On Thu, 9 Feb 2006, Shmulik Regev wrote:

> I guess changing the implementation to honor the caching option is not
> planned, right? :)

I don't normally change things that are working with no reported problems, and
this is no exception. You are the first to bring this up.

> Maybe adding a function to explicitly triggering cache pruning will be
> sufficient. I overcame this problem by using an entirely new multi handle.

What problem did you face because of this?

> I'm no tech writing expert, but maybe a note saying : "Every DNS result is
> cached with regards to the current connection it uses. Specifically when
> using the multi interface with persistent connections each connection will
> cache its DNS results for its entire lifetime". I assume this is also the
> case with easy handles that are reused.

Yes, AFAIR this is the same no matter what interface you use.

-- 
  Commercial curl and libcurl Technical Support: http://haxx.se/curl.html