On Fri, 11 Nov 2011, Yang Tse wrote:

> 591, 592, 593, 1206, 1207 and 1208 currently fail due to QUIT command
> not being sent.
>
> Given that in these test cases above server is sending preliminary positive
> replies and in some of them transient negative replies over the control
> connection, shouldn't we try to send QUIT command ? Or should we just adjust
> test definitions so that it matches existing behavior?

As long as the control connection is still fine we should indeed send QUIT. I
don't consider it a very important bug though.

> Tests 594 and 1209 do not complete in expected time (runs for ever?). These
> two cases don't even get a preliminary positive reply from the server, so
> maybe test definition should not expect QUIT command for these two.

Hm. Maybe. I need to check them out closer first to see what I really think!

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