Subscription coverage is live in production end to end.
What shipped
- New coverage forecast for subscription minting: status, current balance, mint price, raw mint capacity, allocated forecast mints, consecutive fully funded intended drops, funded-through / next-unfunded drops, and minimum / recommended top-up guidance.
- User-facing coverage hero and forecast details on profile Subscriptions, with responsive desktop/mobile layouts and explicit neutral states.
- Coverage status is anchored to consecutive fully funded intended drops beginning with the immediate next intended drop. A later unfunded drop does not downgrade otherwise healthy current coverage. Partial current coverage is ACTION_REQUIRED; users with no current eligibility get NO_CURRENT_ELIGIBILITY.
- Ambiguous funded_mints was replaced with mint_capacity = floor(balance / mint_price) and allocated_mints = mints assigned to forecasted intended drops.
- Demonstrated intent is derived from normalized balance, mode, top-up, subscription, final, and redeemed data. Logs are not parsed as intent.
- Coverage notifications and push delivery are reconciled asynchronously and routed only when the consolidation resolves to exactly one canonical profile--never guessed or broadcast.
Implementation
Backend PR #1839: https://github.com/6529-Collections/6529seize-backend/pull/1839
Reviewed head: 346592abaee02e65256eb88f34cfbe0e5f8ca388
Production merge/composed SHA: 50fb69f8777c432571a2d3c967ce9352d6aa6a52
Frontend PR #3476: https://github.com/6529-Collections/6529seize-frontend/pull/3476
Reviewed head: 8857bbbd2a9fb8057f513849369e68a6027ab8c9
Production merge/composed SHA: 4fa1cb94ca1029dd42d54e15551e5d5e4cfd12fd
Release evidence
Staging-qualified exact manifest: 06accb27-727d-401e-a06f-ee41f2f822b9
Production release train: 97dd7b76-4fbd-4909-97d6-69a28d19819e
Final production manifest: 5385e503-a8a4-464a-bc54-d29fc94afd97
The seven backend units deployed in this schema-safe DAG:
1. dbMigrationsLoop
2. api, pushNotificationsHandler, subscriptionsTopUpLoop, transactionsProcessingLoop, ownersBalancesLoop (each explicitly depends on dbMigrationsLoop)
3. subscriptionsDaily (depends on dbMigrationsLoop and all five stage-2 units; it now also owns the reconciliation Lambda)
4. frontend (depends on the completed backend release)
Production runs
- dbMigrationsLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/30237352994
- api: https://github.com/6529-Collections/6529seize-backend/actions/runs/30237492109
- ownersBalancesLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/30237491032
- transactionsProcessingLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/30237491079
- subscriptionsTopUpLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/30237491049
- pushNotificationsHandler retry: https://github.com/6529-Collections/6529seize-backend/actions/runs/30238003946
- subscriptionsDaily + reconciliation: https://github.com/6529-Collections/6529seize-backend/actions/runs/30238122274
- frontend: https://github.com/6529-Collections/6529seize-frontend/actions/runs/30238286949
- production-safe exact-manifest E2E: https://github.com/6529-Collections/6529seize-frontend/actions/runs/30238532935
Validation
- Release Bus marked both candidates and the train PRODUCTION_DEPLOYED; live controls remain RUNNING.
- The production frontend announces exact SHA 4fa1cb94ca1029dd42d54e15551e5d5e4cfd12fd and recovered to Ready/Green with two healthy instances.
- Production seizeAPI, subscriptionsDaily, and subscriptionCoverageReconciliationLoop all advertise backend SHA 50fb69f8777c432571a2d3c967ce9352d6aa6a52 and report Active/Successful.
- Reconciliation is packaged inside the existing allowlisted subscriptionsDaily stack, with reserved concurrency 1 and enabled 1-minute dirty-work plus hourly sweep schedules. Live invocations are completing cleanly.
- Production profile lookup resolved punk6529bot to exactly one canonical consolidation. The live coverage endpoint returned the complete contract and the correct NOT_SET_UP neutral state.
- Independent read-only UI smoke passed at desktop and 390x844: coverage hero, settings, and upcoming drop schedule render; there is no horizontal overflow and no console error.
Recovery note
The first pushNotificationsHandler run stalled in Release Bus authorization while the API was concurrently being replaced. It was cancelled before AWS credentials were issued or deployment began. The bus correctly classified it as infrastructure, preserved all successful sibling evidence, retried only that operation, and the retry passed. No rollback or production data repair was needed.
Remaining bounded risk: the public punk6529bot profile is currently NOT_SET_UP, so the independent production browser smoke exercised the neutral path rather than a live funded/action-required wallet. Exact-manifest staging and production E2E covered the behavioral status matrix and authenticated release contract.
Wave
Follow The Repo
Generate Summary
Choose a timeframe to see summary coverage.
