OpenEuphoria

Re: Euphoria 2.5

• Posted by Robert Craig <rds at RapidEuphoria.com> Sep 07, 2004
• 608 views

William Heimbigner wrote:
> > The front-end (scanning/parsing) is now written in Euphoria.
> > It's translated by EtoC to C, compiled, and linked with the
> > hand-coded back-end. The back-end is in carefully hand-coded C as before
> > for full speed. So theoretically the front-end (only) is a bit
> > slower. But you won't notice it unless you have a very old machine,
> > and a very large program, and you watch the start-up time carefully.
> >
> 
> If it means it risks slowing it down, why did you rewrite part of it? It
> doesn't quite make sense. What is the AdVanTage of having it be 30%
> translated euphoria?

* It's the same front-end as the open source Eu in Eu 
  interpreter uses (avoids dual maintenance).

* The old C-coded front-end had lots of very ugly, complicated
  code to do with allocating/reallocating numerous data structures.
  The new front eliminates all that by using append() etc.

* I can test changes to the new front-end in interpreted mode,
  (subscript checking, uninitialized variable checking etc.), 
  not to mention profiling, interactive tracing etc.

* I like programming in Euphoria a lot more 
  than I like programming in C.

* the speed difference is tiny and will get tinier as
  machines get faster
  
Regards,
   Rob Craig
   Rapid Deployment Software
   http://www.RapidEuphoria.com