add pure in-memory shapedb

[magnetised]

No description provided.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 88.77%. Comparing base (58c06d8) to head (3fb82e9).
⚠️ Report is 3 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3934      +/-   ##
==========================================
+ Coverage   87.21%   88.77%   +1.56%     
==========================================
  Files          25       25              
  Lines        2394     2415      +21     
  Branches      599      607       +8     
==========================================
+ Hits         2088     2144      +56     
+ Misses        304      269      -35     
  Partials        2        2              

Flag	Coverage Δ
packages/experimental	87.73% <ø> (ø)
packages/react-hooks	86.48% <ø> (ø)
packages/start	82.83% <ø> (ø)
packages/typescript-client	94.01% <ø> (+2.13%)	⬆️
packages/y-electric	56.05% <ø> (ø)
typescript	88.77% <ø> (+1.56%)	⬆️
unit-tests	88.77% <ø> (+1.56%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[blacksmith-sh]

Found 1 test failure on Blacksmith runners:

Failure

Test	View Logs
Elixir.Electric.Postgres.ReplicationClientTest/ test ReplicationClient against real db flushes are advanced once confirmed, and advance beyond the last seen txn once up-to-date	View Logs

[alco]

Note: This review was produced by Claude (Opus) and vetted by @alco.

Correctness Issues

1. "After restart" test coverage gap (shape_cache_test.exs)

The @describetag skip: not ShapeStatus.ShapeDb.persistent?() on the "after restart" describe block skips the entire block when InMemory is active. But the test "can recover consumers after restart" was specifically reworked to use the new restart_shape_cache_only/2 helper, which keeps the InMemory shape db alive across the restart. That test should be exercisable with InMemory — the whole point of the helper change is to avoid needing persistence. As written, it will never run because the describe-level skip tag blankets it.

Either that test should be moved out of the describe block, or it should get @tag skip: false to override the describe-level tag.

2. Dead SQLite config in config.ex

config.ex still computes defaults from ShapeDb.Sqlite.Connection.default!(:synchronous) and ShapeDb.Sqlite.Connection.default!(:cache_size). These flow into shape_db_opts in the main opts keyword, but MonitoredCoreSupervisor now does Keyword.take(opts, [:stack_id]) for the InMemory supervisor — shape_db_opts is never extracted. These config entries are dead code. If the SQLite module ever fails to compile (e.g., exqlite dep removed), these will break config.ex at compile time.

3. persistent?/0 on Sqlite.Supervisor is orphaned

persistent?/0 was added to both Sqlite (the main module) and Sqlite.Supervisor. The delegation from ShapeDb goes to @implementation which is InMemory, so ShapeDb.persistent?/0 returns false. Nobody calls Sqlite.Supervisor.persistent?/0. It's unclear why it's on the Supervisor rather than only on the Sqlite module.

Things that look correct

• The ETS table layout and key design ({:shape, handle}, {:comparable, comparable_hash}, :count) is clean and avoids key collisions.
• comparable_to_hash/1 matches the hashing logic in ShapeDb.Sqlite.Query, so shapes are identified consistently across implementations.
• handle_for_shape_critical/2 correctly aliases to handle_for_shape/2 since ETS reads are always consistent.
• validate_existing_shapes/1 correctly removes incomplete shapes and adjusts the counter.
• All SQLite module renames under ShapeDb.Sqlite look mechanically correct — module names, aliases, and function references are all updated consistently.
• The InMemory test suite is thorough and mirrors the ShapeDb test structure well.
• The restart_shape_cache_only / stop_shape_cache_and_db split in shape_cache_test.exs is a good factoring.