restore: support restoring SQL wrapping with TiDB special comments

[tangenta]

What problem does this PR solve?

TiDB Binlog and Lightning use regex to substitute the TiDB-specific keyword with special comments:

FeatureIDAutoRandom:   regexp.MustCompile(`(?P<REPLACE>(?i)AUTO_RANDOM\b\s*(\s*\(\s*\d+\s*\)\s*)?)`),
...
FeatureIDForceAutoInc: regexp.MustCompile(`(?P<REPLACE>(?i)FORCE)\b\s*AUTO_INCREMENT\s*`),

The regex is fragile because it is easy to get things wrong when the table names or column names contain these keywords, for example:

CREATE TABLE AUTO_RANDOM (id int);

is incorrectly rewritten to

CREATE TABLE `/*T![auto_rand] AUTO_RANDOM AUTO_RANDOM */ ` (id int);

Comparing with pattern matching, using ast.Restore to wrap special comments is not prone to error.

What is changed and how it works?

• Collect TiDB-specific features into tidb/features.go as the single source of truth.
• Add a restore flag: RestoreTiDBSpecialComment, support wrapping with TiDB special comment.
• Remove SpecialCommentsController.

Check List

Tests

• Unit test
• Integration test

Code changes

• Has exported function/method change
• Has interface methods change

Side effects

NA

Related changes

• Need to cherry-pick to the release branch

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• kennytm
• tiancaiamao

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[kennytm]

what is the reason removing SpecialCommentsController and exposing the SpecialComments map directly 🤔

(i mean i don't really like to keep SpecialCommentsController, but allowing anyone to modify the global SpecialComments sounds prone to abuse)

[tiancaiamao]

The old code is more general, the parser is not specific to TiDB and you can register the feature to control it's behavior ...

[tiancaiamao]

OK, since you are the original author of this code, if you thinks it's OK, I have no problem with it.

[tangenta]

the parser is not specific to TiDB

But the keywords(AUTO_RANDOM, CLUSTERED_INDEX...) are TiDB-specific. It is reasonable to attach these keywords to fixed feature IDs.

you can register the feature to control it's behavior

The register mechanism is not that convenient. If the user registers a new feature ID(/*T![new_feature_id] */), he must also modify the code in the parser to add a new keyword to wrap. If the user doesn't need the new keyword, he should use /*T! xxx */ instead.

[kennytm]

/merge

[ti-chi-bot]

@tangenta: Your PR was out of date, I have automatically updated it for you.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the ti-community-infra/tichi repository.

[kennytm]

/merge cancel

[kennytm]

/merge

[ti-chi-bot]

This pull request has been accepted and is ready to merge.

Commit hash: d0b457f