Okay, I took a quick stab at this today, removing that section where the variables were initialized to 888. On my system several new test failures appear without this (which indicates to me that somebody is relying on these lousy initialization values. Here is the list of tests that my system says are failing now:
t/pmc/class.t:22-23 t/pmc/exceptionhandler.t:7 t/pmc/hash.t:99 t/pmc/managedstruct:8 t/compilers/pge/06-grammar.t:1-16 I'm going to take a look later to see if the tests are wrong or if something internally is relying on this weird behavior. --Andrew Whitworth _______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
