[MLton-devel] cvs commit: -exn-history reimplemted

[Matthew Fluet]

> > No, I meant leave the implementation of MLton.Exn.history the same:
> 
> Duh.  Thanks for explaining what should have been obvious.  :-)
> 
> But the the benchmarks still show that compiling -exn-history true can
> hurt even when MLton.Exn.history isn't used.

No, MLton.Exn.history is always used in the top-level handler.  Even when
the user program doesn't use it explicitly.  What's missing is sufficient
analysis that proves that the top-level handler isn't reachable, hence the
history isn't needed.  (Yes, it's the halting problem (or worse, because
some of the benchmarks read in external data).)

I agree that Mton.Exn.history should be like Assert.assert -- it should be
a "nop" when -exn-history false is used.  I think it's arguable whether or
not the top-level handler should print the exn history.  Certainly, it
would be painful to wrap ever top-level evaluation with a handler if the
user wants that information.  On the other hand, it seems like overkill to
add another option that determines this behaviour.  Particularly since the
reason someone compiles with -exn-history true is likely because they are
trying to track down some uncaught exception. ;)



-------------------------------------------------------
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone?  Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel