Mike Shaver wrote:
>On Wed, May 01, 2002 at 09:05:03PM -0700, Doug Turner wrote:
>
>
>>>The two solutions that we have on the plate now is (a) call res_init
>>>IFF it is has a known threadsafe implementation. This would mean, on
>>>linux for example, only certain glibc's would work.
>>>
>>>
>>res_init in glibc 2.01 and 2.2.5 is *not* threadsafe.
>>
>>
>
>Aren't there thread-safe variants, mentioned in the URL I provided before?
>res_ninit, or some such. (And what's the failure mode for the lack of
>threadsafety? It sounds like just using the resolver interfaces from
>multiple threads will hork us, since they don't share a resolver state
>if they use the deprecated, non-threadsafe res_init-and-friends APIs.)
>
>That URL again (second hit for google:res_init+thread+safe):
>
>http://sources.redhat.com/ml/libc-hacker/2000-07/msg00487.html
>
>Mike
>
>
>
res_ninit() is *exactly* what we need.
