David Hill

ACS front is missing StrCpy, which is needed for compliance. (And because it's potentially useful.) This should be done using a header declaration like Print or CreateTranslation.

Would be nice to have subnormal support in the softfloat implementation. Should be possible, but I would want to ensure it doesn't negatively affect the performance of normalized numbers. Possibly...

Need support for the float formats in printf (%e, %f, %g). The absence of which is becoming increasingly conspicuous.

Current implementation can't represent special float values (NaN/infinity/subnormals), which is an unfortunate limitation. Will require replacing use of mpf_class for IR::Value_Float's value.

While support for complex types is optional in C11, it would be nice and shouldn't be too difficult with the right method of implementation.

Being able to have a pointer type that crosses address spaces would be potentially useful in edge cases. Currently planning on doing a three-word solution. One word for type, one...

Currently, compound literals work by generating an object, then making a `*(obj = init, &obj)` expression with it. This has several unfortunate consequences. Particularly the fact that it takes the...

This would allow use of the return statement in scripts which yield a value, as otherwise they internally have a return type of void. Should be grammatically unambiguous to simply...

It really is time to replace all those global functions in Platform/Platform.hpp. Platform::Info should be a pared down analogue to Bytecode::Info, and the latter should probably be constructed with a...