Escape special characters in Enum/DateTime type names

[slabko]

ClickHouse renders string literals inside type names - Enum labels (Enum8('foo' = 1)) and DateTime/DateTime64 timezones (DateTime('Europe/Amsterdam')) - with C-style escape sequences for control and grammar-significant bytes. Previously clickhouse-cpp neither produced these escapes when generating a type name (Type::GetName()) nor decoded them when parsing a type string, so any label/timezone containing ', \, or a control byte could be corrupted or mis-parsed.

This PR makes escaping symmetric in both directions for the following bytes:

byte	escaped
0x00 NUL	\0
0x08 BS	\b
0x09 TAB	\t
0x0A LF	\n
0x0C FF	\f
0x0D CR	\r
0x5C \	\\
0x27 '	\'

Decode (clickhouse/types/type_parser.cpp)

• The single-quoted-string lexer now unescapes escape sequences instead of copying bytes verbatim, via a new kUnescapeMap plus explicit handling of \' and \0.

Encode (clickhouse/types/types.cpp)

• New EscapeStringLiteral() helper backed by kEscapeMap.

• Applied in EnumType::GetName(), DateTimeType::GetName(), and DateTime64Type::GetName().

• Tuples are intentionally out of scope. Tuple field-name identifiers use a different quoting scheme (backticks) and were reverted to the existing implementation; only a benign non-escape tuple round-trip test is included.

• DateTime/DateTime64 escaping is client-side only in practice. The server validates timezone names against the tz database, so a timezone containing escape sequences can't be stored server-side; the escape/unescape logic is still implemented and unit-tested for correctness.

[Copilot]

Pull request overview

Adds symmetric escaping/unescaping of ClickHouse C-style escape sequences inside single-quoted string literals that appear in type names (Enum labels and DateTime/DateTime64 timezones), preventing mis-parsing/corruption for values containing ', \, NUL, or control bytes.

Changes:

• Encode: introduce EscapeStringLiteral() and apply it to EnumType::GetName(), DateTimeType::GetName(), and DateTime64Type::GetName().
• Decode: update TypeParser’s single-quoted-string lexer to unescape supported sequences into a scratch buffer.
• Tests: add unit + integration coverage for escaped Enum labels and DateTime/DateTime64 timezones, plus CreateColumnByType round-trips.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
clickhouse/types/types.cpp	Adds escaping map + helper and applies escaping when rendering Enum/DateTime type names.
clickhouse/types/type_parser.cpp	Adds unescape map and updates quoted-string tokenization to decode escape sequences.
clickhouse/types/type_parser.h	Clarifies token StringView lifetime and scratch-buffer purpose for processed identifiers/strings.
ut/types_ut.cpp	Adds unit tests verifying escaping in Type::GetName() for Enum labels and DateTime timezones.
ut/type_parser_ut.cpp	Adds parsing tests for escaped Enum labels and escaped DateTime/DateTime64 timezones (including all supported escapes).
ut/CreateColumnByType_ut.cpp	Adds CreateColumnByType tests for escaped inputs and render→parse round-trips; includes tuple header for tuple round-trip test.
ut/client_ut.cpp	Adds live-server integration test for Enum escape round-trip via server-reported schema.

---

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

[slabko]

This is, unfortunately, true and can result in errors different from what they really are if the type name is incorrect. However this fix require a separate PR.