It seems that some systems (e.g. fairly consistently in some recent
Solaris autobuilds) would manage to get to the connect phase before the
progress callback was called, resulting in a CURLE_COULDNT_CONNECT
error. Reworked the test to point at a test server that never returns a
full result so the progress callback always gets a chance to be called
before the transfer can complete in some other way.
# Server-side
<reply>
+<data nocheck="yes">
+HTTP/1.1 204 PARTIAL\r
+X-Comment: partial response to keep the client waiting\r
+</data>
+<postcmd>
+wait 10
+</postcmd>
</reply>
# Client-side
<client>
<server>
-none
-</server>
-<features>
http
-</features>
+</server>
<tool>
lib1513
</tool>
# this server/host won't be used for real
<command>
-http://localhost/1513
+http://%HOSTIP:%HTTPPORT/1513
</command>
</client>