Re: Secret new routine_id() feature. (Attn: Rob)
CoJaBo wrote:
>
> Patrick Barnes wrote:
> >
> > On Wed, 08 Dec 2004 19:00:54 -0800, Robert Craig
> > <guest at rapideuphoria.com> wrote:
> > > Thanks for reporting this bug.
> > > It will be fixed in the beta release.
> > > It looks like only the interpreters (PD and RDS) have the bug.
> > > Translated code, bound code, and shrouded code seem to work correctly.
> > > Even in 2.4 and earlier, many programs were parsed completely,
> > > or almost completely, before routine_id() was executed.
> > > I think it's better from a language design point of view that
> > > routine_id() looks only at earlier routines, not later.
> > > My reasons are:
> > > 1. I like to torture people.
> > > 2. Check the EUforum archive for the other reasons.
> >
> > Hah!
> > People can and will write hard-to-read code Rob, there's nothing you
> The code I've had to write to work around this problem
> is far worse then routines out of order!
> At least allow routine_id() to fing routines defined
> later!!!!! PLEASE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
>
>
> > or anyone can do to stop them. The 'elegance' you think comes from
> > arranging functions in order is nothing compared to a program that
> > doesn't need to resort to hacks in order to function, or even a
> > program that has its functions arranged in LOGICAL order, rather than
> > chronological.
> >
> > Restricting forward-reference is seen as a limitation. It's ALWAYS
> > seen as a limitation. If a user runs into the 'no forward-reference'
> > problem, they don't go "gee, RDS is looking out for me best
> > interests", do they? NO! They say "Damn, why doesn't this stupid
> > language let me define my functions where I like?! What is it, C circa
> > 1982?!"
> >
> > Stop pretending you're the arbiter of good coding style, and you'll
> > improve the Euphoria language by it. Those who have a need to get
> > around the forward-reference problem you've *imposed* on the language
> > will thank you.
> >
> > --
> > MrTrick
> >
> >
I guess I will agree with everyone else in that routine_id() should be
able to return a unique integer id number at any section of code, instead
of just after defining the functions/procedures. I see nothing wrong
with being able to take advantage of fully parsed code at startup in this
case. and as it seems.. keeping this feature 'bug' would be in almost
everyones interest, I prosume?
|
Not Categorized, Please Help
|
|
