fix assignment for nullptr var_value<matrix> and for assigning expressions

[SteveBronder]

Summary

Fixes #2977 by forcing arena_matrix to perform a hard copy when using operator= from expressions made from other var matrices.

Tests

Tests have been added to check assigning to a var_value<Matrix> when the inner vari is nullptr and for var_value<Matrix> when the rhs is an expression and the lhs is just initialized to NaN. Tests. can be run with

python3 ./runTests.py ./test/unit/math/rev/core/var_test.cpp

Side Effects

Nope

Release notes

Bugfix for uninitialized var_value<Matrix> types and assignment when the rhs of the assignment is an expression

Checklist

• Copyright holder: Simon's Foundation

  The copyright holder is typically you or your assignee, such as a university or company. By submitting this pull request, the copyright holder is agreeing to the license the submitted work under the following licenses:
  - Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
  - Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

• the basic tests are passing

  • unit tests pass (to run, use: ./runTests.py test/unit)
  • header checks pass, (make test-headers)
  • dependencies checks pass, (make test-math-dependencies)
  • docs build, (make doxygen)
  • code passes the built in C++ standards checks (make cpplint)

• the code is written in idiomatic C++ and changes are documented in the doxygen

• the new changes are tested

[WardBrian]

@SteveBronder this sounds kind of similar to #2895, which we 'fixed' in stan-dev/stanc3#1323.

Do we think stan-dev/stanc3#1323 is still necessary, or does this also resolve the older issue?

Also, does this resolve stan-dev/stan#3244? Edit: it does

[WardBrian]

A few small things but overall looks good. I've also plugged this branch into some downstream tests of models and confirmed it is working there now

[WardBrian]

Can you a comment just stating that this is intentionally different from the above test and why (e.g., that y did not exist before and therefore does not have its own memory which can be set to 0)

[SteveBronder]

Sure, I also wrote another test here to show that the assign_nan test has the same behavior for both kinds of matrices

[WardBrian]

Looks good to me. Thanks!