fix: Like::TestString to align with Java LIKE semantics

[lxy-9602]

Purpose

No Linked issue.

Description:

• Validate LIKE escape sequences like Java Paimon
• Treat trailing \ and invalid escapes as errors
• Make _ match one UTF-8 character instead of one byte
• Keep % matching any sequence
• Remove alloca and use std::vector<bool> instead

Tests

• Add invalid escape sequence tests
• Add escaped \ and % tests
• Add UTF-8 multi-byte _ matching tests
• Add line terminator semantics tests

API and Format

Documentation

Generative AI tooling

Aone Copilot (Claude)

[Copilot]

Pull request overview

Fixes Like::TestString to match Java Paimon's LIKE semantics: strict escape validation, UTF-8 code-point-aware _ wildcard, and Java regex line-terminator handling. Also replaces the alloca-based DP buffer with a std::vector<bool>.

Changes:

• Parse pattern in two phases: validate escapes (only \_, \%, \\ permitted; trailing \ and other escapes raise Status::Invalid) and merge consecutive %.
• Decompose both pattern and field into UTF-8 code points so _ matches one character (not one byte) and rejects Java line terminators (\n, \r, U+0085, U+2028, U+2029); % still matches anything. Min-length quick reject now counts _ as a required char.
• Replace alloca / unique_ptr<bool[]> DP storage with std::vector<bool>; add tests covering invalid escapes, escaped \/%, multibyte _, and line-terminator semantics.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/paimon/common/predicate/like.cpp	Rewrites Like::TestString for Java-compatible escape validation, UTF-8 _ semantics, line-terminator handling, and std::vector-based DP.
src/paimon/common/predicate/predicate_test.cpp	Adds four TEST_F cases for invalid escapes, escaped backslash/percent, UTF-8 multibyte _, and line-terminator semantics.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[zjw1111]

+1