http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49815
--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-07-22 14:14:07 UTC --- Ah, I see. adjust_insn call isn't guarded by MAY_HAVE_DEBUG_STMTS, therefore if -fvar-tracking-assignments gives up due to insanely large hash tables, it won't be reset during vt_finalize. BTW, wouldn't it be better to model the insn completely rather than partially? In particular, in addition to the arguments being copied have also the copy/clobber for all the other arguments, except for the sp = sp - something which would be another set.
