Re: also with res_mkquery

[Robert Millan]

note, i'm adding CC to the bug log and bug-hurd mailing list

On Fri, Apr 11, 2003 at 03:22:40PM +0200, Stephan Trebels wrote:
> Isn't the bug rather, that the res_mkquery test is invalid in autoconf? 
> Even in my linux box, I get:
> 
> stephan@stephan-lap:~$ objdump -T /lib/libresolv.so.2 | grep res_mk
> 0000592d  w   DF .text  00000079  GLIBC_2.0   res_mkquery
> 0000592d g    DF .text  00000079  GLIBC_2.2   __res_mkquery
> 
> This seems to indicate, that the test will fail for a pure 2.2
> libresolv, or am I too naive here?
> 
> The test binary generated linked using your tested C code, produces:
> 
> stephan@stephan-lap:~$ objdump -Rt a.out | grep res_mk
> 08048324       F *UND*  00000079              res_mkquery@@GLIBC_2.0
> 080495a0 R_386_JUMP_SLOT   res_mkquery
> 
> Is there any "official" requirement, that res_mkquery needs to be a
> callable symbol?  All I read in the manpages was, that after three
> includes, the res_mkquery function is callable, which it probably is.

I don't know what to say about this. but note that the manpages are
not a reliable reference, the official documentation for Glibc is
the info libc document (which doesn't say anything about res_mkquery)

-- 
Robert Millan

make: *** No rule to make target `war'.  Stop.

Another world is possible - Just say no to genocide