tsbockman

Unless someone has added floating-point `union` support to DMD since I last checked, this PR has the same problem which stalled mine: [it breaks some parts of `std.math` which currently...

Specifically, this will break `isInfinity` and `isNaN` at CTFE for `double` and `float`: ```D import std.math, std.stdio; void main() { enum float x = float.nan; enum a = isInfinity(x); enum...

That makes sense.

I'm not following druntime development closely now, but I am still available if anyone needs a review or some advice for core.bitop.

The concept here seems right to me, although I can't comment on the implementation. I [bumped the old forum thread](https://forum.dlang.org/post/[email protected]) for this issue just in case someone else sees a...

@schveiguy > There are reasons to use `static if` vs `if`. For instance, `static if` forces CTFE for a function call, and also does not introduce a scope. Although annoying,...

This issue is making IntelliJ almost unusable for me on some projects: D's compile-time meta programming capabilities are intentionally Turing complete, with unbounded time and memory requirements. Trying to recompile...

This version of `core.checkedint` still passes my [preciseops](https://github.com/tsbockman/CheckedInt) test suite, which is considerably more thorough than the druntime unittests. Performance-wise, I find it a bit slower with DMD than the...

(I should note that my tests assume no one has gotten around to implementing intrinsics for `core.checkedint` on DMD yet; I need to retest if this is incorrect.) The bitwise...

I am running the benchmark found in `main.d` of my [CheckedInt](https://github.com/tsbockman/CheckedInt) project. It uses checked arithmetic to do various make-work calculations like prime finding, integer square root, the collatz sequence,...