Re: [Nmh-workers] proposed patch for shell metacharacter failure in nmh-1.7

[Paul Vixie]

Ken Hornstein wrote:
> > > That's not right, it should be:
> > >
> > >       while ((pp = strchr (pp, '\''))&&   buflen>   3) {
> > pointers aren't booleans. in BSD style as used in BIND, this would be:
> >
> > while ((pp = strchr(pp, '\'')) != NULL&&  buflen>  3) {
>
> Forgive my stupidity ... but this is just a matter of style, correct?
> AFAIK, there is no problem from a language perspective.  If there is a
> problem from the language perspective, I would like to understand it.

the same underlying machine code will be generated.

however, it is both clearer, and also, more likely to generate a 
compiler error at least a warning, if our assumptions are wrong.

> (I am neutral on the style issue, myself).

i came to C from PDP-11 assembler, and so anything that sets the 
condition codes could be used for a subsequent branch-if instruction, 
and so i used to write "if (thing1) thing2" because i "knew" that no 
test instruction was needed. later we got -Wall -Werror; i changed.

i would rather give the compiler an extra chance to yell at me. pointers 
aren't booleans, in spite of C's ability to treat them that way. i had 
this fight with keith bostic when we integrated Cron into 4.4, and he 
won, and he was right.

--
P Vixie