Evan Martin

Notes to self, there are two broad situations we care about: 1. User generates build files the first time, invokes n2. 2. User invokes ninja generator manually atop existing build....

Argh, sorry! I have reproed locally, taking a look.

Please let me know if that works for you, I hope I got it right this time...

I am back to thinking about this again because Meson also triggers this. [Relevant code](https://github.com/mesonbuild/meson/blob/27028bd3b173539fd93eed880ad31a73fc7d1719/mesonbuild/backend/ninjabackend.py#L701) only invokes `-t restat` if there is a `.ninja_log` file in the build dir, which...

(The reason the memory can move is if the call_x86 code eventually allocates more memory, then we resize the underlying Vec, which invalidates any pointers into it.)

We have some functions defined in assembly. To link them on Mac they are named e.g. `_tramp64` but you write it without the underscore in the Rust extern declaration, this...

@cadmic got retrowin32 running on x86-64 Mac, which means we have at least all the CPU-initialization bits in place such that actual x86 hardware believes us. So in theory all...

Came here because I saw this linked from the docs, two small comments: - [the mention in the docs](https://jj-vcs.github.io/jj/latest/revsets/#the-all-modifier) says "This option is planned to be removed." but it wasn't...

Now that I've typed that all out, possibly the right answer is that these `.cv_file` bits generated from Rust are just wrong and this should be fixed there.

Workaround: use a target with the same assembly syntax but with the host system so the paths are sane, e.g. `x86_64-apple-darwin` on Mac.