Conversation
Mirrors the dbval minimalism pass: - `clojure.data/diff` would have to realize both databases entirely in memory; the `Diff` extension now throws an informative exception (and is kept only so diff does not fall back to diffing the record fields). Its comparators were also silently broken: they still compared UUID entity ids with `Integer/compare`. - Remove `query_v3`, the datom comparators (`cmp-datoms-*`, `defcomp`, `combine-cmp`, `diff-sorted`, `cmp`, `value-cmp`), `case-tree`/ `case-pick`/`vpred`, the dead `-search` bindings, the mutable `idx` field on `Datom`, and the broken `core/settings`. - Drop `nippy` (never used) and `persistent-sorted-set` (its array helpers are now the tiny `slateval.arrays` namespace). - `empty-db` docstring now documents `:db-file`/`:object-store-url` instead of persistent-sorted-set options. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
`tempid` and `resolve-tempid` had no callers; `d/db` and `transact` stay. The Datascript `squuid` implementation is removed — entity and transaction ids use com.yetanalytics/colossal-squuid (`gen-id`). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The storage layer is SlateDB via its JVM uniffi bindings, so the CLJS variant (disabled since the fork) cannot work without an entirely different backend; git history keeps it recoverable. - Delete `slateval.js`, `deps.cljs`, `externs.js`, `release-js/`, `test/js/`, `slateval.test.cljs` and the cljs/js test+bench scripts. - Strip every `#?`/`#?@` reader conditional (keeping the :clj branches) and rename all `.cljc` sources and tests to `.clj`. - Remove the CLJS-only macro layer: `defn+` (now plain `defn`), `if-cljs`, `cljs-env?`, `patch-tag`; `defrecord-updatable` keeps only its JVM implementation. - Drop the :cljs alias and transit-cljs from deps.edn. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Remove `bench_datomic/`, `test_datomic/`, the :datomic alias and the datomic peer dependency, plus the wrapper scripts. `script/bench.sh` replaces bench_clj/bench_all; test_all.sh is gone — CI runs `test_clj.sh` directly. - Rename the runner to `slateval.bench.run` and adapt it to the SlateDB/UUID/strict-snapshot world: transaction benches replay against a fresh store per iteration with `*batch*` 1 (one-tx-per-op benches use a 100-people dataset — a SlateDB commit flushes durably at ~100ms); read benches build a 20k store once; entity lookups use sampled UUIDs and lookup refs; freeze/thaw dropped (jsonista/cheshire removed). - Bugfix found by the bench: storing an empty vector as a value NPEd in `tuple` (& rest args are nil for zero components, Tuple.addAll requires a List). With regression test. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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.
Summary
Applies the same minimalism steps as dbval's conn-without-atom PR (the parts that are about removing code — the conn/deftype architecture changes were not ported, see below). Net: −3,220 lines, three dependencies dropped.
Drop
clojure.data/diff+ Tier 1 dead codediff-similarnow throws: a diff would have to realize both (potentially larger-than-memory) databases. The extension is kept only so diff doesn't fall back to diffing the record's fields. Its comparators were also silently broken — they still compared UUID entity ids withInteger/compare.query_v3(~1,000 lines), the datom comparators (cmp-datoms-*,defcomp,combine-cmp,diff-sorted),case-tree/case-pick/vpred, dead-searchbindings, the mutableidxfield onDatom, the brokencore/settings.slateval.arrays).Tier 2 compat shims
tempid/resolve-tempidremoved (hollowed out by UUID ids; no callers); the duplicate Datascriptsquuidremoved — ids come from com.yetanalytics/colossal-squuid viagen-id.d/dbandd/transactstay.ClojureScript/JS removal
The storage layer is SlateDB via JVM uniffi bindings, so the disabled CLJS variant can't work without a different backend. All
#?reader conditionals stripped, every.cljc→.clj,slateval.js/deps.cljs/externs.js/release-js//test/js/deleted, the CLJS-only macro layer (defn+,if-cljs,patch-tag) removed (defrecord-updatablekeeps its JVM implementation), :cljs alias and transit-cljs dropped.Bench refactor, Datomic comparison dropped
bench_datomic/,test_datomic/, the :datomic alias and datomic peer dependency are gone. The runner is nowslateval.bench.run(viascript/bench.sh), adapted to UUID ids and strict snapshots: transaction benches replay against a fresh store per iteration; one-tx-per-op benches use a 100-people dataset since a SlateDB commit flushes durably (~100ms each); read benches build a 20k store once. All bench categories smoke-tested.Bugfix found by the bench: storing an empty vector as a value NPEd in
tuple(& restargs are nil for zero components,Tuple.addAllrequires a List). Fixed with a regression test.Not ported (deliberately)
dbval's conn-is-not-an-atom redesign, the DB/FilteredDB deftype conversion, and reference-identity semantics. slateval's
as-of/since/history/with-dry-runfeatures are built on the record shape and snapshot semantics, so that's a separate decision with its own review.Test plan
main: 159 tests, 1045 assertions, 0 failures.datafy/pprintload-checked (not covered by the suite); every bench category smoke-tested.🤖 Generated with Claude Code
https://claude.ai/code/session_014dX8tTR4yFh5SyBGw3atpo