Here's a stab at a consolidation of the SSL detection heuristics into
configure. Source files aren't changed by this patch, except for setup.h and
the various config*.h files.  Within the configure script, OPENSSL_ENABLED is
used to determine if SSL is being used or not, and outside configure,
USE_SSLEAY means the same thing; this could be even further unified some day.

Now, when SSL is not detected, configure skips the various checks that are
dependent on SSL, speeding up the configure process and avoiding complications
with cross compiles.  I also updated all the architecture- specific config
files I could see, but I couldn't test them.

be broken, but at least they'll be less broken than they are now."

as suggested by Kai Sommerfeld

reported on Solaris) problem where the size_t check fails due to the SSL libs
being found in a dir not searched through by the run-time linker.
patch-tracker entry #1081707.

--vernum output wasn't zero prefixed properly (as claimed in documentation).
This is fixed in maketgz now.

in this file automatically on releases using the maketgz script.

- removal of getdate.c
- Added hostares.c, hostasyn.c, hostip4.c, hostip6.c, hostsync.c,
hostthre.c, inet_ntop.c, nwlib.c, parsedate.c, sterror.c, strtoofft.c

I have tested the build on 10.3, and will build on 10.2.8 in the next days.

libcurl always and unconditionally overwrote a stack-based array with 3 zero
bytes. I edited the fix to make it less likely to occur again (and added
a comment explaining the reason to the buffer size).

isn't set to encrypted properly

even if no errorbuffer is set.

the multi interface

libcurl without cookie support. This is mainly useful if you want to build a
minimalistic libcurl with no cookies support at all. Like for embedded
systems or similar.

response. Previously, libcurl would re-resolve the host name with the new
port number and attempt to connect to that, while it should use the IP from
the control channel. This bug made it hard to EPSV from an FTP server with
multiple IP addresses!

at a chunk boundary it was not considered an error and thus went unnoticed.
Added test case 207 to verify.