OpenEuphoria

Re: [EuForum] Definition of insanity...

• Posted by Patrick Barnes <mrtrick at gmail.com> Nov 23, 2004
• 521 views

On Tue, 23 Nov 2004 03:16:02 -0800, Tone =C5=A0koda <guest at rapideuphoria.co=
m> wrote:
> posted by: Tone =C5=A0koda <tskoda at email.si>
>
> Patrick Barnes wrote:
> >
> > Doing the same thing twice and getting different results...
>
> For this kind of bugs a "history debugger" would be nice: a debugger whic=
h records all statements executed and info what values variables had. Then =
when program crashes or something works wrong you can stop program executio=
n and debug for back, ie when debugging you can jump back to any statement =
in code which was executed, not only (one statement) forward, like in norma=
l debuggers. You could also search for a condition, ie var=val.
>
> The problem with such debugger would be that it program execution work qu=
ite slow, both because it would use execute.e and would have to record all =
that info to hard drive, and it would produce gigabytes large debug files.
>
> I'm not sure if it would be really that useful, considering it is needed =
quite a lot of work to create such debugger. But it could be it would be VE=
RY useful.

It would... but I've found the reason for the uncertainty...

There is a bug in IPC.ew that is not queueing data but overwriting it
if a procedure is remotely called multiple times before it's
processed.
Fixing it requires fixing IPC.... a prospect I *don't* relish.

--
MrTrick
----------