On Fri, 8 Mar 2013, Early Ehlinger wrote:

> I believe I'm being affected by this issue:
>
> http://www.redhat.com/archives/libvir-list/2012-September/msg01960.html
>
> At random intervals, I will get EXC_BAD_ACCESS inside Curl_resolv_timeout on
> iOS,using the curl_easy API from multiple threads. I suspect that the
> description Matthias Bolte provides is what I'm seeing, and setting
> CURLOPT_NOSIGNAL to 1 does seem to make the problem go away. Of course,
> it's a random segfault, so I can't be 100% certain.

Threads and signals really don't mix very well.

> Is anybody looking at an alternative to the sigsetjmp/siglongjmp timeout
> implementation? If this really is what I'm encountering, I would hate to
> give up DNS timeouts to keep the app from crashing.

We have two alternatives already: build with the threaded-resolver backend or
the c-ares powered one.

I can't think of any alternative to signals to abort the stock synchronous
name resolver function. Can you?

-- 
  / daniel.haxx.se
-------------------------------------------------------------------
List admin: http://cool.haxx.se/list/listinfo/curl-library
Etiquette:  http://curl.haxx.se/mail/etiquette.html