[BUG] Confusion between postifx and binary use of `&` and `*` when keyword is part of the second argument.

[tsoj]

trafficstars

Describe the bug

main: () -> int = {
    b := 2 & int(3);
    std::cout << b << std::endl;
}

Compiling this results in an error:

[1/5] Generating _cppfront/main.cpp
FAILED: _cppfront/main.cpp /home/tsoj/Dokumente/Projects/Starthinker/build/_cppfront/main.cpp 
cd /home/tsoj/Dokumente/Projects/Starthinker/build && /home/tsoj/Dokumente/Projects/Starthinker/ext/cppfront/target/bin/cppfront /home/tsoj/Dokumente/Projects/Starthinker/src/main.cpp2 -o /home/tsoj/Dokumente/Projects/Starthinker/build/_cppfront/main.cpp
/home/tsoj/Dokumente/Projects/Starthinker/src/main.cpp2...
main.cpp2(2,10): error: ill-formed initializer (at '2')
main.cpp2(2,5): error: ill-formed initializer (at 'b')
main.cpp2(1,1): error: unexpected text at end of Cpp2 code section (at 'main')
main.cpp2(1,0): error: parse failed for section starting here

The same happens if I change the & to a *. However, if I change the line to

b := 2 & (int(3));

it works.

Maybe the problem is that here https://github.com/hsutter/cppfront/blob/c618ed507573e935d351b2918a6f1121a6e5bb5c/source/parse.h#L6058-L6073 it is not checked if the next token is one of the keywords that also suggest a binary & instead of a postfix one (i.e. this, const_cast, all the fundamental types, probably more that I am missing right now)

EDIT: I guess this also happens with cases like a := 1 * -(1); though here the error happens in the cpp1 compilation. Or

a: Hello = (); // Hello has a () operator
b := a&;
c := b*();