualAlloc'ed memory regions. They are not the real > >>>heap blocks returned by Heap32First/Heap32Next. These are filtered > >>>out by checking for flags& 2 (**). > >>Sorry, I cut too much context out of the diff. That's > >>heap_info::heaps, which indeed refers to heap regions which we > >>identified by checking flags&2 (that's why it needs the heap_id > >>inside it now, where it didn't before) (++) > >So you think something like heap_chunks is better? > Maybe. I actually only noticed because the code snippet you > originally sent also used 'heaps' as an identifier and the two > clashed when I pasted it in ;)
I've renamed it to heap_vm_chunks and checked my patch in. I'm going to follow up on the cygwin-developers list. Now that we have your excellent maps output, it's pretty easy to see where the potential problems for fork are. Corinna -- Corinna Vinschen Please, send mails regarding Cygwin to Cygwin Project Co-Leader cygwin AT cygwin DOT com Red Hat