Allison Randal wrote: > Moritz Lenz wrote: >> In the pdd27mmd branch rakudo segfaults on >> t/01-sanity/07-simple-multisubs.t. >> >> The backtrace shows at least 2k nested calls to Parrot_MultiSub_get_iter >> (I didn't print out the full trace...), so I guess it's recursing >> infinitely, and then runs out of stack space. > > Curiously, that test was passing a week ago, and the only thing that's > really changed since then is merging trunk into the branch. This may be > a merge problem. Looking into it.
Ah, got it. Some new Rakudo code from trunk must have been calling MultiSub's 'get_iter' vtable function. It had been modified in the branch to do a manhattan sort before creating an iterator, then call 'VTABLE_get_iter' to create the iterator (creating an infinite recursion). Changed that to a call to SUPER, so it correctly calls the parent (ResizablePMCArray) 'get_iter', instead of recalling itself. Test now passes. Allison _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
