Make the wake digest/HMAC suite RUN and PASS in the real Woodpecker CI
runner (Alpine/musl, root), then hard-require the HMAC legs in CI.
Root cause of the runner-only D1/D4/D5/D6 failures (a masked-local run
passed, so it was runner-specific): a TOOLCHAIN divergence, not locale or
root. digest.sh's _scrub_ctrl used GNU-sed `\xNN` hex-escape byte matching.
The CI runner is node:24-alpine, whose sed is BusyBox — BusyBox sed REJECTS
a `\xNN` character range ("bad regex ... Invalid character range"), aborting
the whole scrub sed and silently VOIDING the scrub. Every scrubbed value
collapsed to empty, cascading into D1 (blank locators), D4 (no scrub/redact,
SHA blanked), D5 (blank agent prefix), D6 (blank [digest] class). Confirmed
by reproducing the exact 9-assertion failure in the ci-base image as root.
Fix (at the correct layer — a wake digest must render identically on any
runner):
- digest.sh _scrub_ctrl: patterns are now LITERAL bytes (printf %b), matching
byte-identically under GNU sed (glibc dev) and BusyBox sed (Alpine CI).
Verified identical output on both. Contract preserved: two-tier trust,
exit-4 hard-locator FAIL-LOUD, secret-scrub, and 40-hex SHA preservation
all unchanged — deterministic, not weakened.
- test-wake-digest-hmac.sh D4: replaced PCRE `grep -qP` (BusyBox grep has no
-P; the `&&` silently skipped the check in CI) with portable literal-byte
`grep -E` ranges (two disjoint bidi/zero-width ranges, excluding legit
U+2014 em-dash).
CI enablement:
- Dockerfile.ci + .woodpecker/ci.yml test step: add openssl (the non-circular
HMAC signer) so H1/H2, beacon B12, install I8 can run. The apk add in the
test step covers PR pipelines before ci-base rebuilds.
- Flip the 3 openssl skip-guards (digest whole-file, beacon B12, install I8)
to HARD-REQUIRE openssl when CI is set (Woodpecker CI=woodpecker) and FAIL
loud if absent; KEEP the skip for openssl-less local dev.
- manifest.txt: wake 0.6.2 -> 0.6.3 (digest.sh scrub portability; precedent).
Red-first verified in the ci-base container (root): D4 catches a broken
redaction, H1 catches a tamper that doesn't break the MAC, B12/I8 catch a
corrupted signer.
Closes#912
Part of #892
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_0158NZqN2n2ymKFeJAZ4GUCb
Publish pipeline:
- Add publish-npm step to .woodpecker/publish.yml — publishes all
@mosaic/* packages to Gitea npm registry on main push/tag
- Requires gitea_npm_token Woodpecker secret (package:write scope)
- publish-npm runs after build, parallel with Docker image builds
- pnpm publish resolves workspace:* to concrete versions automatically
Package configuration:
- All 20 packages versioned at 0.0.1-alpha.1
- publishConfig added to all packages (Gitea registry, public access)
- files field added to all packages (ship only dist/)
- @mosaic/forge includes pipeline/ assets in published package
Meta package (@mosaic/mosaic):
- Now depends on @mosaic/forge, @mosaic/macp, @mosaic/prdy,
@mosaic/quality-rails, @mosaic/types
- npm install @mosaic/mosaic pulls in the standalone framework
Build fixes:
- Fix forge and macp tsconfig rootDir: '.' -> 'src' so dist/index.js
resolves correctly (was dist/src/index.js)
- Exclude __tests__ and vitest.config from build includes
- Clean stale build artifacts from old rootDir config
Required Woodpecker secret:
woodpecker secret add mosaic/mosaic-stack \
--name gitea_npm_token --value '<token>' \
--event push,manual,tag
Each step was re-running pnpm install independently, and all quality
steps (typecheck, lint, format, test) ran in parallel. On merge commits
with more accumulated code this pushed the CI runner over its memory
limit (exit code 254 = OOM kill).
Fix:
- install once, share node_modules via Woodpecker workspace volume
- sequential execution: install → typecheck → lint → format → test → build
- corepack enable in each step (fresh container) but no redundant install
