[m-rev.] for review: module to read trace counts

[Peter Wang]

On Tuesday 11 January 2005 03:10 pm, Zoltan Somogyi wrote:
> On 11-Jan-2005, Peter Wang <wangp at students.cs.mu.OZ.AU> wrote:
> Since the declarative debugger also needs to be able to read trace counts,
> the new module should be in the browser directory, and part of the mdbcomp
> package. Any parts of the compiler, e.g. the trace_port type in
> trace_params.m, that trace_counts.m depends on should also be made part of
> the mdbcomp package.

Ok.

> Using the lexer seems a bit heavyweight to me. The declarative debugger
> will need to read in possibly dozens or hundreds of trace summaries, and it
> needs to do so at interactive speeds. Can you tell us how long this code
> takes to read in some typical sized input files?

I don't know what size is typical, but for a summary I generated from the
compiler itself, it takes just under one second (real time) to read in on an 
unloaded swordfish.

  % wc -l .mercury_trace_counts
  26439 .mercury_trace_counts

> > +            default:
> > +                /* This assumes isalnum is the same as char__isalnum. */
> > +                if (isalnum(*c) ||
> > +                    strchr(" !@#$%^&*()-_+=`~{}[];:'\"<>.,/?\\|", *c))
>
> Where did that string come from? It looks OK, I am just curious.

It is the equivalent of term_io__is_mercury_punctuation_char.  I'll add a 
comment to that effect.

Peter

--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------