Drops
-
-
https://media1.giphy.com/media/v1.Y2lkPTQ2YzEwZWNja2Vxa2QwanA3Y3B3dmZnbGtxemlvb2hsZGUwNWg5bjk4aW4zOGRrZSZlcD12MV9naWZzX3RyZW5kaW5nJmN0PWc/zIatAEDb9NwwAd3cYE/giphy.gif
-
Deployment overview: 4.70.1 (production SHA c5b225d9b59508e8b192009cc968e3cd7aa17adf) What shipped 1. AWS Device Farm mobile QA regime (PR https://github.com/6529-Collections/6529seize-frontend/pull/3091, follow-ups #3102/#3105/#3109) - The 6529 mobile app is a thin Capacitor shell (6529-core-mobile) whose WebView loads https://6529.io, so mobile QA now covers both halves: the deployed frontend on real device browsers, and the native shell contract itself. - New workflow .github/workflows/device-farm-qa.yml (weekly Mondays 04:00 UTC + on-demand): mobile-web Appium smoke on physical Android Chrome + iPhone Safari devices; a native Android pack that builds a debug APK of the shell in CI and validates launch, WebView boot of 6529.io, and mobile6529:// deep links; plus Device Farm built-in fuzz (2,500 events, fixed seed) for crash detection. - Suites in tests/device-farm (self-contained npm package, npm pack + bundleDependencies); ops/scripts/devicefarm-bootstrap.sh provisions the Device Farm project/pools idempotently; regime doc at ops/docs/developer/device-farm-qa.md (packs, IAM policy, cost model, triage table). - AWS fully provisioned and live: project 6529-mobile-qa (us-west-2), rule-based device pools, dedicated scoped IAM user devicefarm-ci, repo secrets set. - deploy-6529 gains an optional non-gating post-production dispatch of the real-device pack for mobile-touching releases. 2. Site-wide title fixes (PRs https://github.com/6529-Collections/6529seize-frontend/pull/3089, https://github.com/6529-Collections/6529seize-frontend/pull/3077, https://github.com/6529-Collections/6529seize-frontend/pull/3088) - ~50 routes' generateMetadata titles aligned with the client-intended titles (Network, Open Data, Tools, Collections, Brain, About families), fixing the post-Next-16 streamed-metadata clobber; dynamic titles (unread counts, NFT names) now re-assert via an ownership-aware observer. 3. Cleanups (PR https://github.com/6529-Collections/6529seize-frontend/pull/3090; #3073 staged earlier) - any_casts baseline lowered to 4 (repo-health workstream E), delegation read path typed. Merges and deploys - PR #3091 merged to main at 862110454; production deployed from main c5b225d9b: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28747257689 (GET /api/version verified). - Staging validations on 1a-staging: a58b3e902 (run 28745808496) and add58a80c (full candidate, run 28746932642), both green, smoke 14/14 each (desktop + mobile chromium). - Production smoke after deploy: 14/14; /api/version matches c5b225d9b. Real-device evidence (first-ever Device Farm runs for this repo) - Final run https://github.com/6529-Collections/6529seize-frontend/actions/runs/28751172726: mobile-web smoke PASSED on physical Google Pixel 8 (Android 14), Pixel 9 Pro (Android 15), and Apple iPhone 15 Pro (iOS 17.3.1) against production. ~2 device-minutes per Android job at $0.17/min — the weekly pack costs roughly $15/week. - Getting there took three live-run iterations, all now documented in the repo: Device Farm rejects the {{amazon_linux_2}}-style host tokens shown in AWS docs (TEST_SPEC_INVALID_YAML_FILE); the action's cleanup needs devicefarm:DeleteUpload; the repo's "selected actions" allowlist needed exact-SHA patterns for the SHA-pinned actions (CodeQL demands pins, the allowlist only had @v1 — startup_failure with zero jobs until reconciled); and Safari's WebDriver on real iPhones both returns from url() before navigating and occasionally swallows a navigation outright — the suite now waits for pathname + content and retries a dropped navigation. Incidents and gates hit - One production dispatch (run 28746615360) was correctly refused by the "main did not advance" preflight after PRs #3077/#3090 merged mid-deploy; the release was re-staged (add58a80c), re-validated, and redeployed cleanly. An earlier dispatch for #3089 (run 28746183253) was likewise cancelled when #3091 advanced main — the deployment-bus preflight working as designed under concurrent release threads. Known follow-ups - MOBILE_REPO_TOKEN secret (fine-grained read-only PAT for 6529-core-mobile) needs a human to mint it; until then the native Android pack self-skips with a notice while the web pack runs fully. - Optional MOBILE_GOOGLE_SERVICES_JSON secret for Firebase push in QA shell builds. - Native iOS on real devices deferred (needs Apple signing infra); iPhone coverage currently via real-device Safari. Release note: 4.70.1 posted to 6529 Releases (drop #1166017).
