metagn

This isn't optimized out, it's ignored. `exportc` should error on `const`. You should be using `let`.

For the package failures you can use what I did in #20199 [here](https://github.com/nim-lang/Nim/pull/20199/files#diff-536ce5b720c0d86d37341fb21a90eb4f23b408da9003f1aa8bb9704de74e8972) and add `nimble install -y arraymancer@#HEAD;` before the test commands of the packages that depend on arraymancer.

Can we use the [date command](https://unix.stackexchange.com/questions/151547/linux-set-date-through-command-line) to test it? edit: lol, [the last commit](https://github.com/nim-lang/Nim/pull/20338/commits/dfcdb6ec2ab6a5fa53b6a99294a84fd122be8f8d) says it was made on Sep 12, 2038

Wouldn't it be better to have 2 separate types in the `options` module? Or is this just a module rename that happens to justify one behavior more?

I've had a little-thought-out, probably unproductive opinion about this for a while and I've decided this being tangentially brought up in a forum post just now is a good enough...

Are there places where we can refactor the use of `assert` into proper errors? Also, `assert`/`doAssert` is used in nearly every test/runnable example. Should `std/assertions` be implicitly imported in these?...

> could you please elaborate more? For example: https://github.com/nim-lang/Nim/blob/e0be71cd046e458e87498d2dd9cb69babe4f17e7/lib/impure/nre.nim#L488-L489 A user can give an `endpos` that is outside of the string bounds. If they do this they will get an...

I just wrote https://github.com/nim-lang/RFCs/issues/468, all library authors would have to do is add `{.pure.}`/whatever annotation to their enums.

`--` expects code from the input stream I believe. So 1. it should not allow `:values`, and 2. it should error when a filename is given to the command.

I believe the PR is ready now. I can undo the following things if needed: * deprecated symbols removed by `nimPreviewSlimSystem`, I just thought the best way to remove these...