Running with -b and -p gives the same output as no flags, which if I read Brett right, means that there's something else wrong.
Regards.
On Sunday, September 21, 2003, at 03:44 PM, Brent Dax wrote:
--Will Coleda: # When run as "parrot -t tcl.pbc foo.tcl 2> /dev/null", it immediately # exits, printing nothing to standard out. (if I peek at the output to # stderr, I see it's executing something. =-)
Parrot uses the "basic" runloop for trace. There are a couple other
features--bounds checking and profiling--that cause this runloop change.
See if parrot -b and parrot -p give you the same results; if they do,
your bug is runloop-dependent.
Everyone: I'm thinking we should have a set of flags that govern *just* runloop selection--something like a JIT flag, a prederef flag, a cgoto flag, and a switch flag. These would be set to a (platform-dependent) default, and could be cleared and manipulated by the embedder through a special subroutine.
I also think the command-line interface needs an overhaul to reflect this.
Any thoughts on that?
Will "Coke" Coleda will at coleda dot com
