Hi Alex,

At 2026-02-21T20:56:05+0100, Alejandro Colomar wrote:
> On 2026-02-21T11:41:42-0600, G. Branden Robinson wrote:
> > At 2026-02-21T16:02:52+0100, Alejandro Colomar wrote:
> > > +.SH RETURN VALUE
> > > +.IR s+strlen(s) .
> > 
> > Too cute, in my opinion.  Use English.  :)
> 
> The thing is, at first I thought, am I going to repeat the same exact
> words as in the DESCRIPTION?
> 
> DESCRIPTION
>      strnul() returns a pointer to the terminating null byte in the
>      string s.
> 
> RETURN VALUE
>      strnul() returns a pointer to the terminating null byte in the
>      string s.
> 
> I could remove the DESCRIPTION altogether...  What would you do?

Fair.  In this case, I'd probably just say...

See \[lq]DESCRIPTION\[rq] above.

I don't have much sympathy for hackers who complain that the "RETURN
VALUE" section didn't tell them what they needed to know when the
one-sentence "DESCRIPTION" _that they didn't read_ would have.

At a certain point, documentation writers have no defense from people
who refuse to read.

> I might as well write it as &s[strlen(s)] if pointer arithmetic is the
> confusing part.  :)

Hah!

> > I assume that the string library reforms you're pursuing are
> > intended in part to be adopted by newcomers to C.
> 
> I intend old programmers to use it too.

Goooooood luck.  As we've seen, to take gets(3) away from some people
has required killing it off from the standard C library itself.

Some old programmers get really good.  Some just get old.

> I guess you're expecting a patch to groff once this is in a branch of
> gnulib you're using.  ;)

No, not expecting.  :)

Replacing our bespoke `strsave()` with `strdup()` seems like a higher
priority to me.

https://savannah.gnu.org/bugs/?66518

Regards,
Branden

Attachment: signature.asc
Description: PGP signature

Reply via email to