🚀 go-pugleaf

In article <13025@smoke.BRL.MIL> gwyn@smoke.BRL.MIL (Doug Gwyn) writes:
>... empty subject.  There is no implementor dumb enough to attempt to make
>such a claim ...

This is why I said Blair's statements were vacuously true.


Here's the recap, for those who tuned in late:

* A lot of people don't really care whether you use for(;;) or while(1), and
  wish that such discussion, if it is to take place at all, should be in
  alt.religion.computers.

* Blair happens to prefer while(1) for some reason.  Most of us don't care
  why.

* Evidently someone once told Blair of a compiler that produced slightly
  better code for the `for' statement.  I believe I've seen such a compiler.

* The better-code argument is pretty weak, since any compiler that overlooks
  such an obvious optimization is doing such a poor job that you'd better not
  be using it for anything where you care about compiler efficiency.  At best,
  you can use this argument as a tie-breaker.  (E.g. this is why I originally
  settled on using `++i' rather than `i++'.)

* Blair noted that the ANS, in describing the `for' statement, uses wording
  that (correctly) describes `for(;;)' as equivalent to `while(1)'.  He
  mistakenly thought that this wording was more than just a convenient way to
  describe the semantics, and used this to rebut the better-code argument.

* It turns out the K&R 1 actually does contain a similar clause, not that it
  matters.

* No sane vendor, in the process of upgrading a Classic C compiler to ANSI C,
  is going to change the way code is generated for `for' or `while' on the
  basis of this wording.  Thus, it remains true that `for(;;)' is slightly
  less likely to generate a useless compare.

* A lot of people don't really care whether you use for(;;) or while(1), and
  wish that such discussion, if it is to take place at all, should be in
  alt.religion.computers.

Karl W. Z. Heuer (karl@ima.ima.isc.com or harvard!ima!karl), The Walking Lint
(Isn't this subject in your KILL file yet?)