I fixed what I thought you meant, check the commit to see if I was right.
(you mentioned a different option at the end).

-Sterling

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> I was reading some code concerning the glabal cache and I saw something that
> maybe lead to a little bug.
>
> When we set CURLOPT_DNS_CACHE_TIMEOUT to a handle it do that:
>
> int use_cache = va_arg(param, int);
> if (use_cache) {
> Curl_global_host_cache_init();
> }
>
> data->set.global_dns_cache = use_cache;
>
> and we use global_dns cache only in Curl_global_host_cache_use, this function
> is only used just before calling Curl_perform ( for multi_perform it's
> different but I think this bug will be here too ) and the code look like
> this:
>
> if (!data->hostcache) {
> if (Curl_global_host_cache_use(data)) {
> data->hostcache = Curl_global_host_cache_get();
> }
> else {
> data->hostcache = Curl_hash_alloc(7, Curl_freednsinfo);
> }
> }
>
> it seem okay but In the case of the first time we called perform without the
> CURLOPT_DNS_CACHE_TIMEOUT, and we set after the perform the
> CURLOPT_DNS_CACHE_TIMEOUT option, It seem that at the beginning of the
> perform function the handle will allready have a hostcahe, so it will not set
> the global cache and it will not use it in future perform operation.
>
> Am I right?
>
> - --
> Jean-Philippe Barrette-LaPierre
> Maintener of curlpp: When TRYING(but still in good way) to get the best of C++
> (http://www.sourceforge.net/projects/curlpp)
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.7 (GNU/Linux)
>
> iD8DBQE+GHReb/24YeGNKWQRAoquAJ9TYLtfmgP43Ju4VkkQkdynbSLR7QCfZBeS
> /5uI9JMh7PjLknhMr2MzxWk=
> =Lnwm
> -----END PGP SIGNATURE-----
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
>

-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
Received on 2003-01-06