svn commit: r231817 - head/usr.sbin/IPXrouted

[Eitan Adler]

On Thu, Feb 16, 2012 at 2:43 AM, Hiroki Sato <hrs at freebsd.org> wrote:
> Eitan Adler <eadler at FreeBSD.org> wrote
>  The oldroute is still used in the #ifdef section at the end of this
>  function.

:( I will try to come up with a fix this weekend.

>  This file was from NetBSD and this change just increased the diff.
<insert rant about contrib code not living in contrib>

> I do not understand why the prototype declaration of MCReadCat() was
>  left while the main body was removed.

My fault, I should have removed both (if the commit was valid in the
first place).

>  IMHO it is not a good idea to remove unused code partially like this
>  or unused code in one from another project if you just want to
>  suppress a compiler warning.

I have a slightly larger goal in mind: a usable world built with
gcc46. How soon, if ever, we will see this goal realized is a
different question. It would be nice to not be restricted to one (or
two) compiler options.

>  It may suppress the warning, but
>  remaining bits in comments and/or conditional parts where the
>  compiler does not complain make developers confused.  I think it is
>  more harmful than the warning.

I need to be more vigilant about ensuring that I am not removing
useful code. This warning is either indicative of a bug or results in
useless code. There are some cases where fixing the warning results in
problems but I try to avoid them, but a few inevitably slip through.


-- 
Eitan Adler
Source & Ports committer
X11, Bugbusting teams