Bump junit-jupiter from 5.8.0 to 5.8.1 - #81

dependabot · 2021-09-27T04:09:28Z

Bumps junit-jupiter from 5.8.0 to 5.8.1.

Release notes

Sourced from junit-jupiter's releases.

JUnit 5.8.1 = Platform 1.8.1 + Jupiter 5.8.1 + Vintage 5.8.1

See Release Notes.

Commits

9883ffa Release 5.8.1
68b08a7 Fix unintended exceptions in assertLinesMatch
248ac22 Add missing public constructor
ba7f6a7 Polish 5.8.1 release notes
5eca40d Prepare 5.8.1 release notes
34217e4 Introduce support for text blocks in @CsvSource
155a25f Deprecate @UseTechnicalNames
49e1901 Polishing 5.8.1 release notes
3242bf6 Make ReflectionSupport.findNestedClasses thread-safe
7896665 Add #2722 to release notes
Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

@dependabot rebase will rebase this PR
@dependabot recreate will recreate this PR, overwriting any edits that have been made to it
@dependabot merge will merge this PR after your CI passes on it
@dependabot squash and merge will squash and merge this PR after your CI passes on it
@dependabot cancel merge will cancel a previously requested merge and block automerging
@dependabot reopen will reopen this PR if it is closed
@dependabot close will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually
@dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
@dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
@dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)

Bumps [junit-jupiter](https://github.com/junit-team/junit5) from 5.8.0 to 5.8.1. - [Release notes](https://github.com/junit-team/junit5/releases) - [Commits](junit-team/junit-framework@r5.8.0...r5.8.1) --- updated-dependencies: - dependency-name: org.junit.jupiter:junit-jupiter dependency-type: direct:development update-type: version-update:semver-patch ... Signed-off-by: dependabot[bot] <support@github.com>

….junit.jupiter-junit-jupiter-5.8.1 Bump junit-jupiter from 5.8.0 to 5.8.1

…IONS ArcadeData#81 withdrew that Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

The forty-operation query set (#82d) with its per-table counts, the one measurement set every table reports (ArcadeData#89), answer equivalence as a rule, an invariant and two field traps (ArcadeData#88), durability matched at the relaxed end with the engines that cannot be (ArcadeData#81), the October start order and the preview switch (ArcadeData#83, ArcadeData#84, ArcadeData#86), and the comparator re-pin list with its two pinning defects (ArcadeData#87). FAIRNESS gains F10 durability, F11 close cost (the number fairness_check already prints), and F12 equivalence; F10 and F11 were taken by code before the file documented them, so equivalence takes the next free number rather than renumbering either. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

…Data#82 query set, ingest/index split, bench_host and instrument on every row (DECISIONS ArcadeData#74, ArcadeData#81, ArcadeData#82, ArcadeData#84) Lanes: TPC-C payment beside new-order and three line-item analytics on every document engine; a 3-hop filtered read and a delete pass on every graph engine; two TSBS queries on every time-series engine; ingest_s and index_s on ArcadeDB, pgvector, Neo4j, Milvus, and LanceDB; every row records durability and instrument; sqlite-vec on the SQLite PRAGMAs; MongoDB timed writes at w=1, j=false; PostgreSQL and PG+AGE read synchronous_commit from the server. Checks: make_paper_tables refuses two instruments in one table; fairness F8 refuses a 2026-10 row without a durability in class. Exporter: the new columns appear only when a lane's rows all carry instrument 2026-10. The runner half of this instrument (bench_host, the PostgreSQL and SurrealDB server flags, the paper-tier BENCH_HOST refusal) sits in 7a26139 on this branch and is re-applied in its own commit at rebase time. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

…ne, and the SurrealDB server flag that was never read DECISIONS ArcadeData#81 asks for the engine's own answer and the string a lane writes onto a row is a published claim, so each was checked on the laptop (2026-09-14) against the pinned image or wheel. The evidence sits in bench_common above the strings it justifies: ArcadeDB GlobalConfiguration.TX_WAL_FLUSH, read out of the running engine: default 0, current 0, "0 = no flush". SQLite PRAGMA read back (wal, 1); strace, 50 commits -> 8 fsync. DuckDB strace, 50 commits -> 55 fsync; duckdb_settings() at 1.5.5 has no commit-sync knob, only checkpoint thresholds. LadybugDB strace, 50 auto-commit writes -> 56 fdatasync; ladybug 0.20.4's Database() takes no sync option. MongoDB server 8.2.12 reports journalCommitInterval 100 ms, and an implicit default of w:majority with journaling. ArangoDB server 3.12.11 answers database.wait-for-sync false, rocksdb.use-fsync false, rocksdb.sync-interval 100; a new collection reads back waitForSync false. QuestDB server 9.1.1 SHOW PARAMETERS: cairo.commit.mode nosync, from the default. SurrealDB strace A/B on the embedded core: SURREAL_SYNC_DATA unset gives 6 fsync at both 50 and 250 commits, set gives 56 and 256. Neo4j SHOW SETTINGS at 2026.07.1 has no durability or sync setting, so it cannot be relaxed. THE ONE THAT WAS WRONG. The served SurrealDB arms were started with SURREAL_DATASTORE_SYNC_DATA=never and their rows said so. The 3.2.4 binary contains no "SYNC_DATA" and no "SURREAL_DATASTORE" token, and none of the 110 SURREAL_* variables it does expose names sync, WAL, fsync, or durability: the flag was never read. It labelled four arms as relaxed while changing nothing, which is the BENCH_GAV=0 shape this harness already has a comment about. The flag is gone from runner.py, and because what 3.2.4 does at commit could not be established, the string says "behaviour at commit is not verified" instead of claiming a class. durability_class gains a third answer, "unverified", and fairness_check refuses it on any backend outside UNVERIFIED_ALLOWED, so an unchecked default cannot spread from these four arms to a fifth. Also: one string per engine, defined in bench_common, so two lanes cannot describe one engine differently; and the Kuzu name is gone from the whole tree (the engine is LadybugDB, package ladybug), including the four places outside this work that still carried it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2 Swept in with it, because they were in the tree: the first half of the DECISIONS ArcadeData#86 skeleton publish (BENCH_SKELETON in export_web, make_paper_tables, fairness_check, page_check, and refresh_web_page's --skeleton). Nothing exercises it yet; the commit that runs it follows.

