Artillery
Half the release notes are a Playwright version bump; the other half is cloud-runner damage control.
A side-by-side editorial comparison of Vitest and Manticore Search — release velocity, themes, recent moves, and the top alternatives to consider.
A major version being built in the open, one breaking change at a time.
Every visible release is a 5.0.0 beta, and every one leads with a Breaking Changes section. The changes are the kind that only land in a major: mocks now clear by default before each test, toHaveTextContent became strict, the benchmark public API was rewritten, config lookup no longer walks ancestor directories, and webdriverio was dropped outright. Roughly one beta every three weeks.
One real release hides inside a stream of per-commit tags: UUID document IDs arrive.
The feed carries two granularities at once — individual commit tags like a dependency bump or a revert, and a single curated release-notes post for 28.6.6 that gathers the actual user-facing work. That release is where the substance sits: UUID document IDs for real-time tables, including generated or explicit IDs and UUID-based filters, REPLACE, UPDATE, and DELETE, plus an ordered and limited GROUP_CONCAT() that keeps the top N values per group. Everything else in the window is version bumps to the Buddy sidecar and the backup tool, a revert, and compatibility repair.
Every visible release is a 5.0.0 beta, and every one leads with a Breaking Changes section. The changes are the kind that only land in a major: mocks now clear by default before each test, toHaveTextContent became strict, the benchmark public API was rewritten, config lookup no longer walks ancestor directories, and webdriverio was dropped outright. Roughly one beta every three weeks.
The through-line is tightening defaults and shrinking surface area. Vitest is removing implicit behavior users had come to rely on — ancestor config discovery, loose text matching, mocks persisting between tests — and consolidating packages, with @vitest/runner inlined and no longer published separately. Browser mode is getting the most net-new investment: trace views in the editor panel, nested mark traces, locators.exact on by default, a session ID required for orchestrator requests. Notably, a large share of these commits are co-authored with coding agents, credited by name in the release notes.
The beta series is still introducing breaking changes as of beta.7, so a stable 5.0.0 does not look imminent; expect at least another beta or two before the breaking-change section stops growing.
The feed carries two granularities at once — individual commit tags like a dependency bump or a revert, and a single curated release-notes post for 28.6.6 that gathers the actual user-facing work. That release is where the substance sits: UUID document IDs for real-time tables, including generated or explicit IDs and UUID-based filters, REPLACE, UPDATE, and DELETE, plus an ordered and limited GROUP_CONCAT() that keeps the top N values per group. Everything else in the window is version bumps to the Buddy sidecar and the backup tool, a revert, and compatibility repair.
New surface area keeps arriving through the Buddy sidecar rather than the core daemon — the latest bump brings Unix socket support — which makes Buddy the component to watch for capability changes. The second thread is fallout from the authentication work: making USER a syntax token broke MySQL client status commands, and 28.6.6 had to restore USER() parsing. The 28.8.1 revert of UTF-8 table and field names, undone because it was merged by mistake, is a reminder that the tag stream is a working branch and not a curated changelog.
UTF-8 table and field names were pulled as an accidental merge rather than rejected on their merits, so the likeliest next move is that support returning deliberately in a later tagged release. Expect further SQL-compatibility patches in the same vein as the USER() fix as more of the auth surface meets real MySQL clients.
Other DevOps products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either Vitest or Manticore Search.
Half the release notes are a Playwright version bump; the other half is cloud-runner damage control.
A terminal Git client that ships release candidates roughly once a year.
Four minor versions in five months, and the shell's remaining work is almost all interaction detail.
Alacritty ships twice a year and spends it on config syntax, Wayland and IME.
Mamba is adding the reproducibility controls conda users have been asking of the solver.
Swagger UI's release stream is dependency bumps and nothing else.
See all Vitest alternatives → · See all Manticore Search alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. Manticore Search is currently shipping more aggressively (velocity 5.0 vs 2.5), with 0 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. Manticore Search is currently shipping more aggressively (velocity 5.0 vs 2.5), with 0 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other DevOps products to evaluate alongside.
Top Vitest alternatives in DevOps are ranked by recent ship velocity. Browse the "Vitest alternatives" section above for the current picks, or visit /alternatives/vitest for the full list with editorial commentary on each.
Top Manticore Search alternatives in DevOps are ranked by recent ship velocity. Browse the "Manticore Search alternatives" section above for the current picks, or visit /alternatives/manticoresearch for the full list with editorial commentary on each.