On Sun Jul 06 22:04:15 2008, songmaster wrote: > Ok, I'm done with trying to understand and debug this any further. It > needs someone who knows how the Parrot interpreter stores return > arguments and how the tail-call stuff is implemented as well. > > At least I found that the segfault was only a side-effect of this > problem, see RT #56636. Once that fix is applied the segfault will no > longer occur, but the tail-call code output will not be correctly sorted > since the underlying problem still exists.
The attached pir code (updated from the original post) no longer segfaults at r38198, but it does produce an incorrectly ordered array: $ ./parrot broken_tailcall_sort.pir 4 5 3 2 5 1 1 2 3 4 5 5 2 4 5 3 1 5 This should be looked into and fixed once the pcc_rewiring branch has been merged back into trunk.
broken_tailcall_sort.pir
Description: Binary data
_______________________________________________ http://lists.parrot.org/mailman/listinfo/parrot-dev
