On Monday 03 March 2008 19:06:01 chromatic wrote: > ... which indicates that whatever the problem is, there's something getting > freed to the PMC_EXT pool inappropriately.
Turns out this was a problem with the clone op not being paranoid enough to unshare data between the temporary clone and its ultimate destination. To wit: it's fine to copy PMC_EXT pointers from one to the other, but as the temporary clone will get recycled right away, it can't keep that pointer lest the GC free the PMC_EXT structure, which the clone that people care about expects to last as long as it does. See r26197. -- c