-
Deploy overview — 4.71.0 title-architecture release (2026-07-05) What shipped - The Next 16 title-clobbering bug is fixed end to end. Since the upgrade, the App Router's streamed metadata re-inserted the SSR <title> after hydration, so any page whose client useSetTitle differed from its generateMetadata title showed the bare SSR string (traced with a MutationObserver: SSR title at ~254ms, client title at ~644ms, streamed metadata re-clobber at ~660ms, sticks). - Fix is two-part, merged in dependency order: 1. https://github.com/6529-Collections/6529seize-frontend/pull/3089 (merge dd4a969ee) — generateMetadata now emits the final client-intended title on every static route (~35 routes: Open Data/Network/Tools/Collections/About/Brain section suffixes), about-section titles composed from the same i18n keys as the client (plus a documentTitles.memes override killing "About The Memes | About"), NextGen trailing-space title fixed, and all Playwright/jest title expectations moved to the canonical strings (several asserted stale "| 6529.io" titles that matched neither side). 2. https://github.com/6529-Collections/6529seize-frontend/pull/3077 (merge c5b225d9b) — TitleContext ownership + MutationObserver re-assert so explicitly claimed titles (mint page enrichment, wave unread counts, notification counts) survive metadata commits; route-default placeholders never fight server metadata. Reopened after an earlier same-day closure: the closure's objection (flipping static titles / breaking smoke packs) was ordering-dependent and #3089 resolves it. Hardened before merge with an SPA-navigation staleness guard (stale observers bail once location moves; regression test fails without the guard). - https://github.com/6529-Collections/6529seize-frontend/pull/3088 (merge 96b2718d7) — media pack accepts both mint-title forms as gate hygiene (duplicate #3097 closed as superseded). - Riders on the same production candidate (merged by parallel lanes): #3090 delegation read-path typing (no behavior change), #3091 AWS Device Farm QA workflow (CI-only). Deploys - Staging (static half) 1a-staging@9981e5e: run 28746193374 → Staging E2E run 28746481103: ALL 12 packs green — first fully green run of that gate since it was introduced. - Staging (full set) 1a-staging@add58a80c: run 28746932642 → Staging E2E run 28747212535: ALL 12 packs green. - Production main@c5b225d9b: run 28747257689 succeeded (EB healthy, version label verified); /api/version serves c5b225d9b59508e8b192009cc968e3cd7aa17adf. Validation evidence - SSR title sweep, 46 routes: 46/46 canonical on staging (both cycles) and production post-deploy. - MutationObserver hydration probes: on staging, static routes render one stable title from first paint; /the-memes/mint enriches to "Mint #517 | artist manifesto | The Memes" at ~1.6s, streamed metadata clobbers at ~1.84s, the observer re-asserts within 1ms and it holds. Re-run on production post-deploy: the original bug route /open-data/network-metrics shows a single stable "Consolidated Network Metrics | Open Data" from t=66ms (previously three-way flip settling on the bare SSR title), and the mint page re-assert holds at t=877ms. Production media pack 5/5. - Jest: 225 suites / 1188 tests green at the release head; 182 suites re-verified at the #3077 merge head; typecheck:ci green. - Local pages+home Playwright pack: 21 passed on the release branch vs 12 passed on main baseline (the 9 fixed are all title assertions; remaining failures fail identically on main and are env-shaped: api.6529.io 429s against dev serve, data waits). Incidents / notes - PR CI flake: "keeps reviewbot admin protected without operator auth" failed once on #3089's run; local repro traced to a poisoned .next dev cache (branch switching), clean-cache run 7/7 green, CI rerun green. No app change. - Lane coordination: merges were deliberately held while the 4.70.0 (touch-first) production deploy was in flight to avoid failing its late origin/main equality check; two agent lanes then converged on byte-identical staging syncs and a single production dispatch for c5b225d9b. - Known follow-ups: none blocking. The tolerant mint-title regex stays as gate hygiene even though enrichment is now deterministic.
-
🚀 Backend deploy — Lambda runtime bump to nodejs22.x **What shipped** PR #1714 (https://github.com/6529-Collections/6529seize-backend/pull/1714), merge 144347a9 to main. Follow-up to the #1712 dependency cleanup: the updated AWS SDK v3 logged NodeVersionSupportWarning on every mediaResizerLoop invocation (SDK releases after early Jan 2027 require Node >= 22). **Audit result**: 53 serverless-managed loops and seizeAPI were already on nodejs22.x. Three stragglers found and fixed: - mediaResizerLoop — was nodejs18.x; runtime lives only in AWS config (deployed via update-function-code, no serverless.yaml), now pinned via `--runtime nodejs22.x` in the deploy workflow (generate-deploy-config.mjs + regenerated deploy.yml) - nextgenMediaProxyInterceptor (Lambda@Edge) — was nodejs20.x; runtime pin added to its deploy.sh before publish-version - dropMediaIngestStorage — serverless.yaml bumped nodejs20.x → nodejs22.x (resources-only stack, cosmetic) **Deploys (in order)** 1. staging dropMediaIngestStorage from 1a-staging b29bb035 — success, no-op stack change as expected: https://github.com/6529-Collections/6529seize-backend/actions/runs/28746600446 2. prod mediaResizerLoop from main 144347a9 — success, function now Runtime nodejs22.x / Active: https://github.com/6529-Collections/6529seize-backend/actions/runs/28746670306 3. prod nextgenMediaProxyInterceptor from main 144347a9 — success, published Version 71 on nodejs22.x, CloudFront /mainnet/metadata/* + /testnet/metadata/* rewired to :71: https://github.com/6529-Collections/6529seize-backend/actions/runs/28746729719 **Validation** - Both lambdas built end-to-end in CI and locally before merge; all PR checks + review bots green (6529bot: "Good to merge", no findings in any lane) - mediaResizerLoop: config verified nodejs22.x, post-deploy invocations clean (no NodeVersionSupportWarning, no errors) - nextgenMediaProxyInterceptor: live smoke on media.generator.6529.io/mainnet/metadata — real token metadata 200, pending-token placeholder path 200 through Edge v71 - main synced back into 1a-staging (948c809b) **Follow-ups / risk** - None outstanding. Runtime bumps are config-only, no code changes; rollback is a redeploy with the previous runtime. - Remaining Node-runtime item repo-wide: none below nodejs22.x anywhere now.
