duphandle doesn't preserve CURLOPT_RESOLVE

[ernstp]

I did this

struct curl_slist *host = curl_slist_append(NULL, "foo.net:80:127.0.0.1");

CURL *curl = curl_easy_init();
curl_easy_setopt(curl, CURLOPT_VERBOSE, 1L);
curl_easy_setopt(curl, CURLOPT_RESOLVE, host);

CURL *c2 = curl_easy_duphandle(curl);
curl_easy_setopt(c2, CURLOPT_URL, "http://foo.net/");
curl_easy_perform(c2);

I expected the following

I expected my CURLOPT_RESOLVE to still have effect on c2. However it does not.
It's interesting that the message
"Added foo.net:80:127.0.0.1 to DNS cache"
never appears so it looks like it's neither done for the first curl handle nor c2.

Moving the setopt for RESOLVE to c2 works.

I can't really find anything in the documentation that would hint to this behaviour.

curl/libcurl version

curl 7.55.1 (x86_64-pc-linux-gnu) libcurl/7.55.1 OpenSSL/1.0.2g zlib/1.2.11 libidn2/2.0.2 libpsl/0.18.0 (+libidn2/2.0.2) librtmp/2.3
Release-Date: 2017-08-14
Protocols: dict file ftp ftps gopher http https imap imaps ldap ldaps pop3 pop3s rtmp rtsp smb smbs smtp smtps telnet tftp
Features: AsynchDNS IDN IPv6 Largefile GSS-API Kerberos SPNEGO NTLM NTLM_WB SSL libz TLS-SRP UnixSockets HTTPS-proxy PSL

operating system

Linux/Ubuntu

[bagder]

Thanks!

My tentative fix above also adds a test for this. I think the only thing to add is to use caution with this since now both handles use the same list so you must make sure to not free the list until both are done with it!

[bagder]

Thanks!

My tentative fix above also adds a test for this. I think the only thing to add is to use caution with this since now both handles use the same list so you must make sure to not free the list until both are done with it!

[ernstp]

Right, but that caution applies in a lot of other similar cases as well, right? Like adding headers etc. So it shouldn't surprise anyone...

[ernstp]

Right, but that caution applies in a lot of other similar cases as well, right? Like adding headers etc. So it shouldn't surprise anyone...

[bagder]

Absolutely, it just struck me when I did the change and modified the test code to use this...

[bagder]

Absolutely, it just struck me when I did the change and modified the test code to use this...

[bagder]

Thanks!

[bagder]

Thanks!