Re: goto (was browser output)

"Michael Carman" <mjcarman@mchsi.com> wrote in message
news:clousn$9421@onews.rockwellcollins.com...
> Laura wrote: 
>
> There *are* occasions when goto is necessary, just like there are
> occasions when using symbolic references is necessary. Both situations
> are very rare, but it does happen.
>
> That said, Perl has three flavors of goto, which are not all the same:
>
>    goto LABEL
>    goto EXPR
>    goto &NAME
>
> The first two are the traditional "jump to someplace else" that's
> generally considered bad. goto &NAME is very different: it replaces the
> current subroutine with the one referenced. While the first two forms
> are shunned, the third is quite common (e.g. in an AUTOLOAD).

I notice that File::Compare and File::Copy both make extensive use of goto
LABEL.

Anyone care to comment on whether it's really needed in these modules?

Tintin <tintin@invalid.invalid> wrote in comp.lang.perl.misc:
>
> "Michael Carman" <mjcarman@mchsi.com> wrote in message
> news:clousn$9421@onews.rockwellcollins.com... 
>
> I notice that File::Compare and File::Copy both make extensive use of goto
> LABEL.
>
> Anyone care to comment on whether it's really needed in these modules?

The question is not if it's needed, but if it is used reasonably.

Goto is evil when used as the primary means of flow control.  At the
time Dijkstra mounted his campaign against "goto", there were
languages (most notably Fortran), that *had* no other large-scale
flow control ("if" only worked on one statement).  That used to make
a mess of programs.  The campaign wasn't so much against the use of
"goto" by individual programmers (they often had no choice), but
to foster the design of "structured" languages that didn't force
"goto" upon their users.

These days there are no more mainstream languages that aren't
"structured", so nobody talks about "structured programming" anymore,
but everyone does it.  In that framework, the occasional "goto" is
no big deal, and it can be the operation of choice.

Anno

Tintin (tintin@invalid.invalid) wrote on MMMMLXXVI September MCMXCIII in
<URL:news:2ubo3fF27cah6U1@uni-berlin.de>:
^^
^^  I notice that File::Compare and File::Copy both make extensive use of go
to
^^  LABEL.
^^
^^  Anyone care to comment on whether it's really needed in these modules?


It's not really needed. But I don't understand the question.
I've noticed that File::Copy makes extensive use of 'while' and
lexical variables. Anyone care to comment on whether it's really
needed in these modules?

Almost any construct is not "really needed", because there's always
another way to do the same thing.


Abigail
--
:;$:=~s:
-:;another Perl Hacker
:;chop
$:;$:=~y:;::d;print+Just.$:

Anno Siegel (anno4000@lublin.zrz.tu-berlin.de) wrote on MMMMLXXVI
September MCMXCIII in <URL:news:clqm9i$m3q$1@mamenchi.zrz.TU-Berlin.DE>:
//
//  The question is not if it's needed, but if it is used reasonably.
//
//  Goto is evil when used as the primary means of flow control.  At the
//  time Dijkstra mounted his campaign against "goto", there were
//  languages (most notably Fortran), that *had* no other large-scale
//  flow control ("if" only worked on one statement).  That used to make
//  a mess of programs.  The campaign wasn't so much against the use of
//  "goto" by individual programmers (they often had no choice), but
//  to foster the design of "structured" languages that didn't force
//  "goto" upon their users.
//
//  These days there are no more mainstream languages that aren't
//  "structured", so nobody talks about "structured programming" anymore,
//  but everyone does it.  In that framework, the occasional "goto" is
//  no big deal, and it can be the operation of choice.


Let me also point to Donald Knuths "Structured Programming with goto
Statements", where he shows that using goto and structured programming
aren't multiple exclusive.



Abigail
--
A perl rose:  perl -e '@}-`-,-`-%-'