David Faure wrote:

> Ohhhh. Didn't know about that one. Maybe Thiago intended to use that 
> one instead of socklen_t (for the same reasons) ? The name actually looks
> better than ksize_t (which was introduced before socklen_t existed at all,
> so ksize_t was either int or size_t).

Actually, the problem is that both socklen_t and size_t exist, which
could, in theory, mean different things. So I created that ksocklen_t
just for the purpose of a socket length, because I couldn't rely on
socklen_t existing.

But, defining it to ksize_t would be fine as well, since ksize_t is
defined to be the third parameter of getsockname. On most recent
systems, that IS socklen_t.

> Defining ksocklen_t to ksize_t and using it everywhere _seems_ like
> the right thing to do, but I'm not 100% sure.
> Hmm, you're saying Thiago isn't available for comments ? :(

I wasn't available. My mail was down because of DNS problems. I have 
a workaround in place.

> Tell me if you want me to commit your patch, or if the above makes
> any sense to anyone :}

I didn't see the patch, but if it defines ksocklen_t to be ksize_t, 
that's fine by me. I just think that having ksocklen_t to be defined 
(rather than change all references to ksize_t) is better.

-- 
  Thiago Macieira - UFOT Registry number: 1001
  thiagom@mail.com talha@geocities.com
    ICQ UIN: 1967141  PGP: 0x8F2978D5 and 0xEA9037A5
      Registered Linux user #65028

158/392. We'll be searching for one individual among thousands of 
drones. But she's one of us...And I'm not about to let her go. -- Capt. 
Janeway, "Dark Frontier, Part II"

 
>> Visit http://master.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<