On Tue, 11 May 2010, Howard Chu wrote:

>> I've now verified that this indeed the case. Test case 400 etc fail due to
>> FTPS breakage, and the breakage is because the code now always uses the
>> same send/recv functions for both FTPS connections which it can't assume...
>>
> Oops. Use two sets of send/recv function pointers in the conn struct then?

Yes exactly, indexed with the 'num' passed in to those functions.

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