[ILUG] Re: [linux-edu] war games?

[Fergal Daly]

Is most of this thread happening in private or am I just missing some of 
the responses?

Anyway, back to the (off) topic. GOTO isn't always a bad thing. Very handy 
when you're allocating a bunch of resources sequentially and you want to 
stop as soon as any allocation fails. This can of course be done with 
nested ifs but if you're allocating 10 different things, that's 10 layers 
of nesting.

They're also good for some complicated if/then/else stuff that would 
otherwise need repeated code.

Also break (or last and next for Perl) are just dressed up GOTOs and 
they'll only get you nicely out of 1 level of nesting of a loop.

They have their place. By the way, if you have an aversion to GOTOs don't 
go anywhere near procmail's source, it's mind boggling,

Fergal

At 16:25 21/06/00, Philip Reynolds wrote:
>John P . Looney's [valen at tuatha.org] 16 lines of dribble included:
>:>On Wed, Jun 21, 2000 at 04:01:46PM +0100, Noel Carroll mentioned:
>:>> I wasn't knocking it... its just that in my case people used it for
>:>> everything.   It was all anti-flow structure.   You still have to use GOTO
>:>> in basic and its good and useful for that.  I just wouldn't use it in 
>C/C++
>:>> at all with the exception of exception handling or setjmp/longjmp stuff.
>:>
>:> Look through the Linux source - therer are seventy+ gotos!
>
>70+, there's a hell of a lot more than that in the full Linux source code last
>time I looked (2.2.15 kernel I think it was). Lots and lots and lots of gotos.
>Generally programs can do without goto statements, function blocks, if
>statements, switch statements etc. in C, can nearly always more than make up
>for gotos and make the code more readable as well.