Subject: Re: resolver problems in -current?
To: Soren Jacobsen <firstname.lastname@example.org>
From: Sean Davis <email@example.com>
Date: 10/09/2003 20:00:50
On Thu, Oct 09, 2003 at 12:52:37PM -0700, Soren Jacobsen wrote:
> On 10/09 18:42, Toru TAKAMIZU wrote:
> > On Wed, Oct 08, 2003 at 11:39:15PM -0400,
> > Sean Davis <firstname.lastname@example.org> wrote:
> > > Has anybody else noticed resolver problems in -current? I just updated
> > > today, and once the new userland was installed, things like nslookup worked,
> > > but the plain libc resolver routines stopped working. the error from
> > > gethostbyname corresponted to NO_RECOVERY. Backing up to a userland from
> > > september 19th, with a kernel from today, brought everything back to the
> > > working state they were in before. I haven't done much digging, so does
> > > anybody know the root cause of this? Or better yet, how to fix it?
> > Take a look at a thread of current-users titled
> > "recent 1-6 branch libc and mozilla (etc.)", though I don't know
> > yours is the same problem.
> I was seeing the same behavior as Sean, and the problem is the same as
> that mentioned in the 1-6 thread. All is fine when I use my ISP's
> nameserver instead of a local dnscache. Sean was using dnscache too, I
Yes, I'm using dnscache. Switching to an ISP's nameserver or switching to
BIND are non-options. So what is broken now that is causing the resolver to
lose when using dnscache? I realize that the majority of NetBSD users I've
ever spoken to hate DJBware in any form, but some don't, and making the
system incompatible with it is a very bad thing to do. I assume it's
unintentional, but it needs fixing. I don't want to run on a september 19th
userland forever (and this also means I cannot update my netbsd-1-6 machine,
since the breakage has evidently made it onto that branch as well.)
/~\ The ASCII
\ / Ribbon Campaign Sean Davis
X Against HTML aka dive
/ \ Email!