test: prefer in-memory databases in sqlite tests

[bitpshr]

Refs #64665. Converts the SQLite tests that only exercise SQL behavior (transactions, data types, typed-array/data-view binding) from temporary file databases to :memory:, which is faster and drops the nextDb() temp-file bookkeeping.

Tests that verify on-disk behavior (timeout, database-sync) intentionally keep using files. This is a focused first pass; the mixed-usage files (named-parameters, statement-sync) could follow.

Refs: #64665

cc @araujogui, let me know if this is what you had in mind, happy to extend it to the remaining files.

[nodejs-github-bot]

Review requested:

• @nodejs/sqlite

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.14%. Comparing base (f9715fc) to head (a2aed4d).
⚠️ Report is 23 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #64701      +/-   ##
==========================================
- Coverage   90.14%   90.14%   -0.01%     
==========================================
  Files         741      743       +2     
  Lines      242194   242324     +130     
  Branches    45606    45637      +31     
==========================================
+ Hits       218323   218437     +114     
- Misses      15358    15394      +36     
+ Partials     8513     8493      -20     

see 68 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.

[mcollina]

lgtm

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/75095/

[araujogui]

I believe we can extend the pattern to more test files

[araujogui]

The only tests that need temporary files are DatabaseSync() constructor, open(), and backup().

[bitpshr]

Done, extended it per your guidance. Converted the named-parameter and statement suites plus the throwaway databases in the main test file to :memory:.

I kept files in a few places beyond the constructor/open/backup cases where an on-disk or shared database is actually required: the timeout suite and the cross-worker changeset test (both share a database by path), and the PRAGMA journal_mode = WAL test (WAL reports memory on an in-memory database). Let me know if you'd rather handle any of those differently.