Use 128 bit instead of Symbol for crate disambiguator

[Xanewok]

As discussed on gitter, this changes crate_disambiguator from Strings to what they are represented as, a 128 bit number.

There's also one bit I think also needs to change, but wasn't 100% sure how: create_root_def. Should I change DefKey::root_parent_stable_hash to accept Fingerprint as crate_disambiguator to quickly combine the hash of crate_name with the new 128 bit hash instead of a string for a disambiguator?

r? @michaelwoerister

EDIT: Are those 3 tests mir-opt failing, because the hash is different, because we calculate it a little bit differently (storing directly instead of hashing the hex-string representation)? Should it be updated like in #45319?

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @michaelwoerister (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[Xanewok]

This truncates the hash to 64 bits, should I implement a 128 but encoding in base_n?

[arielb1]

This is pre-existing?

[Xanewok]

Now sure what you mean exactly, but to_smaller_hash() already exists and truncates to 64 bits.
I meant to ask if base_n::encode should be extended to work with Fingerprint/u128, in addition to u64. To be honest, I saw that it accepts u64, so thought that truncating 128->64 might be acceptable in this case

[michaelwoerister]

The code you wrote should have roughly the same effect as before since DefaultHasher also produces 64 bits. We don't want these hashes to be too long because they end up in directory names and long paths/names sometimes cause trouble for various reasons (file system support, operating system API restrictions, etc).

Maybe you could add a comment here, similar to the one that is replaced:

// The full crate disambiguator is really long. 64 bits of it should be
// sufficient.

[michaelwoerister]

Thanks for the PR, @Xanewok! It already looks pretty good.

Browsing through the code though, I think it is a good idea to introduce a new type for the disambiguator, as in:

#[derive(Copy, Clone, Hash, Debug, ...)]
pub struct CrateDisambiguator {
    fingerprint: Fingerprint
}

impl CrateDisambiguator {
    pub fn to_fingerprint(self) -> Fingerprint {
        self.fingerprint
    }
}

You could put that into librustc/session/config.rs for example. It seems slightly cleaner to me if the disambiguator as its own type.

Regarding the mir-opt tests, yes, just update the hash values.

[michaelwoerister]

Thanks for the PR, @Xanewok! It already looks pretty good.

Browsing through the code though, I think it is a good idea to introduce a new type for the disambiguator, as in:

#[derive(Copy, Clone, Hash, Debug, ...)]
pub struct CrateDisambiguator {
    fingerprint: Fingerprint
}

impl CrateDisambiguator {
    pub fn to_fingerprint(self) -> Fingerprint {
        self.fingerprint
    }
}

You could put that into librustc/session/config.rs for example. It seems slightly cleaner to me if the disambiguator as its own type.

Regarding the mir-opt tests, yes, just update the hash values.

[michaelwoerister]

Usually we pass around Fingerprint by value. It implements Copy so that should be easy.

[michaelwoerister]

The code you wrote should have roughly the same effect as before since DefaultHasher also produces 64 bits. We don't want these hashes to be too long because they end up in directory names and long paths/names sometimes cause trouble for various reasons (file system support, operating system API restrictions, etc).

Maybe you could add a comment here, similar to the one that is replaced:

// The full crate disambiguator is really long. 64 bits of it should be
// sufficient.

[michaelwoerister]

There's also one bit I think also needs to change, but wasn't 100% sure how: create_root_def. Should I change DefKey::root_parent_stable_hash to accept Fingerprint as crate_disambiguator to quickly combine the hash of crate_name with the new 128 bit hash instead of a string for a disambiguator?

Yes, you can do that. It would be a bit more consistent this way.

[Xanewok]

@michaelwoerister created a CrateDisambiguator newtype (awfully similar to DefPathHash) in session/mod.rs. config.rs had a lot of CLI/config-related data and logic and it somehow felt better to put this small type directly inside mod.rs, especially since it's used in Session - hope that's okay.
Addressed comments and fixed tests 🚀

[michaelwoerister]

Excellent, thanks a lot @Xanewok!

@bors r+

[bors]

📌 Commit 7fa64bc has been approved by michaelwoerister

[bors]

⌛️ Testing commit 7fa64bc with merge f764eaf...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: michaelwoerister
Pushing f764eaf to master...

[michaelwoerister]

🎉