Subject: use BIND 8 parsing API

We should be using the BIND 8 parsing API (ns_initparse, etc.) rather
than relying on the HEADER struct (which I think is BIND 4).
Apparently, using -DBIND_8_COMPAT to make the HEADER struct visible can
turn off the res_nsearch() prototype, etc.  If we use the ns_* API on
BIND 8 and 9, we can stop using -DBIND_8_COMPAT.

Affected files are lib/krb5/os/dnssrv.c and lib/krb5/os/hst_realm.c.

To: rt@krbdev.mit.edu
Cc: 
Subject: Re: [krbdev.mit.edu #2710] use BIND 8 parsing API
From: Sam Hartman <hartmans@mit.edu>
Date: Wed, 15 Sep 2004 16:52:52 -0400
RT-Send-Cc: 

we should also make a decision on whether we care about bind4 enough
to maintain SRV and TXT record support on bind4 only systems.

To: rt-comment@krbdev.mit.edu
Cc: krb5-prs@mit.edu
Subject: Re: [krbdev.mit.edu #2710] use BIND 8 parsing API 
Date: Thu, 16 Sep 2004 10:19:32 -0400
From: Ken Hornstein <kenh@cmf.nrl.navy.mil>
RT-Send-Cc: 

>We should be using the BIND 8 parsing API (ns_initparse, etc.) rather
>than relying on the HEADER struct (which I think is BIND 4).
>Apparently, using -DBIND_8_COMPAT to make the HEADER struct visible can
>turn off the res_nsearch() prototype, etc.  If we use the ns_* API on
>BIND 8 and 9, we can stop using -DBIND_8_COMPAT.

It seems to me that at least under MacOS X 10.3.5 and Solaris 8, the
ns_initparse() API isn't prototyped in resolv.conf.

--Ken

To: Ken Hornstein <kenh@cmf.nrl.navy.mil>
Cc: rt-comment@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2710] use BIND 8 parsing API
From: Tom Yu <tlyu@mit.edu>
Date: Thu, 16 Sep 2004 11:36:18 -0400
RT-Send-Cc: 

>>>>> "kenh" == Ken Hornstein <kenh@cmf.nrl.navy.mil> writes:

kenh> It seems to me that at least under MacOS X 10.3.5 and Solaris 8, the
kenh> ns_initparse() API isn't prototyped in resolv.conf.

It's in <arpa/nameser.h.>

---Tom

From: tlyu@mit.edu
Subject: CVS Commit

Add DNS resolver glue layer.  Use it.


To generate a diff of this commit:



	cvs diff -r5.471 -r5.472 krb5/src/ChangeLog
	cvs diff -r1.302 -r1.303 krb5/src/aclocal.m4
	cvs diff -r1.466 -r1.467 krb5/src/include/ChangeLog
	cvs diff -r1.50 -r1.51 krb5/src/include/configure.in
	cvs diff -r5.398 -r5.399 krb5/src/lib/krb5/os/ChangeLog
	cvs diff -r1.90 -r1.91 krb5/src/lib/krb5/os/Makefile.in
	cvs diff -r5.5 -r5.6 krb5/src/lib/krb5/os/dnssrv.c
	cvs diff -r5.54 -r5.55 krb5/src/lib/krb5/os/hst_realm.c
	cvs diff -r0 -r5.1 krb5/src/lib/krb5/os/dnsglue.c
		krb5/src/lib/krb5/os/dnsglue.h

From: tlyu@mit.edu
Subject: CVS Commit

Don't assume that presence of res_nsearch() means we have ns_initparse().


To generate a diff of this commit:



	cvs diff -r5.472 -r5.473 krb5/src/ChangeLog
	cvs diff -r1.303 -r1.304 krb5/src/aclocal.m4
	cvs diff -r5.399 -r5.400 krb5/src/lib/krb5/os/ChangeLog
	cvs diff -r5.1 -r5.2 krb5/src/lib/krb5/os/dnsglue.c

To: rt@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2710] CVS Commit
From: Sam Hartman <hartmans@mit.edu>
Date: Tue, 21 Sep 2004 19:06:40 -0400
RT-Send-Cc: 

Except on the affected platforms I'd expect to find ns_initparse.
Certainly my arpa/nameser.h prototypes it.

To: rt-comment@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2710] CVS Commit
From: Tom Yu <tlyu@mit.edu>
Date: Tue, 21 Sep 2004 21:37:09 -0400
RT-Send-Cc: 

The problem is, ns_initparse() is prototyped, but the library doesn't
export the sysbol.  Yay broken OSes.

---Tom

To: rt@krbdev.mit.edu
Cc: krb5-prs@MIT.EDU
Subject: Re: [krbdev.mit.edu #2710] CVS Commit
From: Sam Hartman <hartmans@mit.edu>
Date: Tue, 21 Sep 2004 21:44:21 -0400
RT-Send-Cc: 

>>>>> "Tom" == Tom Yu via RT <rt-comment@krbdev.mit.edu> writes:

    Tom> The problem is, ns_initparse() is prototyped, but the library
    Tom> doesn't export the sysbol.  Yay broken OSes.

That's not what I see:

libresolv.a:ns_parse.o:00000080 T __ns_initparse

To: rt@krbdev.mit.edu
Subject: Re: [krbdev.mit.edu #2710] CVS Commit
From: Tom Yu <tlyu@mit.edu>
Date: Tue, 21 Sep 2004 22:25:07 -0400
RT-Send-Cc: 

>>>>> "Sam" == Sam Hartman via RT <rt-comment@krbdev.mit.edu> writes:

>>>>> "Tom" == Tom Yu via RT <rt-comment@krbdev.mit.edu> writes:
Tom> The problem is, ns_initparse() is prototyped, but the library
Tom> doesn't export the sysbol.  Yay broken OSes.

Sam> That's not what I see:

Sam> libresolv.a:ns_parse.o:00000080 T __ns_initparse

On all-in-one, I see:

/usr/lib/libresolv.so:00009810 t __ns_initparse

Note the lowercase "t" for a local symbol.  The static library appears
to have the symbol, but the shared library reduces it to a local symbol.

---Tom