Still needs to be documented, somehow/somewhere.

The env var OPENSSL_MALLOC_FAILURES controls how often malloc/realloc
should fail.  It's a set of fields separated by semicolons.  Each field
is a count and optional percentage (separated by @) which defaults to 100.
If count is zero then it lasts "forever."  For example: 100;@25 means the
first 100 allocations pass, then the rest have a 25% chance of failing
until the program exits or crashes.

If env var OPENSSL_MALLOC_FD parses as a positive integer, a record
of all malloc "shouldfail" tests is written to that file descriptor.
If a malloc will fail, and OPENSSL_NO_CRYPTO_MDEBUG_BACKTRACE is not set
(platform specific), then a backtrace will be written to the descriptor
when a malloc fails.  This can be useful because a malloc may fail but
not be checked, and problems will only occur later.
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1252)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2208)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2202)

Reviewed-by: NRichard Levitte <levitte@openssl.org>
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2202)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2218)

- On VMS, apps/apps.c depends on apps/vms_term_sock.c, so add it to
  the build
- On VMS, apps/*.c are compiled with default symbol settings,
  i.e. uppercased and truncated symbols, which differs from test
  programs.  Make sure uitest.c knows that with a few pragmas.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2218)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2218)

One of the new tests uses a DH based ciphersuite. That test should be
disabled if DH is disabled.
Reviewed-by: NRichard Levitte <levitte@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2217)

A guard was in the wrong place in the header file.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2212)

It tests both the use of UI_METHOD (through the apps/apps.h API) and
wrapping an older style PEM password callback in a UI_METHOD.

Replace the earlier UI test with a run of this test program
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2204)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

The length passed to tls1_set_sigalgs() is a multiple of two and there are
two char entries in the list for each sigalg. When we set
client_sigalgslen or conf_sigalgslen this is the number of ints in the list
where there is one entry per sigalg (i.e. half the length of the list passed
to the function).
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

The SHA1 sigalgs were inadvertently missed off in the sigalgs refactor.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

The siglen value needs to be initialised prior to it being read in the
call to EVP_DigestSignFinal later in this function.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Declare a variable as static to silence the warning
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Check that signatures actually work, and that an incorrect signature
results in a handshake failure.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

This enables us to make changes to in-flight TLSv1.3 messages that appear
after the ServerHello.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Previously SKE in TLSProxy only knew about one anonymous ciphersuite so
there was never a signature. Extend that to include a ciphersuite that is
not anonymous. This also fixes a bug where the existing SKE processing was
checking against the wrong anon ciphersuite value. This has a knock on
impact on the sslskewith0p test. The bug meant the test was working...but
entirely by accident!
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

TLSv1.3 introduces PSS based sigalgs. Offering these in a TLSv1.3 client
implies that the client is prepared to accept these sigalgs even in
TLSv1.2.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

This also acts as a test for the bug fixed in the previous commit.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

TLSv1.3 freezes the record layer version and ensures that it is always set
to TLSv1.0. Some implementations check this.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

In TLSv1.3 we must use PSS based sig algs for RSA signing. Ignore any
shared sig algs which are PKCS1 based.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

A misreading of the TLS1.3 spec meant we were using the handshake hashes
up to and including the Client Finished to calculate the client
application traffic secret. We should be only use up until the Server
Finished.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

We can't handle these messages yet, so ignore them for now.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

We need to use the length of the handshake hash for the length of the
finished key.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

We were not incrementing the sequence number every time we sent/received
a record.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)

The sigalgs work has made some old lookup tables and functions redundant
so remove them.
Reviewed-by: NRich Salz <rsalz@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/2157)