Otherwise, only part of it gets pulled in through <windows.h> on
original MinGW.

Fixes https://github.com/curl/curl/issues/2361
Closes https://github.com/curl/curl/pull/2721

When size_t is not a typedef for unsigned long (as usually the case on
Windows), GCC emits -Wformat warnings when using lu and lx format
specifiers with size_t. Silence them with explicit casts to
unsigned long.

Closes https://github.com/curl/curl/pull/2721

... not the read buffer size, as that can be set smaller and thus cause
a buffer overflow! CVE-2018-0500

Reported-by: Peter Wu
Bug: https://curl.haxx.se/docs/adv_2018-70a2.html

Closes #2718

...but GCC users lose out on TLS 1.3 support, since we can't weak-link
enumeration constants.

Fixes #2656
Closes #2703

Variable 'output_var' is not used and can be removed.
Function 'collect_true' renamed to 'count_true'.

Closes #2711

... because otherwise not everything get closed down correctly.

Fixes #2708
Closes #2712

Closes #2713

Closes #2706

Closes #2704

telnet.c(1401,28): warning: cast from function call of type 'int' to
non-matching type 'HANDLE' (aka 'void *') [-Wbad-function-cast]

Fixes #2696
Closes #2700

Closes #2698

The code treated the set version as the *exact* version to require in
the TLS handshake, which is not what other TLS backends do and probably
not what most people expect either.

Reported-by: Andreas Olsson
Assisted-by: Gaurav Malhotra
Fixes #2691
Closes #2694

Reported-by: Andreas Olsson
Fixes #2692
Closes #2693

The previous example was a little bit confusing, because SSL* structure
(or other "in use" SSL connection pointer) is not accessible after the
transfer is completed, therefore working with the raw TLS library
specific pointer needs to be done during transfer.

Closes #2690

... since default uses the threaded one and we test the c-ares build
already.

Closes #2689

Since it isn't used either and requires the getnameinfo check

Follow-up to 0aeca417

Closes #2687

... and trim the threaded Curl_resolver_getsock() to return zero
millisecond wait times during the first three milliseconds so that
localhost or names in the OS resolver cache gets detected and used
faster.

Closes #2685

The linker is pretty dumb and processes things left to right, keeping a
tally of symbols it hasn't resolved yet. So, we need -ldl to appear
after -lcrypto otherwise the linker won't find the dl functions.

Closes #2684

... it was previously unchecked in two places and thus errors could
remain undetected and cause trouble.

Closes #2681

Follow-up to b6a16afa

... to work with longer passwords etc. Grow it from a 256 to a 4096
bytes buffer.

Reported-by: Dario Nieuwenhuis
Fixes #2676
Closes #2680

Closes #2673

Fixes #2677
Closes #2679

Follow-up to 2c15693a.

Bug #2674
Closes #2675