[9fans] /sys/src/cmd/chmod.c

[9fans] /sys/src/cmd/chmod.c

[Lucio De Re]

This struck me, looking at recent changes to chmod(1) which allow one
to flag a file as "not to be included in nightly archives":

	for(; *s ; s++){
		switch(*s){
		case 'r':
			mode |= A(DMREAD); break;
		case 'w':
			mode |= A(DMWRITE); break;
		case 'x':
			mode |= A(DMEXEC); break;
		case 'a':
			mode |= DMAPPEND; break;
		case 'l':
			mode |= DMEXCL; break;
		case 't':
			mode |= DMTMP; break;
		default:
			return 0;
		}
	}
	if(*s != 0)
		return 0;

I can't see how *s can be non-zero on exit from the for loop.  Is
my C programming _so_ rusty?!

++L

Re: [9fans] /sys/src/cmd/chmod.c

[Martin Harriss]

Defensive programming?

Martin


Lucio De Re wrote:
> This struck me, looking at recent changes to chmod(1) which allow one
> to flag a file as "not to be included in nightly archives":
>
> 	for(; *s ; s++){
> 		switch(*s){
> 		case 'r':
> 			mode |= A(DMREAD); break;
> 		case 'w':
> 			mode |= A(DMWRITE); break;
> 		case 'x':
> 			mode |= A(DMEXEC); break;
> 		case 'a':
> 			mode |= DMAPPEND; break;
> 		case 'l':
> 			mode |= DMEXCL; break;
> 		case 't':
> 			mode |= DMTMP; break;
> 		default:
> 			return 0;
> 		}
> 	}
> 	if(*s != 0)
> 		return 0;
>
> I can't see how *s can be non-zero on exit from the for loop.  Is
> my C programming _so_ rusty?!
>
> ++L
>

Re: [9fans] /sys/src/cmd/chmod.c

[Zhen Lin]

lucio@proxima.alt.za (Lucio De Re) wrote in message news:<20031231163240.Z25949@cackle.proxima.alt.za>...
> I can't see how *s can be non-zero on exit from the for loop.  Is
> my C programming _so_ rusty?!
> 
> ++L

Evil, non-synchronised write into *s? Magic cosmic rays? Anything
could happen to anything.

Re: [9fans] /sys/src/cmd/chmod.c

[ron minnich]

On Mon, 5 Jan 2004, Zhen Lin wrote:

> Evil, non-synchronised write into *s? Magic cosmic rays? Anything
> could happen to anything.

like the code above changing in some way and the "impossible condition"
gets reached. Gosh, that never happens to *my* code :-)

ron