adopted deepcopy for cloned codecs

create length of array and count occurrence of object in array

made Then language section more consistent by
disambiguating 'from' and 'inside' in its construct,
ignore 'the' in the section and improve error reporting.
Documentation is updated.

scenario and tests aligned

main test support in bls public key aggregation for multidarkroom

adopted a multi-level structure with name of schema included

both array and dictionary definitions work after the name of
a defined schema to trigger the import of a whole table of
schema objects, useful to import multiple public keys checked
for validity, credentials and other serialized schema objects.

This commit adds the zencode integration test and the crypto
integration test for Arm cortex-m targets.

For testing the Zenroom on the cortex-m3 device, the stderr and stdout
stream should be changed. The stderr messages in Zenroom will be
redirect to semihosting console. The stdout message will be written
into the "outlog" file since the QEMU doesn't have the stderr and stdout
stream.

Since the QEMU also can not receive the stdin stream, the test script
will be written into a temp file. The QEMU will open the temp file
via semihosting to get the test scripts.

now the statements to insert and remove elements from array or
dictionaries are the same, autodetection is made correctly and the
CODEC is kept intact. Dictionary test fixes.

also marginal support for internal types: ecp, ecp2 and big

extend array comparison and hash
as well separate implementation over different source files
remove 'eval' function to improve security

also no need to use of encapsulation decoding in schema

test correctly addresses use of named rather than inside