-
Deployment overview — web 4.70.0 (touch-first device detection) What shipped (user-facing): - Hybrid touch laptops (Surface-class: touchscreen + touchpad/mouse) now always get the desktop interface. Root cause: hasTouchScreen was raw maxTouchPoints>0 gating SmallScreenLayout below 1024px, plus ~10 standalone (pointer: coarse)/maxTouchPoints checks suppressing tooltips and swapping hover actions for long-press/bottom-sheets. Detection is now centralized in helpers/touch-first.helpers.ts (touch AND no fine pointer AND no hover, with a phone-UA override so phones keep mobile UI even with a paired mouse; client hints keep tablets+trackpad on desktop). Reactive to capability changes (mouse plug/unplug, posture flips) via media-query subscriptions. - NextGen landing/about horizontal overflow fixed (PR #3080), Network Metrics downloads RSC crash + title fixed (PR #3081), collection page titles aligned (PR #3082). PRs and SHAs: - PR #3075 (touch-first, 28 files) — merge 7d59e7322 = production deployed SHA - PR #3086 (readonly E2E guard: alchemy read-proxy allowlist) — merge 7ba49b03e - Riders from the repo-health campaign: #3080, #3081, #3082, #3084 (docs), #3087/#3079 (any-cast cleanup), #3056 (TODO triage) - Production run: github.com/6529-Collections/6529seize-frontend/actions/runs/28745717051 (Web Deploy - PROD, main @ 7d59e7322, success) - Staging runs: deploy 28742783662/28743892863/28744710645; Staging E2E 28745223584 (success — first fully green run of the new 12-pack workflow) Validation: - Staging: full Staging E2E green (12 packs, ~160 read-only tests, desktop+mobile chromium) on the exact production candidate; release-specific device-profile checks 15/15 across three staging builds (Surface hybrid 950px/1600px -> desktop; plain desktop control -> desktop; phone -> small-screen; phone+paired-mouse -> small-screen; zero page errors). - Production: device profiles 5/5 on 6529.io post-deploy; nextgen-collections + media-mint-detail + home packs 17/17 (both nextgen and mint packs had been failing on the pre-release production baseline — this release made prod strictly greener). - Jest: ~100 suites over the touched surfaces; new decision-matrix tests for LayoutWrapper and the touch-first helpers. Failed gates worked through: - Debt ratchet false positive (": any" inside a comment) — reworded. - 6529bot general lane "Needs changes" (phone+mouse / stylus-hover misclassification) — fixed at the source with the phone-UA override + client-hints tablet handling; follow-up verdict "No new findings". - CodeRabbit: 3 inline findings (stale one-shot detection, capability divergence, Android-tablet UA fallback) — all fixed; Android fallback narrowed to Android.*Mobile. - Responsiveness lane: single native-sim flake on /the-memes (body.capacitor-native timing in the browser-level Capacitor simulation) — evidenced as unrelated (CapacitorSetup untouched; local contract tests pass on both capacitor sim projects; web-desktop/web-mobile clean). - Staging E2E red on first two runs: xtdh readonly-guard block (fixed via PR #3086), nextgen overflow + mint-title (proven pre-existing on the production baseline with byte-identical errors; both since fixed on main by #3080/#3082). Known follow-ups: - Convertibles that keep the touchpad enumerated in folded/tablet posture stay on desktop UI (benign; capability-driven). - 6529bot responsiveness lane may want a rerun/verdict refresh on the merged head; its blocking finding was evidenced as simulation flake.
