Re: [off topic?] Re: [PATCH] strtok --> strsep in framebuffer drivers

[Kari Hurtta]

> > René Scharfe wrote:
> 
> > NAME
> >        strsep - extract token from string
> > [...]
> > RETURN VALUE
> >        The strsep() function returns a pointer to the  token,  or
> >        NULL if delim is not found in stringp.
> > 
> > If strsep returns NULL, and you dereference it -> Oops.
> > 
> > 
> > ! if (!this_opt)
> > 	continue;
> 
> Is that manual page incorrect?
> 
> I would think that 
> 
> 	strsep() returns NULL  when *stringp points to '\0' character
> 	
> 	and that if delim is not found in stringp then stringp
> 	is just advanced to '\0' character of string (and original
> 	*stringp value is returned)
> 
> If that is not true, then these all patches are incorrect.
> Namely last token will be skipped.

Seems that I have also good to talk to mayself....

lib/string.c tells truth.

strsep() returns NULL when *stringp is NULL.

if delim is not found *stringp is set to NULL (and original
*stringp value is returned).

Because kernel code claims that 

 * It returns empty tokens, too, behaving exactly like the libc function
 * of that name. In fact, it was stolen from glibc2 and de-fancy-fied.
 * Same semantics, slimmer shape. ;)

quoted manual page IS incorrect (at least if it describes glibc2 function.)

-- 
          /"\                           |  Kari 
          \ /     ASCII Ribbon Campaign |    Hurtta
           X      Against HTML Mail     |
          / \                           |