Ensure hexval and int don't share BindVar after Normalization

[williammartin]

Description

This PR relates to #14450, adding a test and attempting to resolve what appears to be an incorrect sharing of a bind var between a HexVal e.g. x'31' and an Int e.g. 31. This occurs because the type is not considered as part of the key for the BindVar map.

There was some existing code that attempted to disambiguate literal values as map keys by prefixing a single quote. Since this proved to not be correctly handling all situations, I thought it appropriate to key by the Literal itself. This ensures that a literal's Type is also considered first class in disambiguation.

I chose to use the Literal rather than *Literal because literals with the same type and value are intended to share a BindVar.

There's now slightly more data in the key since it is a struct with a Type field of size Int, but since the key previously was an arbitrary length string, the difference should be negligible. Running the benchmarks didn't seem to indicate anything of concern. I ran them using go test -bench="BenchmarkNormalize" -run="^#" ./go/vt/sqlparser/.

That said, it's totally possible (likely even) that I don't have the full picture here. I just kind of stumbled across this yesterday and have no other experience with vitess so...take it all with a grain of salt! I'm also not at all opposed to doing another approach like the existing single-quote prefix. It just seemed like more edge cases waiting to happen.

I have no opinions on backporting because I'm not an operator and don't know anything about vitess' processes.

Related Issue(s)

Fixes #14450

Checklist

• "Backport to:" labels have been added if this change should be back-ported
• Tests were added or are not required
• Did the new or modified tests pass consistently locally and on the CI
• Documentation was added or is not required

Deployment Notes

[vitess-bot]

Review Checklist

Hello reviewers! 👋 Please follow this checklist when reviewing this Pull Request.

General

• Ensure that the Pull Request has a descriptive title.
• Ensure there is a link to an issue (except for internal cleanup and flaky test fixes), new features should have an RFC that documents use cases and test cases.

Tests

• Bug fixes should have at least one unit or end-to-end test, enhancement and new features should have a sufficient number of tests.

Documentation

• Apply the release notes (needs details) label if users need to know about this change.
• New features should be documented.
• There should be some code comments as to why things are implemented the way they are.
• There should be a comment at the top of each new or modified test to explain what the test does.

New flags

• Is this flag really necessary?
• Flag names must be clear and intuitive, use dashes (-), and have a clear help text.

If a workflow is added or modified:

• Each item in Jobs should be named in order to mark it as required.
• If the workflow needs to be marked as required, the maintainer team must be notified.

Backward compatibility

• Protobuf changes should be wire-compatible.
• Changes to _vt tables and RPCs need to be backward compatible.
• RPC changes should be compatible with vitess-operator
• If a flag is removed, then it should also be removed from vitess-operator and arewefastyet, if used there.
• vtctl command output order should be stable and awk-able.

[vmg]

Good catch! I think by keying on Literal we're going to lose some deduplication in some corner cases... but the previous approach just wasn't correct at all. This is a strict improvement for now.

[vitess-bot]

Hello! 👋

This Pull Request is now handled by arewefastyet. The current HEAD and future commits will be benchmarked.

You can find the performance comparison on the arewefastyet website.

[williammartin]

Yeh @vmg I wondered also if that might be the case. Even then it seems like an allow list of shareable values would be a better place to start than with the default assumption that all are shareable. If you have some ideas of which values are shareable, I'd be happy to explore an implementation in another PR.

Cheers all for the quick review and making me feel valued!