Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.programming > #2218

Re: true loop

Show all headers | View raw

Patricia Shanahan wrote:
> for(;;) is horribly indirect. Yes, I know that an empty continuation
> condition means continue forever, but what is wrong with saying what one
> means?

I quite strongly disagree.  I have always found the "while (true)" idiom to be 
unnecessarily indirect -- it /doesn't/ (IMO) "say what it means".   What I'm 
trying to say is
    loop forever
but what "while (true)" is saying is
    loop while a condition (which happens always to be true) holds"

It may help illustrate my unease with the idiom if you consider:
    while (!false) {...
or even (and I've heard of this one in production code though I haven't seen it 
personally):
   while (THE_POPE_IS_A_CATHOLIC) {...
(which obviously depends on the appropriate definition elsewhere, and is bad 
for that reason /too/).

I might be happier if the idiom was

    while () {...}

but that would depend on a special-case one-off rule for "while".  The 
corresponding idiom using "for" also depends on a  rule, but (unlike the case 
for "while") that rule is not /special/ -- it already exists for each of the 
sub-clauses of the "for" statement anyway.  (In C-family languages).

Going back to the OP("bob")'s original question.  One thing that I haven't seen 
anyone mention is that there are two ways that an infinite loop is used.  One 
is where the loop is actually expected to run "forever" -- as in the heart of a 
scheduler -- or at least until some external processing somewhere breaks the 
loop "invisibly", say by calling exit() -- for instance a message pump loop 
might take that form.  That obviously can't be considered bad practise, but 
isn't all that common either.

But the other way it's used is as the framework for a loop which has no overall 
controlling condition, but which is terminated by conditional processing within 
the loop.  Personally I use that form rather often, since often there isn't a 
single clear "master" condition.  I can imagine that there may be people (the 
ones who object to breaking out of loops and who probably also object to 
multiple "return" statements) who would object to the:
    while (true) { ... processing with breaks/returns... }
idiom (whether expressed with "while" or "for").  If such people do exist then 
they are mistaken (I make this absolute statement with no apologies 
whatsoever).  So my answer to [what might be] the OP's question is that it 
isn't considered bad practice (by anyone who isn't wrong) ;-)

    -- chris 

Back to comp.programming | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

true loop bob <bob@coolfone.comze.com> - 2012-09-11 10:52 -0700
  Re: true loop Daniel Pitts <newsgroup.nospam@virtualinfinity.net> - 2012-09-11 11:17 -0700
  Re: true loop pete <pfiland@mindspring.com> - 2012-09-11 19:12 -0400
    Re: true loop Patricia Shanahan <pats@acm.org> - 2012-09-11 16:28 -0700
      Re: true loop pete <pfiland@mindspring.com> - 2012-09-11 22:23 -0400
      Re: true loop Geoff <geoff@invalid.invalid> - 2012-09-11 19:33 -0700
        Re: true loop Patricia Shanahan <pats@acm.org> - 2012-09-11 19:53 -0700
        Re: true loop Nick Keighley <nick_keighley_nospam@hotmail.com> - 2012-09-12 02:23 -0700
      Re: true loop "Chris Uppal" <chris.uppal@metagnostic.REMOVE-THIS.org> - 2012-09-15 10:56 +0100
    Re: true loop Nick Keighley <nick_keighley_nospam@hotmail.com> - 2012-09-12 02:21 -0700
      Re: true loop pete <pfiland@mindspring.com> - 2012-09-12 18:04 -0400
        Re: true loop Ian Collins <ian-news@hotmail.com> - 2012-09-13 10:22 +1200
          Re: true loop Geoff <geoff@invalid.invalid> - 2012-09-12 19:06 -0700
  Re: true loop Mok-Kong Shen <mok-kong.shen@t-online.de> - 2012-09-13 16:27 +0200
    Re: true loop Nick Keighley <nick_keighley_nospam@hotmail.com> - 2012-09-19 07:04 -0700
      Re: true loop Patricia Shanahan <pats@acm.org> - 2012-09-19 07:30 -0700

csiph-web