-
Deployment overview — web 4.69.1 (2026-07-05): emoji picker search autofocus What shipped - The emoji picker's search input now auto-focuses on open at every entry point (drop reaction picker + composer, desktop portal and mobile dialog), removing the extra click keyboard users previously needed. Note: PR #3066 already rode along in the 4.69.0 production build (main 824c57e25, Web Deploy - PROD run 28741978540) — its merge c956b71b5 was an ancestor — but was not called out in the 4.69.0 notes. Documenting it here as 4.69.1; no additional prod deploy was needed. PRs and SHAs - PR: https://github.com/6529-Collections/6529seize-frontend/pull/3066 (merged to main as c956b71b5) - Shipped in production build main 824c57e25 (Web Deploy - PROD run https://github.com/6529-Collections/6529seize-frontend/actions/runs/28741978540) - Staging: validated on 1a-staging (merge fef038a84), deploy run https://github.com/6529-Collections/6529seize-frontend/actions/runs/28741073885 green; staging E2E run https://github.com/6529-Collections/6529seize-frontend/actions/runs/28741346113 showed only the pre-existing Network Metrics title spec failure (unrelated), consistent with the 4.69.0 exception record. Validation - In-PR: 19/19 unit tests across both picker suites; lint clean; live Playwright confirmed the search input is the active element immediately on open (desktop, mobile emulation, composer). - Production: 6529.io healthy (HTTP 200); change present in the shipped 824c57e25 bundle. Follow-ups - Stale Network Metrics E2E title assertion (shared with the 4.69.0 follow-up).
-
Backend dependency-security release — Dependabot cleanup (PR #1712) WHAT SHIPPED Cleared 56 of the repo's 58 open Dependabot alerts (the only CRITICAL, all 9 HIGH, and every MEDIUM/LOW that isn't chained to the deferred serverless v3 upgrade). Dependency-only change: 41 lambda package.json/package-lock.json files, zero source-code changes. PR: https://github.com/6529-Collections/6529seize-backend/pull/1712 (merged to main as d6119d11) The four batches: 1. esbuild -> ^0.28.1 across 37 lambda dirs (33 alerted + 4 with the same stale 0.27.2 locks not yet flagged). Many lockfiles were stale vs their manifests (locks still resolving esbuild 0.17.x); regenerating them with the CI-pinned npm@10.9.8 is the real fix. Build-time tool only — no runtime behavior change. 2. cloudwatchAlarmsToDiscordLoop: discord.js -> 14.26.4 (fixes ws 8.21.0, lodash 4.18.1) plus an npm override lifting undici to 6.27.0 (discord.js exact-pins 6.24.1). 14 alerts cleared. 3. mediaResizerLoop: @aws-sdk/client-s3 + lib-storage -> 3.1079.0, which removes fast-xml-parser from the dependency tree entirely — this was the repo's only CRITICAL alert (fast-xml-parser 5.3.4). 6 alerts cleared. 4. dbMigrationsLoop: lockfile-only lodash -> 4.18.1 (was 4.17.21 via db-migrate -> winston/async). 3 alerts cleared. Intentionally NOT touched: the 2 root package-lock.json alerts (file-type via @serverless/utils, aws-sdk v2 via serverless) — chained to the serverless v3 upgrade, which stays deferred by owner decision for licensing reasons. REVIEW CI green (backend+API build, SonarCloud, Snyk, DCO). CodeRabbit: no findings. All 6529bot lanes clean on head d24688ad (security / general / auth-api / deploy-actions / db-lambda / media-external); the GLM swarm advisory items were each verified and answered on the PR (lodash 4.18.1 is the registry latest; the s3Uploader async 3.2.4->0.2.10 change is fluent-ffmpeg 2.1.3 following its own pinned range; Lambda runtime nodejs22.x satisfies undici 6.27 engines). DEPLOYMENTS Only 3 of the 41 touched lambdas carry runtime dependency changes and were redeployed; the esbuild-only lambdas need no redeploy (the fix lands with their next regular deploy). Staging (from 1a-staging @ 78f4eb55): - dbMigrationsLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/28743872586 — validated: post-deploy invocation connected to DB/Redis, "No migrations to run", finished in 18.6s. (cloudwatchAlarmsToDiscordLoop and mediaResizerLoop are prod-only services — no staging lane exists.) Production (from main @ d6119d11): - dbMigrationsLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/28743988554 — validated: post-deploy invocation clean, "No migrations to run", 8.6s. - cloudwatchAlarmsToDiscordLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/28744057638 — function Active/Successful (nodejs22.x). Event-driven (fires on CloudWatch alarms); no synthetic invoke was run to avoid a test post into Discord ops channels. - mediaResizerLoop: https://github.com/6529-Collections/6529seize-backend/actions/runs/28744058228 — function Active/Successful. main merged back into 1a-staging (8b59f277) to keep staging current. FOLLOW-UPS / RESIDUAL RISK - The next real CloudWatch alarm will be the first live exercise of the discord.js 14.26.4 + undici 6.27 path; worth a glance at the ops Discord channel when one fires. - 2 Dependabot alerts remain open by design (serverless v3 chain, deferred). - Dependabot will rescan and close the 56 cleared alerts after the merge; if any linger, the lockfile sweep evidence is in the PR.
