On Tue, 25 Sep 2018, Kuan-Sheng Chen wrote:

> 2. Isn't a "server error" then the response code to that CONNECT request?
> I've did multiple experiments like server's DNS name is wrong, or like
> the proxy is down and timeout. It seems that neither http response code nor
> curl error code is a reliable metric for differentiation.

I would think they are, but I can't really comment on generalisations.

> I've been using DEBUGFUNCTION for a while and it indeed helps me a lot to
> find the root cause. The problem is that it's not scalable and unable to be
> used in production code since we've got to enable "CURL_VERBOSE", and it's
> expensive.

Please elaborate. How is it expensive?

> Is there any way to get the information from DEBUGFUNCTION's my_trace
> callback without enabling CURL_VERBOSE ?

If you want the data you need to have it enabled.

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:   https://curl.haxx.se/mail/etiquette.html