clean up dependency tree

[m4rch3n1ng]

this pull request is purely a clean up for your dependency tree and should not change any of the features. i partially used cargo-machete for the low-hanging fruit (which is not super accurate) or just looked at the output from cargo-tree and kind of guessed and then looked if i can remove any of the dependencies.

1. [64364e1]: the rustyline crate by default enables a feature for custom keybinds (custom-bindings) which steel doesn't actually need and this was pulling in a few unnecessary dependencies
2. [3bd2990]: steel-core: remove dependency on serde_derive as you were already depending serde with the derive feature, and remove the pretty dependency. this doesn't actually change much since serde depends on serde_derive and steel-parser depends on pretty, but i think this makes the dependency graph a little cleaner.
3. [a11af18]: remove the dirs and home dependency from steel-core. steel was already pulling in the env_home crate through which and the dirs dependency was actually only used for getting the user home directory, which env_home can do too. if your msrv ever goes above 1.87 (the current nightly) you can actually just use std::env::home_dir() as that was very recently undeprecated (Undeprecate env::home_dir rust-lang/rust#137327).
4. [350d6e5]: similar to point 2 about serde_derive, just in all the other crates.
5. [68e3d64]: remove dev-dependency on serde from steel-interpreter. as it was a dev dependency, this doesn't really affect anything, but it wasn't needed.
6. [fed7492]: i cannot fully replace once_cell with the std equivalents, as std::sync::LazyLock was stabilized only in 1.80 and steel's msrv is 1.76, but i can replace it for the steel-language-server.
7. [371eaa6]: update the colored crate to get rid of lazy_static. since colored was only used by the steel-repl which has an msrv of 1.81 anyway due to rustyline, it should be fine that colored v3 has an msrv of 1.80.
8. [2500789]: steel-core was dependending on the http crate, but wasn't actually using it.
9. [3fcffd4]: there was already a comment in one of the Cargo.tomls about not pulling in the entirety of num and only depending on what was actually needed. this commit does exactly that: it replaces the num dependency and directly relies on num-traits, num-rational and num-bigint
10. [3c090a6]: similar to the commit with the num dependency, but instead with crossbeam where you only needed crossbeam-channel and crossbeam-utils. this change doesn't actually fully remove crossbeam from the dependency tree, as it is still used by termimad, but i opened a pr there as well: breaking: directly depend on crossbeam-channel instead of pulling in all of crossbeam Canop/termimad#69.
11. [75b6bbe]: remove unused env_logger dev dependency in steel-core.

none of these changes are "make-or-break" to me and i would be willing to drop any of the changes if you prefer not to have them.

[mattwparas]

Thanks for this - overall looks good.

At some point in time I had left once_cell in because I was trying to keep the MSRV to match helix at 1.76, however it appears that my branch moved up to 1.81 (not sure when I did that, but that's fine) - I'm wondering if that is the only thing keeping the MSRV at 1.81? I'll do a little poking around before I make a decision here

[m4rch3n1ng]

the 1.81 msrv is mostly due to the home crate, which raises msrv quite frequently. if you want, i can drop the commit replacing the once_cell dependency and replace your dependency on the home crate with something like env_home (or dirs), which i think should allow me to drop the msrv of steel-core back down to 1.76, if you prefer. (steel-repl would still be msrv of 1.81 though, due to a transitive dependency)

[m4rch3n1ng]

one more note: i realized that owo-colors is not actually a drop-in for colored, as colored also only colorizes if the terminal supports it, which would make #312 for example not as trivial to implement. but there actually is a new version of colored that gets rid of lazy_static anyway, and since i cannot bring the msrv of steel-repl under 1.81, it doesn't actually matter that the new colored version has an msrv of 1.80

[mattwparas]

Thanks for the investigation there. I think, if possible, I'd like to try to keep the MSRV for steel core on 1.76 if possible. If its not trivially fixable with the home changes you've already made + removing the commit for the the once cell stuff, then we can merge this as is.

[m4rch3n1ng]

done now. i used the env_home crate to replace home which i think should be good enough? (and it's already in your dependency tree anyway). the msrv goes back up to 1.81 if you enable the git feature, but i don't think i can fix that.

for the once_cell change i decided to keep it for the steel-language-server, but revert it for all the other crates. nvmd that doesn't work, dropped that change completely now.

[m4rch3n1ng]

tbe ci is failing because something on windows is trying to use cmake version less than 3.5. what. i have no idea how to even attempt to fix that lol sorry

[mattwparas]

I saw that happening yesterday, I'll take a look, don't worry about it