Closes #596

.. and enable SSLpinning tests for mbedTLS, BoringSSL and LibreSSL.

Previously, when HTTP/2 is enabled and used, and stream has content
length known, Curl_read was not called when there was no bytes left to
read. Because of this, we could not make sure that
http2_handle_stream_close was called for every stream. Since we use
http2_handle_stream_close to emit trailer fields, they were
effectively ignored. This commit changes the code so that Curl_read is
called even if no bytes left to read, to ensure that
http2_handle_stream_close is called for every stream.

Discussed in https://github.com/bagder/curl/pull/564

This regression landed in 5778e6f5 and made libcurl not act on received
settings and instead stayed with its internal defaults.

Bug: http://curl.haxx.se/mail/lib-2016-01/0031.html
Reported-by: Bankde

This reverts commit 46cb70e9.

Bug: http://curl.haxx.se/mail/lib-2016-01/0031.html

Discussed in https://github.com/bagder/curl/pull/564

Use the ACE form of IDN hostnames as key in the connection cache.  Add
new tests.

Closes #592

- Fix ALPN reply detection.

- Wrap nghttp2 code in ifdef USE_NGHTTP2.


Prior to this change ALPN and HTTP/2 did not work properly in mbedTLS.

Check that the trailer buffer exists before attempting a client write
for trailers on stream close.

Refer to comments in https://github.com/bagder/curl/pull/564

Mistake from commit a464f338

To make sure curl doesn't allow multiplexing before a connection is
upgraded to HTTP/2 (like when Upgrade: h2c fails), we must make sure the
connection uses HTTP/2 as well and not only check what's wanted.

Closes #584

Patch-by: c0ff

- Add to both curl_global_init.3 and libcurl.3 the caveat for Windows
that initializing libcurl via a DLL's DllMain or static initializer
could cause a deadlock.

Bug: https://github.com/bagder/curl/issues/586


Reported-by:  <marc-groundctl@users.noreply.github.com>

MSYS would otherwise turn a /-style path into a C:\-style path.

While still using datacheck mode binary for the inline reply data.

Move the text-mode conversion for reply/replycheck from the verify
section into the load section and add support for 4 more check parts.

Not converting to double caused small timeouts to be skipped.

Not converting to double caused small timeouts to be skipped.

Previously file.txt[CR][LF] would have been returned as file.tx
(without the last t) if filetype is symlink. Now the t is
included and the internal item_length includes the zero byte.

Spotted using test 576 on Windows.

Follow up on b064ff0c, thanks Daniel.