Chip Hogg

Weirdly, even a direct test doesn't seem to expose it! ```cpp TEST(CommonSingleFile, CanFindPrimeFactors) { EXPECT_EQ(detail::find_prime_factor(1'000'003u), 1'000'003u); } ``` Here's a plan. First, we'll leave it open for a little while...

Here's a very concise repro: https://godbolt.org/z/jz1oTh8xx I'll ask around and see if I can get some insights.

(To be clear: I meant a repro of the generic problem, where the signedness comes into play for the lvalue but not the prvalue. Still no repro for the Au...

Meanwhile, on the Au repro front: @StefanoD, do you have a precise compiler version and exact/complete list of compiler flags in your build that do show this error?

Thanks for the update! Enjoy your vacation. 😎 🌴 Meanwhile, I have some comments to share. > * We don't use `PrimeFactorization` in our entire code base You probably don't...

I didn't do this. If you wanted to do this it'd be great, thanks!

Thanks for sharing the links!

Interestingly, mp-units doesn't have this problem, because they don't let you apply prefixes to units that aren't "named units": .

Ideas for test cases: - [x] Include `km + mi + mm` --- the first addition creates a common unit, but the second one should "wipe it out" (because `mm`...

Updated notes for posterity: After the latest (and soon-upcoming) changes, this would involve making a new "abstract operation" for _adding_ a magnitude. It would also violate assumptions in the current...