sqlite: avoid abort on oversized SQL text

[lazerg]

expandedSQL, sourceSQL, result column names and the sqlite.db.query trace callback passed SQLite text to String::NewFromUtf8() without a length. When that text is longer than String::kMaxLength bytes, V8 hits a fatal CHECK instead of returning an empty handle. Passing the length makes the getters and column names throw ERR_STRING_TOO_LONG, the same as oversized column values since #66209, and the trace callback skips the event so the statement still runs.

NullableSQLiteStringToValue() and CreateSQLiteErrorImpl() have the same problem with huge identifiers, but the authorizer and error message paths need their own handling, so I left them for a follow-up.

Fixes: #66487

[nodejs-github-bot]

Review requested:

• @nodejs/sqlite

[github-actions]

Caution

AgentScan found account activity patterns that may be consistent with automation. This is a heuristic, not proof that this pull request was opened by an agent or violates policy. AI-assisted contributions are permitted, but automated tooling must not open pull requests without advance approval, and contributors must personally understand, test, verify, and take responsibility for every submitted change. See the AgentScan analysis, AI use policy, and automation policy for additional context.

[codecov]

Codecov Report

❌ Patch coverage is 85.71429% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.44%. Comparing base (c60762c) to head (68733b9).
⚠️ Report is 80 commits behind head on main.

Files with missing lines	Patch %	Lines
src/node_sqlite.cc	85.71%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #66498      +/-   ##
==========================================
+ Coverage   90.43%   90.44%   +0.01%     
==========================================
  Files         790      791       +1     
  Lines      275435   276520    +1085     
  Branches    52823    53096     +273     
==========================================
+ Hits       249082   250094    +1012     
- Misses      16769    16837      +68     
- Partials     9584     9589       +5     

Files with missing lines	Coverage Δ
src/node_sqlite.cc	82.24% <85.71%> (+0.32%)	⬆️

... and 54 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[mceachen]

The new tests cover expandedSQL and the expanded-SQL branch of TraceCallback. Three other paths this PR changes have no test:

• sourceSQL (SourceSQLGetter)
• column names (ColumnNameToName), reached from the row keys of get(), all(), and iterate(), and from columns()
• the sqlite3_sql() fallback in TraceCallback, used when sqlite3_expanded_sql() returns NULL

Each one aborts v26.10.0, so a test would fail without this change. With this change, the first two should throw ERR_STRING_TOO_LONG:

// 'é' is one byte per character in V8 and two bytes in UTF-8.
const n = (constants.MAX_STRING_LENGTH >>> 1) + 1;
db.prepare(`SELECT length('${'é'.repeat(n)}') AS len`).sourceSQL;
db.prepare(`SELECT length('${'é'.repeat(n)}')`).get(); // column name

The fallback repro is in #66487.

Measured with /usr/bin/time on v26.10.0, these peak at 2.4 GB and 3.5 GB, most of it during prepare(). The fallback repro peaks at 2.7 GB. That is more than the suite's enoughTestMem check (os.totalmem() > 1.75 GB) allows for, so they may belong in test/pummel or behind a higher os.totalmem() check.

#66385 rewrites the column-name cache, so a column-name test would catch a regression there. If you are leaving these out on purpose, could you say why in the PR description?

[github-actions]

Failed to resume CI

✖ Failed to load PR CI job 78171: Unexpected token '<', ..."

Full Auto Start CI output

�[36m⠋�[39m Validating Jenkins credentials
�[36m⠋�[39m Validating Jenkins credentials
✔  Jenkins credentials valid
�[36m⠙�[39m Looking for CI runs for pull request 66498
�[36m⠙�[39m Looking for CI runs for pull request 66498
�[36m⠙�[39m Getting PR from nodejs/node/pull/66498
�[36m⠙�[39m Getting reviews from nodejs/node/pull/66498
�[36m⠙�[39m Getting comments from nodejs/node/pull/66498
✔  Found PR CI job 78171
�[36m⠹�[39m Querying data for job/node-test-pull-request/78171/
�[36m⠹�[39m Querying data for job/node-test-pull-request/78171/
�[36m⠹�[39m Querying API for job/node-test-pull-request/78171/

[DEBUG] [SyntaxError: Unexpected token '<', ..."    
  
  <!DOCTYPE "... is not valid JSON]
✖  Failed to load PR CI job 78171: Unexpected token '<', ..."    
  
  <!DOCTYPE "... is not valid JSON

View workflow run

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/78224/

[lazerg]

I added test/pummel/test-sqlite-sql-text-too-long.js in 68733b9. It covers sourceSQL, column names through get(), all(), iterate() and columns(), and the sqlite3_sql() fallback in the trace callback.

For the fallback, the test sets db.limits.length to the byte length of the string literal. Then sqlite3_expanded_sql() returns NULL, and the test does not need the 240 MB blob. Each of the three blocks aborts on main without this fix. The whole file peaks at 3.8 GB here.