Update rand to 0.9.x

[dmitrmax]

Closes #789

[Kixunil]

ACK a9fe576

Forgot to check CI

[Kixunil]

You need to run update lock files from contrib directory.

[dmitrmax]

Done

[apoelstra]

You also need to run the formatter and fix the compilation errors with the rand feature enabled.

[Kixunil]

Also please squash the commits.

[dmitrmax]

Ok, i did:

1. Fixed compilation errors
2. Squashed the commits
3. Run the rustfmt. But it formatted a lot of things that I've never touched myself. Is it ok?

[dmitrmax]

I reverted rustfmt for the files which are not affected by the transition to rand v0.9.0. Only those touched by me are reformatted now.

[apoelstra]

Seems like you are using the wrong version of the formatter. You need to use a nightly compiler to run cargo fmt.

[dmitrmax]

Done

[dmitrmax]

Fixed the no-std test

[apoelstra]

The wasm failure looks like it's related to a call to memmove in libsecp. That isn't related to this PR. Similarly the Windows CI failure is hard to read but I think it's also unrelated. I'll run my local CI on this and see if it passes.

[dmitrmax]

Ok. Let me know if there is anything more I can do for this PR.

[apoelstra]

Ok, right now we depend on getrandom 0.2 with the js feature enabled. This is done when compiling for wasm to turn this feature on because rand depends on it.

But rand 0.9 does not depend on getrandom 0.2. It depends on getrandom 0.3. It appears that getrandom 0.3 no longer has the js feature, even though the README for rand 0.9 says it does https://crates.io/crates/rand

I'm not sure what exactly to do here, but we should definitely get rid of the getrandom 0.2 dependency.

[dmitrmax]

Ok, I've updated getrandom to 0.3. Somehow that required to update js-sys crate up to recent version 0.3.77. Sorry, I'm not a wasm guy, I've called contrib/wasm.sh and it ran with the following error when executing wasm-bindgen:
failed to find intrinsics to enable clone_ref function

Not sure what it means.

Also I've updated rand_core dependency from 0.6 to 0.9.

I've made a separate commit to simplify the review. If it looks good to you I can squash both commits into one.

[apoelstra]

Can you squash? I think it's easiest to review this in one commit.

[dmitrmax]

Squashed

[apoelstra]

Thanks for iterating so much! This looks great now. Nice to move to syn 2 across the board. Hopefully this works with our MSRV. CI will tell us.

I will also run my local CI on this which will do a bunch more checks, and then we should be good to merge.

[apoelstra]

ACK 993b4b7; successfully ran local tests