feat(img-db): runtime-selectable SQLite backend for the image graph (single-node)#17
Draft
markovejnovic wants to merge 12 commits into
Draft
feat(img-db): runtime-selectable SQLite backend for the image graph (single-node)#17markovejnovic wants to merge 12 commits into
markovejnovic wants to merge 12 commits into
Conversation
- Add Hyper.Img.Db.Repo.Postgres (concrete, priv: priv/repo keeps existing migrations) - Add Hyper.Img.Db.Repo.Sqlite (concrete, same priv path) - Add Hyper.Img.Db.Backend (config-driven selector; selected/0, repo/0, sqlite?/0) - Rewrite Hyper.Img.Db.Repo as a thin runtime facade forwarding Ecto callbacks to Backend.repo() - Add with_low_priority/2 to facade (postgres: statement_timeout tx; sqlite: passthrough) - Rename config block from Repo to Repo.Postgres; add priv: "priv/repo" to both repo configs - Point ecto_repos at Hyper.Img.Db.Repo.Postgres; start Backend.repo() in application tree
Adds SingleNodeGuard GenServer that refuses to boot when peers are already connected and halts the node if a peer joins while SQLite is active. Started conditionally in the application tree only when Backend.sqlite?/0 is true.
Adds a real-SQLite integration test that starts Repo.Sqlite against a
temp file, runs the shared migration, and exercises the three constructs
that could behave differently on SQLite vs Postgres:
- Lease.bump/3 ON CONFLICT upsert (via new bump_with_repo/4 that takes
the repo explicitly, used by tests; bump/3 delegates to it)
- GC prune query: parent_as correlated subquery in delete_all with
RETURNING - confirmed {count, sizes} populates correctly on SQLite
- Image.resolve_chain/1 ordered join (refactored to return a query
instead of executing it, so callers choose the repo)
ecto_sqlite3 note: on conflict update, the returned struct carries a
freshly-generated UUID rather than the existing row's id (ecto_sqlite3
does not RETURNING the stored row after a conflict update). The test
asserts what actually matters: one row persists and the stored expiry
was advanced, verified by Repo.aggregate + Repo.one!.
…tidy pipe - Fix 1: Clarify lease upsert test assertions with precise comments: returned struct check notes it's from changeset (not DB re-read); count==1 assertion proves upsert didn't insert; stored-row re-read proves persistence. - Fix 2: Add log: false to Ecto.Migrator.run to suppress "redefining module" noise. - Fix 3: Replace |> then(&Repo.all/1) with idiomatic |> Repo.all() in chain_sizes/1.
Adds an Image-graph storage backends section to the architecture doc covering both backends, the config snippet to enable SQLite, the SingleNodeGuard boot-refusal and runtime-halt enforcement, per-backend migration commands, and the known upsert-id limitation under SQLite. Updates the README Features bullet to mention the SQLite option for single-node deployments.
…ndle_info clauses
…ph-backend # Conflicts: # mix.lock
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
…ode guard Simplify the SQLite backend per review feedback: - Collapse Repo.Postgres/Repo.Sqlite/Backend into one Hyper.Img.Db.Repo whose adapter is chosen by a single Hyper.Img.Db.Config module (config :hyper, Hyper.Img.Db, backend: :postgres | :sqlite). Restores the default priv/telemetry_prefix, so ecto_repos/CI/migrations use the plain repo name again. - Move the single-node guard out of the DB namespace into a generic, DB-agnostic Hyper.SingleNodeGuard; application.ex decides whether to attach it based on Hyper.Img.Db.Config.sqlite?/0. - Revert test-only scaffolding: Image.resolve_chain/1 executes again and Lease.bump/3 is a single function (drop bump_with_repo/4). - Delete the backend/guard/portability tests. - Move storage-backend docs from architecture.md into the intro cookbook page.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.