[kevinushey]

Fixes #1519.

The class_<T>::property() overload taking a free-function getter (PROP (*)(Class*)) and a member-function setter fell off the end without returning *this. That is undefined behavior; with Apple clang it compiles to a trap, so any module using the overload aborts the R session as soon as it is loaded. See #1519 for the reprex.

Changes:

• inst/include/Rcpp/module/Module_Add_Property.h: add the missing return *this;.
• inst/tinytest/cpp/Module.cpp: add a ModuleProperties class that exposes one property per property() overload (member, const member and free-function getters, combined with member and free-function setters). Only two of the nine overloads were exercised before.
• inst/tinytest/test_module.R: set and read each read-write property, check that the read-only ones read the same value and reject assignment.

Verification:

• The reprex from module property with a free-function getter and member setter crashes at module load #1519 exits with SIGTRAP (code 133) against current master and prints module loaded and [1] 5 against this branch.
• The new test_module.R traps while sourceCpp() loads cpp/Module.cpp against master's headers (the whole file aborts), and passes on this branch (55/55).
• test_module.R, test_modref.R, test_expose_class.R and test_module_client_package.R pass with RunAllRcppTests=yes RunVerboseRcppTests=yes on R 4.6.1 / macOS. (test_module.R still prints Error: Expecting an external pointer: [type=environment]. at exit; that is the dummy-object finalizer fixed separately in make finalize() a no-op for never-constructed module objects #1518.)

Checklist

• Code compiles correctly
• R CMD check still passes all tests (module test files run locally as above; the change is a one-line header fix plus tests)
• Preferably, new tests were added which fail without the change (without the fix, loading the test module traps)
• Document the changes by file in ChangeLog

[eddelbuettel]

👍

Nice 'lines of tests and documentation' / 'lines of new code' ration too :)