[sane-devel] [sane-commit] CVS update of sane-backends/frontend (saned.c)

[m. allan noah]

On Thu, Mar 12, 2009 at 6:42 PM, Ilia Sotnikov <hostcc at gmail.com> wrote:
> On Thu, Mar 12, 2009 at 11:28 PM, Julien BLACHE <jb at jblache.org> wrote:
>> Reviewing the code, did you actually try the gai+IPv6 code?
>
> Actually, yes. I'm always testing all possible scenarios.
>
>> Going in a for() loop iterating over res after getaddrinfo() returned
>> an error doesn't look safe, as I don't remember any guarantee about
>> the value of res in that case and the doc doesn't mention anything
>> about that.
>
> Agree, could be libc implementation specific.
>
>> In the IPv4 code, you'll end up printing an unrelated error message
>> "... (only IPv4 is supported)", because the if() you've modified
>> actually only checks that the returned address is IPv4.
>
> Cosmetic.

Error messages are the only thing we have to debug random user issues.
They should be as accurate as possible.

>> I'm sorry, but I can't call that a good patch, and am reiterating my
>> request for a revert.
>
> You are not the only SANE developer, thus I'm waiting for the consensus.

Julien is the only devel spending any time on the networking code for
the past couple of years. The closest you will come to consensus here
is to convince him that the concept is sound. After that, you must
still produce a patch to primarily his liking.

> P.S. Just wondering how such a simple change could lead to such
> lengthy discussion.

You made half the posts :)

allan
-- 
"The truth is an offense, but not a sin"