Use chunked file reading to avoid loading entire files into memory

[JanTvrdik]

Summary

• Added PatternIterator — a reusable, chunk-aware regex matching iterator over Iterator<string> streams. It accumulates chunks, yields match arrays, supports mid-iteration pattern changes via setPattern(), and handles chunk-boundary safety (waits for more data when a match reaches the end of the buffer).
• Added BaseMultiQueryParser — an abstract base class implementing IMultiQueryParser with parseFile() (chunked 64 KiB reading via fopen/fread), parseString(), and an abstract parseStringStream() method.
• Refactored all three parsers (MySqlMultiQueryParser, PostgreSqlMultiQueryParser, SqlServerMultiQueryParser) to extend BaseMultiQueryParser and use PatternIterator instead of BufferedFileParseTrait. No changes to regex patterns or public API.
• Zero-length regex matches (e.g. \z end-of-content markers) are handled safely — they break out of both loops to prevent infinite iteration and ensure consistent behavior regardless of chunking.
• Added comprehensive PatternIteratorTest (33 tests) covering empty/trivial inputs, single/multi-chunk matching, capture groups, chunk-boundary safety, generator-based streams, single-byte chunks, pattern mutation, error handling, zero-length matches, SQL-like patterns, and comparison with the MySQL query pattern.
• Added randomized chunking test to MySqlMultiQueryParserTest — runs 100 iterations with random chunk sizes (1–256 bytes) against mysql.sql to verify results are identical regardless of chunking.

Test plan

• All tests pass (42 tests)
• PHPStan reports no errors (composer phpstan)
• Verify with a large SQL file (100+ MB) that memory stays bounded

[Copilot]

Pull request overview

This PR refactors the multi-query parser to use chunked file reading instead of loading entire files into memory, aiming to reduce memory usage for large SQL files. The implementation introduces a shared BufferedFileParseTrait that reads files in 64 KiB chunks using fopen()/fread() and refactors all three database-specific parsers (MySQL, PostgreSQL, SQL Server) to use this trait. A safety mechanism is included to prevent false \z regex anchor matches at chunk boundaries.

Changes:

• Introduced BufferedFileParseTrait with chunked file reading logic (64 KiB chunks)
• Refactored MySQL, PostgreSQL, and SQL Server parsers to use the new trait with callback-based pattern processing
• Added safety check to handle \z anchor edge cases at chunk boundaries

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
src/BufferedFileParseTrait.php	New trait implementing chunked file reading with buffering, pattern matching, and memory management logic
src/MySqlMultiQueryParser.php	Refactored to use BufferedFileParseTrait with callback handling for dynamic delimiter changes
src/PostgreSqlMultiQueryParser.php	Refactored to use BufferedFileParseTrait with static callback for query extraction
src/SqlServerMultiQueryParser.php	Refactored to use BufferedFileParseTrait with static callback for query extraction

---

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

[Copilot]

Pull request overview

Copilot reviewed 13 out of 13 changed files in this pull request and generated 10 comments.

---

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