David Rubin
David Rubin
happens because we run `bitSize` on the array, but never rule out a `comptime_int` array. Same thing should happen if `[1]comptime_int` as the destination size.
debug trace on `0.12.0-dev.1839+461f9c283`: ``` thread 916074 panic: reached unreachable code Analyzing empty.zig: empty.zig:e.U %9 = extended(this()) > %10 = type_info(%9) %11 = field_val(%10, "Enum") %12 = field_val(%11, "tag_type") %13...
@mlugg sorry to bother :D whenever you have a chance to review
will merge the latest changes from master to re-run the ci, maybe itll fix the flaky test 🤔
saw the discussion earlier about this 😃 the ci has already passed when i rebased so it wouldn't make much sense now, but i hope it helps in the future.
> Needs a rebase and then another review from @mlugg Can do right now!
> Should there be an equivalent of `__builtin_expect_with_probability`? It can be used to inform backends with [branch weight metadata](https://llvm.org/docs/BranchWeightMetadata.html#builtin-expect-with-probability) and it appears to be used to guide outlining (function extraction):...
i want to re-explore this issue with the x86_64 backend reaching some level of maturity. should be much easier to implement a solution if it's still needed.
Yeah, my fault. Github had a bit of a moment and decided to mark the entire merge commit as belonging to this PR. I don't think we need a 132k...