Re: Guile 1.5.4 beta ready for testing.

[Marius Vollmer]

David Pirotte <address@hidden> writes:

> Marius Vollmer wrote:
> > 
> > Keith Wright <address@hidden> writes:
> > 
> > > It's OK.  I wish I had known to do that when I had the problem.
> > > What about making that the default and let people who know
> > > every option and never make mistakes set --no-debug?
> > > ...or turn it off in their .guile file!
> > 
> > Hmm, well.  I'm tempted to say, yes, why not.  It will make Guile
> > slower by default (about twice as slow), but I find that I turn the
> > debugger on in my programs anyway, and then just leave it on.  The
> > slowness hasn't hurt me yet.
> 
> I totally desagree, debug should be activated strictly on request,
> not being the default

Ok, what we actually have is this: Guile starts up with debugging off,
and it is turned on when entering the read-eval-print-loop.  It is not
turned on when running a program via `-s', `-l', or `-c'.  `.guile' is
only loaded when Guile is about to enter the repl, but debugging is
off when it is loaded.

Thus, a simple change would be to turn debugging on prior to loading
`.guile'.  This will leave `-s', `-l' and `-c' alone, and only turn on
debugging when it would have been turned on anyway, only earlier.

Ok?