-
Deployment overview — web 4.69.0 (2026-07-05) What shipped - "Copy text" action in the wave/DM drop long-press menu on touch devices, next to "Copy link", in both the authenticated and guest action sheets. Payload matches desktop Ctrl/Cmd+C copy exactly (author + time heading, message text, reply/quote context, attachment URLs). - The pure formatting logic (~650 lines) was extracted from useWaveDropsClipboard into helpers/waves/drop-clipboard.helpers.ts and is now shared by desktop copy and the new mobile action; desktop behavior unchanged. - Copy text / Copy link / Copied! labels are message-backed (waves.drop.actions.* with fr-FR/es-ES/de-DE translations); the Copied! confirmation is announced via a visually-hidden role="status" region. PRs and SHAs - PR: https://github.com/6529-Collections/6529seize-frontend/pull/3072 (merged to main as 824c57e25) - Production deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28741978540 (Web Deploy - PROD from main 824c57e25, version verification green) - Staging deploys validated: 14f0714e9 then 005e0d148 (after the a11y fix below); final staging E2E run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28741346113 - main synced back into 1a-staging after the production merge (f059f49f9) Validation performed - Unit: 29 jest tests across the copy components and drop-clipboard helpers, plus tsc/eslint/knip clean. - Staging: full read-only Playwright packs (~160 tests, desktop + mobile Chromium) via the new post-deploy Staging E2E workflow, plus a release-specific device-emulated check (long-press opens the sheet, Copy text has a proper accessible name, clipboard payload verified end to end against a real staging wave). - Production: routes /, /the-memes, /waves, /network load clean under mobile emulation; no 5xx, no critical console errors; the shipped bundle contains the new feature (message-key probe). Incidents and gates worked during the release - Staging validation caught a real WCAG regression in the first staging build: role="status" on the button label removed the accessible name of both copy buttons in Chromium. Fixed (2e5e65852) by moving the live region out of the label, restaged, and re-verified against the deployed build. This never reached production. - The new Staging E2E suite (PR #3071, first completed run today) reported 12 hard failures. All were diagnosed with page-level probes and a production comparison run: 7 fail identically on https://6529.io (NextGen specs expecting a "Featured" heading the site doesn't render; Network Metrics title assertion) = pre-existing spec bugs; 1 is staging-dataset-only (The Memes mint title expects an active mint; staging has none); the rest passed on retry. Release-captain exception recorded; production promotion proceeded on green release-critical surfaces (150+ tests passed including all waves/drops/social/home packs). Follow-ups - Fix the staging E2E spec assertions that fail against both environments (task chip queued; details include exact spec lines and repro commands). - Optional polish deferred from review threads: locale-aware clipboard timestamp formatting and copy-failure toast feedback for both copy actions.
-
Deployment overview - web release 4.68.0 (repo-health campaign train) Production is on main 306e55a3d (staging-validated via 1a-staging 8fced9f36). Prod run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28738949386 (green, EB healthy). This train carried 28 PRs merged since the previous deploy (97cac975a). Highlights: - Redux fully removed: editSlice/groupSlice replaced by EditingDropContext/ActiveGroupContext, three deps dropped (#3047), with new browser E2E coverage for the drop-edit and group-filter journeys (#3070) - Profile i18n work recovered from a collapsed stacked-PR tail and re-landed (#3034, #3035) - Layout unification: stray src/ tree folded into the root layout (#3048), router hygiene (#3050), 55 legacy museum/about pages rebuilt on shared chrome (#3054, #3055) - Type safety: any-casts 358 -> 65, NextGen contract layer typed end-to-end (#3052) - CI hardening: structural-debt ratchet as a required check (#3032), coverage floor on main pushes (#3033, #3038), push-time secret scanning (#3045), nightly branch janitor (#3036, armed), and a new post-deploy Staging E2E workflow running 12 Playwright packs automatically (#3071) - Bootstrap exit verified at zero surface with reintroduction guards (#3042, #3046, #3049) Validation evidence: - Local: full Jest suite (1,959 suites / 11,047 tests) green on the release SHA; new edit-drop sandbox E2E 2/2; sandbox packs green - Staging (8fced9f36): 12-pack battery green across smoke/surfaces/social/groups/delegation/admin-guards/content/deep-links/search/media; collections + network packs have failures that reproduce byte-identically on the pre-train production build (pre-existing, tracked) - Production (post-deploy): combined readonly pack 63/70 with the exact same 7 pre-existing failures as before the deploy (before/after match), social 6/6, groups 5/5 including the new group-filter test against production data Known follow-ups: the pre-existing NextGen collections page overflow at desktop width and one network-metrics test failure now have a dedicated owner thread (root cause located: Bootstrap-era negative-margin rows missing their compensating gutters, concentrated in components/nextGen/collections), alongside the slow collection routes already on the improvement list.
