will be properly built with configure --disable-manual even if the source
file is already present

 threaded name resolves.

for it. It saves one extra copy of the header.

I also added comments for several functions in formdata.c

 o minor --fail with authentication bugfix

we can pass on "odd" characters (newline, CR, tab) like this.

option instead (even though it appends the pid to the file name, making it
harder to figure out its name to parse it after a test has run).

Also made sure we only use valgrind for the actual test command command lines,
not when for example running curl to detect if there are any already running
servers are present etc.

situtations in bindings that still have switch() cases for them

Robert Marlow reported.

  stuff added a few weeks ago.  Turns out that if you specify --proxy-ntlm and
  communicate with a proxy that requires basic authentication, the proxy
  properly returns a 407, but the failure detection code doesn't realize it
  should give up, so curl returns with exit code 0. Test case 162 verifies
  this.

as a problem serious enough to skip the final QUIT command before closing
the control connection. To avoid the risk that it will "hang" waiting for
the QUIT response. Added test case 161 to verify this.

we won't try to QUIT the control connection and risk "hanging" waiting for
a response. Test case 161 verifies this. The quit-sending function was
also made static.

we tell valgrind to use that to send the logfile to

less

160 shows.

We got no data and we attempted to re-use a connection. This might happen if
the connection was left alive when we were done using it before, but that was
closed when we wanted to read from it again. Bad luck. Retry the same request
on a fresh connect!

Deleted the sockerror variable again, it serves no purpose anymore.

to re-use a connection that is closed when re-used and libcurl should then
make a new fresh connection and use instead

length limit of the hostname part of the URL.

most 1.1 connections will be closed. The case where it gets closed anyway
is subject for a new separate test.

it actually will close. This is after all what HTTP 1.1 says a server should
do.