Preserve quoted newlines following backslashes

[repzakske]

Formatting a quoted value containing a backslash followed by LF can remove whitespace and normalize line endings inside that value:

query = "SELECT 'a\\\nb  \r\nc';"
assert sqlparse.format(query) == query

The assertion currently fails: the two spaces after b disappear and its CRLF becomes LF. SPLIT_REGEX uses \\. for quoted escapes, but . does not match LF. Enabling DOTALL keeps the complete quoted value together so the serializer only strips whitespace outside it.

Added 32 regression cases for single/double quotes, LF/CR/CRLF, paired backslashes, and default/whitespace/reindent formatting, including keyword transformations. Ten fail before the fix. Independent SQLite checks preserve all 36 tested values after the fix; the original formatter changes eight.

• ran the tests (pytest): 538 passed on each of Python 3.10–3.14; the baseline's two xfails and one xpass are unchanged
• all style issues addressed (ruff check sqlparse/)
• changes are covered by tests
• changes are documented in the changelog

[carlitogwapito]

Here's a small optional test-only commit, directly on your current head.

It combines keyword uppercasing with whitespace stripping or reindentation, using two quoted values containing backslash/LF, trailing spaces and CRLF. The words select and from inside those values stay lowercase.

Two guards worth keeping: both quoted payloads remain exact, and the outside SELECT/AS transformation really happens. That checks a non-identity formatting operation alongside your existing preservation cases.

On Python 3.14.4, both new cases pass your source. The base transforms the outside keywords but loses both literal payloads. The unchanged keyword-case test passes on both. Focused Ruff checks pass; I haven't run the full suite, dialect/SQL-semantic checks, benchmarks, build or version/OS matrix. No SQL was executed.

To run the added file in your tree: python -m pytest tests/test_keyword_transform_quoted_payloads.py -q.

[repzakske]

Here's a small optional test-only commit, directly on your current head.

It combines keyword uppercasing with whitespace stripping or reindentation, using two quoted values containing backslash/LF, trailing spaces and CRLF. The words select and from inside those values stay lowercase.

Two guards worth keeping: both quoted payloads remain exact, and the outside SELECT/AS transformation really happens. That checks a non-identity formatting operation alongside your existing preservation cases.

On Python 3.14.4, both new cases pass your source. The base transforms the outside keywords but loses both literal payloads. The unchanged keyword-case test passes on both. Focused Ruff checks pass; I haven't run the full suite, dialect/SQL-semantic checks, benchmarks, build or version/OS matrix. No SQL was executed.

To run the added file in your tree: python -m pytest tests/test_keyword_transform_quoted_payloads.py -q.

Thanks!