ASH HAFEZ wrote:
>
> firstly, i'm impressed by the quick responses ...
>
> what i intended to test for was an appropriate program error. I didn't
> expect a non-terminating one though. I don't believe this same
> output occurs in the 1.4 implementation -- although it does seem to halt
> for a while before printing a 2 line program error.
>
> ash
>
> btw, where abouts in the world am i sending these emails?
Hi,
Sorry about the misunderstanding.
I'll look into sorting the error message. The problem is
that in Hugs the fatal failures in matches are not caught
apart from top level, and the top level printing mechanism
(as you observe) can loop.
However, this simplistic aborting mechanism is one reason
Hugs is both compact and can generated code quickly.
I previously looked at adding code to give a reasonable error
message when pattern matching fails, but in does
make the code quite a bit bigger.
Marks suggestion, having a cutoff for the message,
seems a reasonable interim solution.
Where are we from? Have a look at
http:/www.cse.ogi.edu/PacSoft/
Currently, PacSoft looks after Hugs.
Cheers,
Andy
