OpenEuphoria

RE: Euphoria's Future?

• Posted by Henri.Goffin at sbs.be Nov 19, 2001
• 467 views

I'd rather think I'm trapped in a space-time loop that forces me to live the
same events again and again 

More seriously, Rob, your suggestion (use trace to locate the offending
statement) is of no use here since my prog crashes before executing the first
statement. Probably during the init phase (heap allocation or something).

My poor debugging skills on MS platforms have just allowed me to make clear that
the program tries to write to an unallocated memory area. The instruction is: 
mov     [edi+0x4],ebx and edi = 0x007d0ffc. This pattern reminds me of numerous
such errors I saw in the past where some loop executes one time too much and
makes a pointer point just beyond the boundary of an allocated memory block. But
it's just a feeling. After all maybe neither my program nor Euphoria interpreter
are wrong but instead the C-compiler (Watcom, right?) is the culprit.

Last comment. The doc says a bound program is bound with the PD flavor of
exw.exe. So I tried 'pdexw myprog.exw' thinking I would reproduce the problem.
But it also runs OK just like with the registered Euphoria interpreter. So it
seems to be linked with what bind.ex produces.

Regards,

Henri Goffin

> -----Original Message-----
> From:	rforno at tutopia.com [SMTP:rforno at tutopia.com]
> Sent:	Friday, November 16, 2001 1:24 AM
> To:	EUforum
> Subject:	Re: Euphoria's Future?
> 
> 
> Were the planets properly aligned?
> ----- Original Message -----
> From: "Robert Craig" <rds at RapidEuphoria.com>
> To: "EUforum" <EUforum at topica.com>
> Sent: Thursday, November 15, 2001 2:20 PM
> Subject: Re: Euphoria's Future?
> 
> 
> > Henri Goffin writes:
> > > This is simply not true, at least on NT/W2K platforms.
> > > The reason is the bound executable starts in a new console
> > > window that is unaware of the invoking window's environment.
> >
> > With the Euphoria source, you'll be able to build your
> > own exw.exe that is more console-oriented. It should work
> > with STDIN/STDOUT on any system. You just have to edit
> > one line in the .bat file that I'll distribute with it. Maybe I'll make a
> version
> > like that myself and upload it once 2.3 alpha is out.
> >
> > > There is also the problem of a memory violation error
> > > that I have with a bound Euphoria program. The answer of
> > > Robert to my request for help was rather disappointing because
> > > too reminescent of Microsoft's favorite silver bullet:
> > > "reboot your system, maybe the problem will have disappeared".
> > > Who would dare build his/her business on such a basis?
> >
> > On Oct 5, 2001 I replied to you...
> >
> > R> In 2.3 the interpreter will have trace(3) which will allow you
> > R> to log the exact Euphoria statement that was being executed
> > R> when a machine-level crash like this occurred.
> >
> > R> Until then, you can insert print statements to pin down
> > R> where the crash happens.
> >
> > Did you insert any print statements?
> > If not, wait 10 days or so until trace(3) is available.
> > You'll also be able to try the new 2-pass binder,
> > although I doubt that a binding bug is the
> > root cause of your problems.
> > >From my experience these cases usually turn out
> > to be a bug in the application that only shows up
> > when it's bound (or it's Tuesday afternoon, or there's
> > unusual sunspot activity, or ...)
> >
> > Regards,
> >    Rob Craig
> >    Rapid Deployment Software
> >    http://www.RapidEuphoria.com
> >
> >
> 
>