-
Deployment overview - web release 4.68.0 (repo-health campaign train)
-
Frontend release deployed to production (train through 4.67.0) Production deploy from main @ 97cac975a via Web Deploy - PROD (run 28722854601). A prior prod deploy at 19:08 UTC (d9a0864f) already carried most of this train; this deploy adds the wagmi bootstrap change and brings prod level with the staging-validated set. Shipped in this train: - #3023 Remove legacy curation leaderboard filter — deletes the inner curation-group filter; top curation tabs (separate ?curation= pipeline) unaffected; also fixes a latent leaderboard-unmount cache eviction. - #2973 Add year-scoped Memes filters — Year filter + URL year param, year-scoped season dropdown, detail-page year links. - #3026 Polish About and Tools indexes — reworked onto the new About IA (#3024): Subscription Minting rename, arrow cards, removed tools icons/lead, shorter tools meta description; removed now-dead GroupedLinkIndexIcon plumbing. - #3001 Add Join journey — shipped at /join (replaced the placeholder route), help records consolidated to one join.6529 entry, palette aligned to iron/primary, copy tightened. - #3030 Render boosted standalone media URLs as media — image/video/audio/3D-model direct URLs and captioned markdown images render as media; plain webpage URLs incl .html stay on link-preview. - #3028 Slim Wagmi AppKit bootstrap — render the app as soon as the wagmi adapter exists; gate modal opening behind a ready-context; hardened with a ready-failure toast and a 15s timeout on connect-intent waits so a hung relay recovers. - Dependabot: #3004 dev-minor group, #2918 p-retry 6.2.1→8.0.0, #2919 date-fns 2.30.0→4.4.0. Closed (not shipped): #3015 (runtime-minor group — Lexical 0.14→0.46 breaks the composer build) and #2986 (runtime-patch — appkit bump pulls an incompatible @wagmi/core/tempo). Follow-ups queued: exclude lexical/@lexical from the dependabot minor group; do appkit+wagmi as a coordinated manual bump. Validation: - Staging deployed from 1a-staging @ 277614af8 (run 28721419300, success). Prod candidate main is content-identical to that staging tree except 3 repo-health markdown docs. - Staging E2E (test:e2e:staging): 20 passed. Targeted release checks (/join renders, tools arrow cards, Subscription Minting label, Memes year param): 4/4 passed. - 6 staging pack failures were classified against a production baseline: 5 are pre-existing navigation-IA test-vs-markup skew (identical failures on prod: About sidebar strict-mode locator, header-search Wave Score link ×2, View Network Stats link ×2) and 1 is environmental (home test hardcodes title "6529.io"; staging serves "6529 Staging"). None are attributable to this release. Follow-up queued to fix these tests. - Prod E2E against https://6529.io: 9/9 passed (about smoke, the-memes smoke, and 4 release checks: /join journey renders, tools arrow cards, Subscription Minting label, Memes year param). Direct route health: /join /about /tools /the-memes?year=4 all HTTP 200. Known follow-ups: /join is a functional v1 (palette aligned, but no hero/marketing section yet, and the step state machine lacks unit coverage) and will get another pass.
-
Deployed: agent discovery artifacts (PR #3029) What shipped - public/llms.txt + public/glossary.json (27 terms): build-time projections of ops/help/help-index.json, the @help6529 corpus. New scripts/sync-agent-files.cjs runs in prebuild after help-index:sync; drift between corpus and committed artifacts fails CI. - New waves.drop glossary record plus a "glossary" tag on 14 term-defining records (projection rule: kind == glossary OR tag glossary). - robots.txt agent block emitted by next-sitemap transformRobotsTxt, so it survives every build. - proxy.ts: /llms.txt and /glossary.json added to the access-gate static bypass (staging validation caught them 307-ing to /access). Refs - PR: https://github.com/6529-Collections/6529seize-frontend/pull/3029 (merge d9a0864ff) - Staging: deploy-staging runs 28714187715 and 28716355908 (production-equivalent set), files validated 200 with correct content, access gate intact - Production: build-upload-deploy-prod run 28716684468 from main @ d9a0864ff; post-deploy checks green (files 200, core routes 200, production read-only social E2E 6/6) - One staging build failure en route (run 28715997399): stale ToolsIndex/AboutIndex from the staging-side revert of the about/tools polish vs main's merged version; fixed by aligning four files with main. Unrelated to the release content. Known gaps / follow-ups - Staging gated-page Playwright smoke not run locally (needs staging access credential); validated via direct HTTP checks + production read-only E2E instead. - Corpus contract: any edit to ops/help/help-index.json must rerun "6529 run agent-files:sync" in the same PR (CI enforces it).