…eData#86) The October page's shape has to be readable before mini measures anything, so the whole instrument runs on the laptop at micro scale, one repetition, and publishes to the preview route as a skeleton. A skeleton that looks like a measurement is worse than no skeleton, so everything below is a refusal or a label, not an option. WHAT A SKELETON MAY BE. refresh_web_page --skeleton implies --preview, reads its own freeze, and refuses unless every row carries a bench_host that is not the bench host and the sweep tier; a LIVE publish refuses any payload the exporter stamped skeleton. The freeze has its own file, runs_skeleton_laptop .csv, written from its own log (BENCH_RUNS_JSONL), so the campaign's tracked runs_paper.csv cannot be overwritten even for the minutes a publish takes, and the page's source link names the file it really used. WHAT IT SAYS. The payload carries skeleton, a banner, the waivers, and the tables a laptop cannot draw with the reason for each; every table repeats the placeholder warning in its own conditions, because a reader who lands on one table never sees the banner. Sizes are the ones actually measured and marked as such, so no Size column names a corpus the cell never read. WHAT IT STILL HAS TO PASS. F1 (cpuset pinning) and F3 (memory envelope) are waived and named, because both describe the bench host. Everything else runs: the degree match, the close cost, the durability class, the instrument, the engine identity, the recall floors. page_check keeps the structural half of its two typed-number sections -- every pinned sentence must still be present, exactly once -- and reports the values as placeholders instead of comparing them to filler; the DEEP-10M cross-generator section says it does not apply at micro scale. The atomicity counts are NOT waived: they read the skeleton's own rows. A skeleton draws only what it measured. The dense and sparse multipass overlays, the client/server decomposition, and the Python-cost table are all bench-host artifacts, so they are withheld and named rather than published under a laptop banner. Also here, and not skeleton-specific: the page host is read from the rows' bench_host instead of the literal "mini" the exporter has carried since the field did not exist, and every table gains a durability condition generated from its own engines' rows (DECISIONS ArcadeData#81) instead of the one global paragraph describing September's mixed defaults. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

…l (DECISIONS ArcadeData#90) ArcadeData#90 supersedes the single-setting half of ArcadeData#81: every timed write runs twice, once relaxed and once strict. In scope, the six document operations, the three graph writes, and the cross-model transaction; out of scope, bulk ingest and every read path. AN AXIS, NOT A SECOND MEASUREMENT INSIDE A CELL, because most engines cannot switch per operation: ArcadeDB's txWalFlush is per database, SurrealDB's and QuestDB's are server flags. So `runner.py --durability strict` is one flag, the class reaches the client container as BENCH_DURABILITY and the server containers as patched flags, and the row carries `durability_class`, `durability_no_setting` and `durability_server_flags` beside the `durability` string it already carried. The class is part of the run_id and of the canonical key: without that a strict cell writes the same key as the relaxed one beside it and the later silently shadows the earlier, which is the shadowing shape make_paper_tables already documents twice. READ BACK, NOT ASSERTED, ON BOTH SIDES, because a flag that does not take is exactly what ArcadeData#81 was written after. ArcadeDB's value comes from GlobalConfiguration.TX_WAL_FLUSH after the database is open (laptop A/B: the JVM property gives 0 relaxed and 2 strict, and the engine says so); SQLite's from PRAGMA journal_mode and PRAGMA synchronous; ArangoDB's from the collection's own properties()["sync"]; the PostgreSQL family's from SHOW synchronous_commit. stamp_durability takes the engine's string VERBATIM and never re-maps it, so a relaxed answer to a strict request survives to the gate instead of being upgraded into a claim. fairness_check F10b fails that row, and fails a write cell that exists in only one class. FOUR ENGINES HAVE NO KNOB, not three. ArcadeData#90 names Neo4j, DuckDB and LadybugDB; the SurrealDB 3.2.4 server belongs with them and the reason is on the record (ArcadeData#81's evidence: no SYNC_DATA and no SURREAL_DATASTORE token in the binary, and none of its 110 SURREAL_* variables names sync, WAL, fsync or durability). Giving it an invented flag would label its rows strict while changing nothing. The four run once, declare it, and the page prints their one number in both columns. Laptop, l1tpc oltp at micro, one repetition, seven engines, both classes, strict over relaxed on new-order p50: SQLite 48.4x, PostgreSQL 9.8x, SurrealDB embedded 9.0x, ArcadeDB embedded 4.6x, MongoDB 1.8x, ArangoDB 1.7x, DuckDB 1.00x. Every one of the six write digests matched across both classes on every engine, which is the other half of the check: a strict commit changes when a write becomes durable, not what it wrote. Also fixed here, found by the smoke: the lifecycle read digest was computed before the measure() calls, so every row said "issues no read" while read_action_ms sat beside it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

