Repository navigation
v2.0.10 — 2026-10-23
Release Notes
Patch release on the v2.x line under the support policy: two data-integrity fixes, for WAL entries of more than 1,000,000 facts that replay dropped along with every later transaction (#686) and for float constants that matched any float within f64::EPSILON (#706), plus prepared queries that ignored the database's rule limits (#679). File format is unchanged (v7). No API changes. Upgrading is recommended for every v2.x user who commits transactions of more than 1,000,000 facts or matches float constants in patterns.
Fixed
- A float constant in a pattern matches equal floats only (#706), backported from v3.0.0. A float written in a pattern's value position, or bound to a
$slotof a prepared query, matched every stored float within an absolutef64::EPSILON(about2.2e-16), so[?e :m/x 0.0]returned an entity whose:m/xis1e-17, while[?e :m/x ?v] [(= ?v 0.0)]did not. Above 1.0 the epsilon is smaller than the gap between adjacent floats, so only floats smaller than about2.2e-16in magnitude were affected: every one of them matched0.0, and1e-20matched5e-20. Float constants now compare by IEEE==, as joins, predicates and the indexes do:0.0matches-0.0and nothing else changes. The fix covers patterns in queries, rule bodies andnot/orbodies. Nothing stored was wrong. - WAL entries of more than 1,000,000 facts replay (#686), backported from v3.0.0. A
transact,retractorWriteTransactioncommit of more than 1,000,000 facts wrote a WAL entry that replay took as damage: replay stopped there, and the next open cut the WAL off at that entry, so after a crash, kill or close without a checkpoint that transaction and every later one in the WAL were lost, though each had reported success. Replay now takes an entry's fact count as damage only when the bytes after it cannot hold that many facts. A checkpoint before the crash kept the facts. Present since v0.5.0. - Prepared queries use the database's rule limits (#679), backported from v3.0.0.
PreparedQuery::executeran a rule query with the defaultmax_derived_factsandmax_resultsof 1,000,000 whateverOpenOptionsset, so a lower limit set to stop a runaway recursive rule did not apply to prepared queries, and a higher one set for a large derivation failed withINT-020whereexecuteof the same query succeeded. It now uses the database's limits, asexecutedoes.:max-derived-factsand:max-resultswritten in the query text were always applied. Present since v0.19.0.
Tests
- Kani proofs of the value encoding (#627), backported from v3.0.0.
src/storage/index/verification.rsholds 5 Kani harnesses, compiled only undercfg(kani). They prove for every input thatencode_value, the value part of the v7 index keys, gives each ofNull,Boolean,Integer,FloatandRefits own type byte and keeps each one's order, and that every NaN encodes to the same bytes..github/workflows/kani.ymlruns them on pull requests that touch the encoder and nightly onmainandv3. Strings, keywords, decoders, fact pages and the WAL are not covered: the model checker does not finish on them. Test-only;Cargo.tomldeclares thekanicfg so normal builds do not warn.
Known issues
- Same-transaction multi-valued facts can read back as one value (#371, #287), a fact retracted and re-asserted in one
WriteTransactionis lost at commit (#477), a later assertion does not replace an earlier valid-time window (#435),save()is not crash-atomic (#374), and a.walfrom a later checkpoint is replayed onto an older.graph(#535); all are fixed in v3.0.0. All v2.x known issues, with affected versions and workarounds, are listed in the pinned issue #421.
Download minigraf 2.0.10
| File | Platform | Checksum |
|---|---|---|
| minigraf-aarch64-apple-darwin.tar.xz | Apple Silicon macOS | checksum |
| minigraf-x86_64-apple-darwin.tar.xz | Intel macOS | checksum |
| minigraf-x86_64-pc-windows-msvc.zip | x64 Windows | checksum |
| minigraf-aarch64-unknown-linux-gnu.tar.xz | ARM64 Linux | checksum |
| minigraf-x86_64-unknown-linux-gnu.tar.xz | x64 Linux | checksum |