feat(frontend): lossy "purify" column default value

[BugenZhao]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

In #19949 we mentioned that

convert ExprNode back to ast can be a bit tricky.

which is the very missing piece of definition purification.

In this PR, we allow converting persisted DefaultColumnDesc back to ast::ColumnOption::DefaultValue losslessly, by simply storing the encoded value temporarily into the AST.

Apparently this is somehow hacky, as unparsing the AST back to SQL will still lose such information. However, this can still be beneficial for...

• reusing the purification code for get_new_table_definition_for_cdc_table
• resolv-ing ALTER TABLE will refresh previous snapshot values for absent cells #17121 once we adopt the purified definition for replacing job

...as long as we believe that AST is the best intermediate representation for schema change.

Checklist

• I have written necessary rustdoc comments.
• I have added necessary unit tests and integration tests.
• I have added test labels as necessary.
• I have added fuzzing tests or opened an issue to track them.
• My PR contains breaking changes.
• My PR changes performance-critical code, so I will run (micro) benchmarks and present the results.
• My PR contains critical fixes that are necessary to be merged into the latest release.

Documentation

• My PR needs documentation updates.

Release note

[BugenZhao]

• feat(frontend): use purified definition for replacing table #20131
• refactor(frontend): reuse def purification for assembling cdc table def when auto schema change #19997
• feat(frontend): lossy "purify" column default value #19993 👈 (View in Graphite)
• feat(frontend): purification for source def sql #19965
• feat(frontend): initially introduce table def sql purification #19949
• refactor(frontend): extract code for parsing table def sql into single stmt #19948
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[BugenZhao]

Previously we short-circuited the procedure once we find all columns are present. Now we should always go through the step for filling the default value.

[fuyufjh]

LGTM