Cat Stevens

I figured I'd ask around, just in case something already exists or is possible for this purpose: [Atomically write a file without changing inodes (preserve hard link)](https://unix.stackexchange.com/questions/478314/atomically-write-a-file-without-changing-inodes-preserve-hard-link), but I suspect...

"Janus, A Summing Up", at least the ones [amazon](https://www.amazon.com/dp/1939438950/) (sorry) [lists](https://www.amazon.com/dp/B0146V2HPY), appears to have nothing to do with computing, and was essentially written by a journalist of the 20th century....

This is great! Definitely one of the most unusual Forth implementations I've come across.

Related: #2393 (closed for that specific case, but same thing)

Ditto for using `IN: vocab.tests` and `IN: vocab.docs` in `-docs` and `-tests` files, which result in the help and tests not loading, and not saying why they don't load.

I really like this, a road to the VM being cleanly documented, well-defined and stuff. But, you meant to use `-Wall` as in `-W all`, there is no `-Weverything` option...

oh, i think GCC has no `Weverything` but clang does, i've just never seen it documented anywhere.

what is preferable here: ```factor : ( from to step/f seq -- slice ) ``` or ```factor : ( from to seq -- slice ) : ( from to step...

See https://github.com/catb0t/factor/commit/09089bf9e90cd237807d5efa059825404a6cf49f. I'll try rebuilding the image, but for some reason `"sequences" reload` when replacing the old slice implementation with the `step-slice` one (but called `slice` etc), Factor freezes and/or...

This will go in the final commit / PR message, but I should explain: - the `dup dip swap dip swap dip swap` silliness is because i need to reach...