CI Update tests - make all tests pass (hsutter#1324) * CI Update regression tests * CI Include UFCS test compilation failures * CI Revert UFCS test restore for further investigation * CI update mixed-bugfix-for-ufcs-non-local.cpp2 results

Fix default argument lowering (hsutter#1262) There should now be enough test cases to exercise all the places you can put default arguments Also, only emit Cpp1 lambdas as 'mutable' if there are captures (when it matters) so that pure function expressions are not 'mutable' Closes hsutter#1235 Signed-off-by: Herb Sutter <herb.sutter@gmail.com>

Allow function parameter default arguments Closes hsutter#1189 I had been experimenting with not allowing default arguments for function parameters, in part because of the potential for creating order-dependent code; the way to find out whether they're necessary is to not support them and see if that leaves a usability hole. The result of the experiment is that it does leave a hole: There's persistent feedback that default arguments are often useful, and are actually necessary for a few cases including particularly `std::source_location` parameters. As for order independence, there are already ways to opt into creating potentially order-dependent code (such as by deduced return types which depend on function bodies). So I think it's time to enable default arguments, and Cpp2 is still order-independent by default. This example now works: my_function_name: ( fn: *const char = std::source_location::current().function_name() ) = { std::cout << "calling: (fn)$\n"; } main: (args) = { my_function_name(); } // On MSVC 2022, prints: // calling: int __cdecl main(const int,char **) // On GCC 14, prints: // calling: int main(int, char**)

Update version to 0.7.2 

Add support for concatenating string literals with/without interpolat… …ions And test by applying throughout `reflect.h2` => we don't need a separate regression test case for the feature, `reflect.h2` exercises it well

Merge branch 'main' of https://github.com/hsutter/cppfront