Re: C/C++ programmer giving Ada95 a chance -- writing an emulator.

["Jean-Pierre Rosen" <rosen.adalog@wanadoo.fr>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1183 bytes --]


jross <rem.jross@rem.codemecca.com> a �crit dans le message :
38e7e951.8384503@news.shreve.net...
>
> I was making the assumption that a case statement would compile to
> evaluate every case until finding a match -- hence, not efficient.
> Several years ago, I am sure this would have been the case (no pun
> intended).  However, now that I think about it, a compiler these days
> should be intelligent enough to create an indexed jump table where all
> (or most all) consecutive values are provided and create efficient
> execution code.
>
As a point of information, when I was working on Ada/Ed (some years ago
now), I had in my office a PhD thesis on "efficient code generation
techniques for case statements". There was literally a dozen of techniques
possible, depending on how the values were split over the whole range...
Moral: NEVER make assumptions about the way code is generated, unless you
REALLY know what's happening in the compiler (which in practice means, you
are part of the compiler team!)
--
---------------------------------------------------------
           J-P. Rosen (Rosen.Adalog@wanadoo.fr)
Visit Adalog's web site at http://pro.wanadoo.fr/adalog