Ref. dynare#1949

NB: the “overwrite” option only works for the same type of “learnt_in” value,
i.e. an integer-valued “learnt_in” cannot replace a date-valued “learnt_in” and
vice versa.

This information can be computed from MATLAB/Octave. This will facilitate the
introduction of dates in shock sequences.

– New “heterogeneity_dimension” statement
– New option “heterogeneity” to “var”, “varexo”, “parameters”, “model” and
  “shocks” statements
– New “SUM()” operator in “model” block

This is cleaner than using exit() or assert() for that purpose.

In particular, use a std::variant to store the values of options. This ensures
that a given option name can have only one value (previously, for a given
option name, it was possible to store several values as long as they were of
different types).

As a consequence, add many [[maybe_unused]] attributes for legitimate
cases (mainly in virtual functions).

In particular, use this feature in many loops which feature a special treatment
for the first iteration, using a boolean variable (replacing iterator
manipulation). By the way, also use std::exchange() to simultaneously test the
value of this variable and update it.

For use with perfect_foresight_with_expectation_errors_setup.

Ref. #68

Ref. dynare#1791

Ref. dynare!1844, dynare!1851

dynare#1650

Adapt the code accordingly.

Use ReplaceShorterLiterals option to enforce replacement of all strings.

https://clang.llvm.org/extra/clang-tidy/checks/modernize-raw-string-literal.html

The syntax makes clearer the distinction between a function call and an
initialization. Also, it prevents bugs due to implicit type conversions.