warnings

assume that the curl file is a proper binary, it is often a script file
produced by libtool.

for a reason I don't know

initial basic valgrind-log scan for memory leaks it could detect.

trace if a test results in a 'core' file and gdb is present

libcurl.

from time to time!

anywhere. We need to write a better ftps-server for test purposes and then
we can re-introduced FTPS tests.

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.

  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.

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

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

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.

attached. And some cosmetic fixes.

the execution of the tests