The main change here is the timer value that was wrong, it was given in
usecs (ms * 1000), while the itimerspec struct wants nsecs (ms * 1000 *
1000). This resulted in the callback being invoked WAY TOO OFTEN.

As a quick check you can run this command before and after applying this
commit:

    # shell 1
    ./ephiperfifo 2>&1 | tee ephiperfifo.log
    # shell 2
    echo http://hacking.elboulangero.com

 > hiper.fifo

Then just compare the size of the logs files.

Closes #3633
Fixes #3632
Signed-off-by: Arnaud Rebillout <arnaud.rebillout@collabora.com>

- no need to have them protocol specific

- no need to set pointers to them with the Curl_setup_transfer() call

- make Curl_setup_transfer() operate on a transfer pointer, not
  connection

- switch some counters from long to the more proper curl_off_t type

Closes #3627

 - use better variable names to explain their purposes
 - convert logic to curl_multi_wait()

When a transfer is done, the resolver thread will be brought down. That
could accidentally generate an error message in the error buffer even
though this is not an error situationand the transfer would still return
OK.  An application that still reads the error buffer could find a
"Could not resolve host: [host name]" message there and get confused.

Reported-by: Michael Schmid
Fixes #3629
Closes #3630

clarify redir - "in absurdum" doesn't seem to make sense in this context

Closes #3631

in the same sftp_done function in both SSH backends. Simplify them
somewhat.

Pointed out by Codacy.

Closes #3628

Introduced in 8b6314cc, but not used anymore in current code. Unclear
since when.

Closes #3626

Without it set, we would unwillingly triger the "HTTP error before end
of send, stop sending" condition even if the entire POST body had been
sent (since it wouldn't know the expected size) which would
unnecessarily log that message and close the connection when it didn't
have to.

Reported-by: Matt McClure
Bug: https://curl.haxx.se/mail/archive-2019-02/0023.html
Closes #3624

Everyone calls it Secure Transport, now we do too.

Reviewed-by: Nick Zitzmann

Closes #3619

But use the MSYS2 shell rather than the default MSYS shell because of
POSIX path conversion issues. Classic MinGW is only available on the
Visual Studio 2015 image.

Closes https://github.com/curl/curl/pull/3623

Add a MinGW-w64 build using CMake's MSYS Makefiles generator.
Use the Visual Studio 2015 image as it has GCC 8, while the
Visual Studio 2017 image only has GCC 7.2.

Closes https://github.com/curl/curl/pull/3623

Follow-up to 8eddb8f4.

If the cookieinfo pointer is NULL there really is nothing to save.

Without this fix, we got a problem when a handle was using shared object
with cookies and is told to "FLUSH" it to file (which worked) and then
the share object was removed and when the easy handle was closed just
afterwards it has no cookieinfo and no cookies so it decided to save an
empty jar (overwriting the file just flushed).

Test 1905 now verifies that this works.

Assisted-by: Michael Wallner
Assisted-by: Marcel Raad

Closes #3621

Closes #3421

This allows the compiler to pack and align the structs better in
memory. For a rather feature-complete build on x86_64 Linux, gcc 8.1.2
makes the Curl_easy struct 4.9% smaller. From 6312 bytes to 6000.

Removed an unused struct field.

No functionality changes.

Closes #3610

Closes #3616

Closes #3611

Follow-up to 180501cb

Reported-by: Marcel Raad
Fixes #3609
Closes #3620

Switch all Visual Studio 2015 builds to Visual Studio 2017. It's not a
moving target anymore as the last update, Update 9, has been released.

Closes https://github.com/curl/curl/pull/3606

The Visual Studio 2017 image has Visual Studio 2015 and 2017 installed.

Closes https://github.com/curl/curl/pull/3606

Currently, we're using the default Visual Studio 2015 image for
everything.

Closes https://github.com/curl/curl/pull/3606

Instead of using a fixed 256 byte buffer in the connectdata struct.

In my build, this reduces the size of the connectdata struct by 11.8%,
from 2160 to 1904 bytes with no functionality or performance loss.

This also fixes a bug in schannel's Curl_verify_certificate where it
called Curl_sspi_strerror when it should have called Curl_strerror for
string from GetLastError. the only effect would have been no text or the
wrong text being shown for the error.

Co-authored-by: Jay Satiro

Closes #3612

Regression brought by a52e46f3 (shipped in 7.63.0)

Closes #3613

It's passing now.

Closes https://github.com/curl/curl/pull/3615

Remove the Visual Studio 2012 and 2013 builds as they add little value.

Ref: https://github.com/curl/curl/pull/3606
Closes https://github.com/curl/curl/pull/3614

Closes #3591

- Add an additional check for LDAP that also checks for OpenSSL since
  on AIX those libraries may be required to link LDAP properly.

Fixes https://github.com/curl/curl/issues/3595
Closes https://github.com/curl/curl/pull/3596

Add support for Ephemeral elliptic curve Diffie-Hellman key exchange
algorithm option when selecting ciphers. This became available on the
Win10 SDK.

Closes https://github.com/curl/curl/pull/3608

Failing to do so would make the CURLINFO_TOTAL_TIME timeout to not get
updated correctly and could end up getting reported to the application
completely wrong (way too small).

Reported-by: accountantM on github
Fixes #3602
Closes #3605

From within the timer callbacks. Recursive is problematic for several
reasons. They should still work, but this way the examples and the
documentation becomes simpler. I don't think we need to encourage
recursive calls.

Discussed in #3537
Closes #3601

The macro itself has been removed in commit
11974ac8.

Closes https://github.com/curl/curl/pull/3604

since wolfSSL PR https://github.com/wolfSSL/wolfssl/pull/717 (shipped in
wolfSSL 3.10.2 and later) it sends these curves by default already.

Pointed-out-by: David Garske

Closes #3599

and the two remaining #ifdefs for it

Closes #3600

- Split off connection shutdown procedure from Curl_disconnect into new
  function conn_shutdown.

- Change the shutdown procedure to close the sockets before
  disassociating the transfer.

Prior to this change the sockets were closed after disassociating the
transfer so SOCKETFUNCTION wasn't called since the transfer was already
disassociated. That likely came about from recent work started in
Jan 2019 (#3442) to separate transfers from connections.

Bug: https://curl.haxx.se/mail/lib-2019-02/0101.html
Reported-by: Pavel Löbl

Closes https://github.com/curl/curl/issues/3597
Closes https://github.com/curl/curl/pull/3598

As seen in the MinGW autobuilds. Caused by commit
f26bc29c.