[PATCH] URI RR Support for resolv_wrapper

[Alexander Bokovoy]

On Wed, 17 Aug 2016, Nathaniel McCallum wrote:
> On Wed, 2016-08-17 at 11:37 +0300, Alexander Bokovoy wrote:
> > On Tue, 16 Aug 2016, Nathaniel McCallum wrote:
> > > 
> > > One thing to note is that we have to #define ns_t_uri since glibc
> > > doesn't yet provide it. However, I made a patch for that too:
> > > 
> > > https://sourceware.org/ml/libc-alpha/2016-08/msg00465.html
> > > 
> > > Specifically, our desire for this patch is to test the upstream URI
> > > record usage that will be hopefully going into MIT Kerberos:
> > > 
> > > https://github.com/krb5/krb5/pull/481
> > > 
> > 
> > > 
> > > From 2903c724896828ef29a6f2e4871efa44f6c36807 Mon Sep 17 00:00:00
> > > 2001
> > > From: Nathaniel McCallum <npmccallum at redhat.com>
> > > Date: Tue, 16 Aug 2016 17:38:32 -0400
> > > Subject: [PATCH] Add support for the URI DNS Resource Record type
> > > 
> > > For more information, see RFC 7553.
> > > 
> > > Signed-off-by: Nathaniel McCallum <npmccallum at redhat.com>
> > > ---
> > >  doc/resolv_wrapper.1     |   1 +
> > >  doc/resolv_wrapper.1.txt |   1 +
> > >  src/resolv_wrapper.c     |  95
> > > +++++++++++++++++++++++++++++++++++++++++
> > >  tests/fake_hosts.in      |   2 +
> > >  tests/test_dns_fake.c    | 107
> > > +++++++++++++++++++++++++++++++++++++++++++++++
> > >  5 files changed, 206 insertions(+)
> > > 
> > > diff --git a/doc/resolv_wrapper.1 b/doc/resolv_wrapper.1
> > > index 5219713..fccdc5a 100644
> > > --- a/doc/resolv_wrapper.1
> > > +++ b/doc/resolv_wrapper.1
> > > @@ -92,6 +92,7 @@ A       dc\&.cwrap\&.org 127\&.0\&.0\&.10
> > >  AAAA    dc\&.cwrap\&.org fd00::5357:5f0a
> > >  CNAME   kerberos\&.cwrap\&.org dc\&.cwrap\&.org
> > >  SRV     _kerberos\&._tcp\&.cwrap\&.org kerberos\&.cwrap\&.org 88
> > > +URI     _vpn\&.cwrap\&.org https://vpn\&.cwrap\&.org/VPN
> > >  .fi
> > >  .if n \{\
> > >  .RE
> > > diff --git a/doc/resolv_wrapper.1.txt b/doc/resolv_wrapper.1.txt
> > > index 12f96d0..63e40d4 100644
> > > --- a/doc/resolv_wrapper.1.txt
> > > +++ b/doc/resolv_wrapper.1.txt
> > > @@ -47,6 +47,7 @@ For example:
> > >      AAAA    dc.cwrap.org fd00::5357:5f0a
> > >      CNAME   kerberos.cwrap.org dc.cwrap.org
> > >      SRV     _kerberos._tcp.cwrap.org kerberos.cwrap.org 88
> > > +    URI     _vpn.cwrap.org https://vpn.cwrap.org/VPN
> > >  
> > >  *RESOLV_WRAPPER_DEBUGLEVEL*::
> > >  
> > > diff --git a/src/resolv_wrapper.c b/src/resolv_wrapper.c
> > > index 48018be..a07086a 100644
> > > --- a/src/resolv_wrapper.c
> > > +++ b/src/resolv_wrapper.c
> > > @@ -73,6 +73,8 @@
> > >  #define ns_name_compress dn_comp
> > >  #endif
> > >  
> > > +#define ns_t_uri 256
> > > +
> > Can you wrap this #define with #ifndef?
> 
> When else would ns_t_uri be defined? It isn't in any public header. And
> my patch to glibc was rejected based on the fact that nameser.h will be
> deprecated. So it will never be defined anywhere except internally.
Ok, that detail wasn't in the original email and I didn't read the whole
thread on libc-alpha.

LGTM, then.

-- 
/ Alexander Bokovoy