It seems good to me.

It seems good to me.

No, please don't do that. It looses the feature of the resolver. You should change as follows like NetBSD did:

#ifndef DEBUG
#define DEBUG
#endif

Oops, I have submitted it accidentally at the middle. X-(
It seems you forgot to include the diff for resolv.h in your previous patch.
Other than that, it seems good to me.

There are some codes in res_setoptions() which touch statp->_u._ext.ext but we disabled them. I think it is better to do something like following in res_setoptions():

@@ -599,9 +595,7 @@
 {
        const char *cp = options;
        int i;
-#ifndef _LIBC
        struct __res_state_ext *ext = statp->_u._ext.ext;
-#endif

Could you move reload_period as well? We should not touch struct __res_state as much as possible to ease merging further upstream changes.

In D4239#90029, @bapt wrote:

Fix with locale ja_JP.UTF-8

It seems good to me.

In D2216#33, @thomas wrote:

Indeed, the weak aliases make sense, and avoid the need to bump the major. I'll produce an update patch. Thanks!

No, we don't need to bump shlib major until removing weak_reference.
The weak_reference is for keeping ABI backward compatibility.

The use of `__' is intentional, here. It seems you are confusing ABI
with API. Though it changes the ABI of libmd, as far as the
applications include the corresponding header file, it doesn't change
the API.
Use of symbol versioning is overkill for this purpose, IMHO.