Adrian Boyko

The current `valid_is_comparand` examines individual comparands, in isolation. As a result, it can only detect "always false" is-expressions and is unable to detect "always true" is-expressions such as: * `c1...

Any thoughts on the possibility of having builtin complex primitive types? I think the RFC is proposing that all complex values will be object instances, which seems like it would...

Did some grunt work in gdb that might be helpful. This is the sequence of calls in the loop, in execution order, with the first two AST args for each....

@mfelsche and I are continuing to look into this issue. I've found that the troublesome infinite recursion can be avoided (at least for the particular Pony code cited in this...

The suggestion was to replace the `eq_typeargs` call in `exact_nominal` with a while loop copied from elsewhere in the code. The while loop was supposed to perform basic syntactic comparison...

@jasoncarr0, I think that push_assume and pop_assume are virtually no-ops without check_assume, so I restored it and didn't produce any traces that don't include check_assume. I just went back to...

@mfelsche, The order matters because the subtyping rules for intersections don't necessarily need to check all types that are intersected to determine that a subtype relation does or does not...

@jasoncarr0, the typeargs are sometimes typeparamrefs, but exact_nominal expects nominals.

Do the `stb__` variables in `imgui_draw.cpp` also need to be dealt with? ``` static unsigned char *stb__barrier_out_e, *stb__barrier_out_b; static const unsigned char *stb__barrier_in_b; static unsigned char *stb__dout; ```

I had the same problem and just killed it. You'll probably see that it has already created the .test_coverage.dart script in your test folder. Unfortunately, .test_coverage.dart also hangs.