Re: wcasestrstr patch incorrect?

[Jürgen Kahrs]

Andrew J. Schorr wrote:

> The patched code results in this 'for' loop:
> 
>                       for (j = 0; j < needle_len; j++, start++) {
>                               wchar_t h, n;
> 
>                               h = towlower(*start);
>                               n = towlower(needle[j]);
>                               if (h != n)
>                                       continue;
>                       }
> 
> Coming at the end of the 'for' loop, doesn't the continue statement
> have no effect?  It seems to me like the previous 'goto out'
> code was correct...

Yes, you are right. The "continue" should be reverted
to the "goto". Arnold probably intended to do a
"continue" on the outer loop (i) (which would be
equivalent to a "goto"). But he forgot that the
"continue" always works on the inner loop (j).