On Friday 29 August 2008 21:47:41 Stephen Weeks wrote: > I tried using resizablepmcarray's append() in cardinal and I got 'double > fre or corruption' on program exit. > > I tried to fix it, but I'm not sure about GC issues. > > As far as I can tell, the attached shouldwork.patch (which adds > GC_WRITE_BARRIER before the write) should fix the problem, but it doesn't > change anything. > > The attached doeswork.patch actually does fix the issue, but has some > obvious issues. > > As far as I can tell, shouldwork.patch is doing the same thing as > calling VTABLE_set_pmc_keyed_int does, so I'm not sure why it doesn't > work. > > I couldn't reproduce it with ResizablePMCArrays directly, though. I'll > keep working to see if I can generate a small test case. > > Anyone have any ideas?
Do you have a small PIR program which demonstrates the problem? Does r31933 help? -- c