sizeof(ip6_addr_t) > 16 depending on compilation flags, possibly reading
more data than available.

Fix by using the actual address' sizeof for memcpy.

When running the test on `arduino-mega2560` printing the float failed
and was printed as ` ?`.

    Calculated  ? bits of entropy from 10000 samples.

Replace using `printf` floating point printing by using `fmt/print_float`.
Now the test succeeds on `arduino-mega2560`.

As `print_float` does not buffer and is used with `printf` the output
should be flushed before calling it if `fflush` is available.

cpp check is reporting the following error

    error (shiftTooManyBitsSigned):
    Shifting signed 32-bit value by 31 bits is undefined behaviour

Fix by ensuring the `1` is an `uint32_t` before shifting.

Improved readability of if statements. This issue was pointed out by cppcheck.

This driver currently only supports the SPI connected DS3234, and only
for initializing 1 Hz square wave output on the SQW pin, nothing else.

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

Updated the test to use NETOPT_MAX_DPU_SIZE instead of the deprecated
NETOPT_MAX_PACKET_SIZE

After being cleared, the memory can be `0xFF` for `atmega` platforms for
example.

Fix the test to take EEPROM_CLEAR_BYTE into account.

`strcmp` was used to check that the 4 bytes were empty but was not
testing what it should as an empty string is `0` bytes long.

The whole must be verified.

Use `memcmp` where it is obvious it is not doing a string comparison.

Casting to `const char *` is not necessary as input, the type are
already the right ones.

It could not be before as `result` was with the wrong type.

Results type should only be an array of char, not an array of pointers.

To make the CI happy ;-).

Several parts were inspired by a driver written by:
  George Psimenos (gp7g14@soton.ac.uk)
  Steffen Robertz (steffen.robertz@rwth-aachen.de)

Add a test case to the crypto tests testing the chacha20poly1305
algorithm with the test vectors from rfc 8439.

Add positive and negative tests for data bits, stop bits and parity:

- 7E1, 7O1
- 8E1, 8O1
- 8N2

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>

Add command mode that will be used like this:

mode <dev> <data bits> <parity> <stop bits>

This command must be called after init otherwise
the UART won't be fully initialized.

Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com>

The msp430 toolchain is missing an `fputs()` implementation. This commit makes
them use the `printf("%s", str);` instead of `fputs(str, stdout);`, which is
semantically equivalent (but has more overhead).