-
Morphit v1.12.7
StableAll checks were successfulmorphit-ci / TypeScript typecheck (sweep all workspaces) (push) Successful in 47smorphit-ci / apps/web svelte-check (svelte-kit sync + svelte-aware tsc) (push) Successful in 39smorphit-ci / Integration tests (real Postgres 16) (push) Successful in 41smorphit-ci / ansible-lint (playbook quality gate) (push) Successful in 16smorphit-ci / Smoke suite (run-smokes.sh, triple-pulse) (push) Successful in 29m10smorphit-release / Build + publish release tarball (push) Successful in 35m19sreleased this
2026-08-17 22:48:16 +00:00 | 16 commits to main since this releaseMorphit v1.12.7
Theme: federated instances stop showing a false "Build integrity check failed" banner, tor-only visitors never touch the clear net, and health/backup reporting on a fresh node is honest and self-healing.
Fixed
Federated instances no longer show a false "Build integrity check failed" banner. Until now, each operator rebuilt the web frontend locally during install/upgrade, and because those builds aren't byte-reproducible across machines, an honest instance's assets didn't match the hashes published on-chain — tripping the tamper banner on every federated node. The release now ships one canonical, prebuilt frontend, and instances deploy those exact bytes instead of rebuilding (falling back to a local build only if the prebuilt is ever absent). Every instance now serves identical assets that match the on-chain pin. The frontend is generic — each instance still loads its own identity (name, addresses) at runtime — so one build serves the whole federation.
A visitor on a tor-only site never opens a clearnet connection. When a page is served from a
.onion/.i2paddress, the app now uses a hidden-service-only RPC pool with no clearnet fall-through at all — a Tor/I2P visitor's browser can never reach out to the clear net, not even as a fallback. Clearnet instances are unchanged (they still try hidden nodes first for Tor-Browser visitors, then clearnet)."Blurt RPC connectivity" is reported honestly on a fresh/offline node. The indexer counted an RPC endpoint as healthy merely because it wasn't in a cooldown — which is true of every endpoint the instant the node starts, before any probe has run. So a fresh or offline node briefly claimed RPC connectivity it didn't have. An endpoint now counts as healthy only after a real success, so a not-yet-connected node honestly reports zero.
A fresh node's first backup now lands on its own. On a fresh (especially offline) node the indexer can take a couple of minutes to finish migrating its schema; the first backup would fire before the schema existed, correctly skip, and then wait until the next daily run — leaving the operator at "no backup yet." The install now retries the first backup until a real dump lands.
morphit-ops registerrun offline fails fast instead of hanging. The registration broadcast now times out after 15 seconds with a clear message rather than blocking on an unreachable RPC.Notes
- No database migration in this release.
- Instances install/upgrade faster now (no local frontend rebuild in the common case).
- If you operate an instance: after upgrading, your served
/verify.jsonwill match the canonical on-chain hashes, so the build-integrity check passes.
Downloads
-
Source code (ZIP)
2 downloads
-
Source code (TAR.GZ)
1 download