…, ts) index Three comparator arms for the time-series lane, on SQLite's footing: no time-series type, one table with a datetime field, the tag, the three metrics, and a composite (host, ts) index defined before the load. surrealdb_ts (SDK 2.0.0, core 2.3.10, SurrealKV) and surrealdb_ts_server (3.2.4 on RocksDB, the pinned digest) run one SurrealQL text: time::floor for the minute and hour buckets, d'...' datetime literals for the ranges, and the grouped-ordered-limited query as a subquery, because core 2.3.10 sorts by the group key ascending after GROUP BY and would take the wrong five buckets (BUGS F31 again); 3.2.4 answers both spellings in the same time. The served arm goes through the reconnecting client (DECISIONS ArcadeData#91), the timed loop drops a sample taken across a reconnect, and every SurrealDB row records `reconnects`. arangodb_ts (3.12.11, the pinned digest) holds ts as epoch milliseconds, a persistent index over ["host", "ts"], DATE_TRUNC for the buckets, and the collection is created with waitForSync at the cell's class and read back (DECISIONS ArcadeData#90, ArcadeData#81). Laptop smoke through runner.py (ts100, one rep, BENCH_QITER=3, cpuset 0-11, results/runs_ts_plain.jsonl, untracked): all three cells ok, every query returned the expected shape (1 / 60 / 12 / 1200 / 32,944 / 5 rows, 100 hosts x 12 buckets), and equivalence_check over these rows plus the skeleton's reports every digest equal to the engines already on the table (32 groups agreeing, 0 failures; the one KNOWN entry is the pre-existing arcadedb_ts_native_server q_groupby). Laptop numbers are smoke, not record. Registries: runner BACKENDS and LANES["l4"], export_web display names, L4_CANON_LABELS, order and deployment, fairness_check UNVERIFIED_ALLOWED for the served SurrealDB arm, and one COMPARATORS.md role per engine. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

The first October landing failed, and the failure was worth having. e4 was never registered in the three places a lane has to be registered, so load_canonical discarded all five rows qOI produced: the freeze came out empty and make_paper_tables had nothing to build from. It went unnoticed because the E4 TABLE reads results/e4decomp_<pin>/ rather than rows, so nothing downstream complained. The PAPER_SCALES comment warned about exactly this and named two precedents -- lifecycle and l4, 117 rows worth ~18 h of bench time, discarded before any table, figure or gate saw them. e4 is the third, and the gate that would have caught the next problem is the thing the omission was hiding: registered -> the rows reach F10 -> F10 fails five e4 rows for recording no durability at all. Which is correct. The lane runs an engine, the engine has a txWalFlush setting, and every other lane reads it back out of the engine rather than asserting it (ArcadeData#81, ArcadeData#90). e4 recorded nothing. Fixed in e4_decomp, and qOK re-runs the lane to get rows that carry it -- three minutes of machine time, since qOI ran 17:12:20 to 17:15:02 for one engine and five reps. So the landing is not blocked on a decision, it is blocked on a re-run that costs three minutes at the end of the chain. The artifacts qOI produced are fine and the table still reads them. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JB6Hg77dQVqABoTJmiUnV2

dependabot Bot added dependencies Pull requests that update a dependency file java labels Sep 27, 2021

robfrank self-requested a review September 27, 2021 07:08

robfrank approved these changes Sep 27, 2021

View reviewed changes

mergify Bot merged commit 23c80e6 into main Sep 27, 2021

dependabot Bot deleted the dependabot/maven/org.junit.jupiter-junit-jupiter-5.8.1 branch September 27, 2021 07:09

tae898 pushed a commit to humemai/arcadedb-embedded-python that referenced this pull request Jun 28, 2026

Merge pull request ArcadeData#81 from ArcadeData/dependabot/maven/org…

11ea7a1

….junit.jupiter-junit-jupiter-5.8.1 Bump junit-jupiter from 5.8.0 to 5.8.1

Navigation Menu

Uh oh!

Bump junit-jupiter from 5.8.0 to 5.8.1 - #81