-
Backend deploy note - review follow-ups shipped (3 PRs) The polish items and deferrals recorded in today's backlog-clearing release are now themselves shipped. All three PRs went through full bot review (including manually triggered general/security/deploy-actions/auth-api/db-lambda lanes) to clean verdicts, then staging deploy + validation before production. PR #1702 "Polish Meme edition size floor plumbing" (merge SHA 86ff619e) - Manifold lazy-claim contract address + ABI extracted to a shared constants module (was duplicated between the floor resolver and mint announcements service). - MEMES_EDITION_SIZE_FLOOR_CAP=310 moved to constants with its rationale documented in code. - Floor refresh widened from "latest Meme only" to "latest Meme + any Meme minted in the last 30 days", so a loop outage spanning a drop self-heals instead of leaving that drop's floor stuck at backfilled supply. New unit tests cover the outage-recovery, outside-window, and missing-mint-date cases. - https://github.com/6529-Collections/6529seize-backend/pull/1702 PR #1704 "Upgrade json-2-csv to v5" (merge SHA 8812dbe2) - Clears the moderate CSV-injection advisory on json-2-csv@3 (the last root audit finding fixable without touching the serverless v3 chain; npm audit fix cannot apply it because the fix is semver-major). - v5 renamed json2csvAsync to a synchronous json2csv; all seven call sites updated (API CSV downloads, dbDumpsDaily, ratings export, rememesLoop, royalties, subscriptionsDaily, tdhLoop uploads). Output verified byte-identical to v3 for the shapes used, including the empty-array edge case; full build + test suite green. - https://github.com/6529-Collections/6529seize-backend/pull/1704 PR #1700 "Activate pinned npm via Corepack in PR CI" (merge SHA 9f015599) - Makes the packageManager npm@10.9.8 pin binding in PR CI rather than advisory: Corepack shims npm, installs the pinned version, and the build fails if the resolved npm does not match the pin (expected version read from package.json, hash-pin suffixes handled, missing-pin guarded with a clear error). Self-validated on its own CI run: "npm resolved to 10.9.8 (pinned 10.9.8)". - CI-only; no service deploys. - https://github.com/6529-Collections/6529seize-backend/pull/1700 Deployments - Staging (from 1a-staging, all success): nftsLoop 28712344040, mintAnnouncementsLoop 28712344682, api 28712345208, tdhLoop 28712345787; staging API validated healthy. - Production (from main, all success): nftsLoop 28712439657, mintAnnouncementsLoop 28712439923, api 28712440233, dbDumpsDaily 28712440519, rememesLoop 28712440850, royaltiesLoop 28712441171, subscriptionsDaily 28712441474, tdhLoop 28712441797. - Production validation: API healthy (200s) and a live CSV download endpoint (aggregated-activity with download_page) returned well-formed CSV through the new json-2-csv v5 path. - main synced back into 1a-staging (7883ea41). Remaining known risk (deliberately deferred) - The final 6 npm audit findings (2 low / 4 moderate) all chain to serverless v3 (via @serverless/utils, file-type, aws-sdk v2, dashboard-plugin). The fix is serverless 4, a semver-major with a commercial licensing model change - flagged for an owner decision rather than an autonomous bump. Incidents: none. All twelve deploy runs green on first attempt.
-
@[GelatoGenesis] - review requested: 5 backend performance PRs (nothing merged, nothing deployed) A full performance audit of 6529seize-backend produced 5 PRs of unambiguous, behavior-identical optimizations. Every change is either "same queries, fewer of them" or "same result, no quadratic rescan" - no schema, API-contract or output changes intended anywhere. Each PR body walks through the equivalence argument per change, and the risky rewrites are property-tested against the verbatim pre-optimization algorithms. 1) PR #1699 - tdhLoop: deletes a dead copy+filter+sort block that ran per meme NFT in calculateRanks (result was never read), replaces two O(n^2) wallet scans in TDH consolidation with one-pass grouping, hoists invariant rank lookups, and guards consolidateMissingWallets' dedup before its DB calls. Property test proves output identical to the old algorithm. https://github.com/6529-Collections/6529seize-backend/pull/1699 2) PR #1701 - ownersBalancesLoop: owners grouped by wallet once instead of filtered per address (the reset-mode path was quadratic in the extreme), consolidation keys resolved in 5000-address batched view queries instead of one query per address, consolidated balances computed once per unique consolidation, reset path selects only wallets. https://github.com/6529-Collections/6529seize-backend/pull/1701 3) PR #1703 - subscriptions: new batched eligibility lookup (1 MAX + 1 IN query) replaces 2 queries per subscription in both the public upcoming-meme-counts endpoint (thousands of queries per request before) and the daily job; full-table balance scans per subscription become maps; airdrop-address lookups skip unfunded subscriptions. This one touches the final-subscription flow, so it deserves the closest look. https://github.com/6529-Collections/6529seize-backend/pull/1703 4) PR #1705 - indexing loops: nextgen token traits fetched per collection instead of per token (~10k queries -> 3), nfts supply via one GROUP BY instead of a query per ERC1155 token, market-stats volumes via one GROUP BY per contract, and O(n^2) rank/count scans replaced with sort/frequency passes (property-tested). https://github.com/6529-Collections/6529seize-backend/pull/1705 5) PR #1706 - api: drop-mapping's six independent reads overlap in one Promise.all (was 3 sequential stages on the hottest endpoints); PFP S3 client reused instead of constructed per upload. https://github.com/6529-Collections/6529seize-backend/pull/1706 Status: all 5 green on CI (build, DCO, SonarCloud quality gate, Snyk), CodeRabbit findings addressed or answered, focused Jest suites added where the rewrites were subtle. Also flagged in PR notes (not changed): fetchAllNftOwners' double .where() silently drops the contract filter when both args are passed - pre-existing, needs a human call. Per instructions these stay unmerged and undeployed until you've reviewed. Suggested review order: #1699 -> #1701 -> #1703 (highest scrutiny) -> #1705 -> #1706.
-
@[GelatoGenesis] - 3 more perf PRs added to the review queue (same rules: behavior-identical, nothing merged, nothing deployed) Follow-up batch from the same audit, approved by 6529. These are the "Tier 1" items - same risk profile as the first five, with equivalence tests against the verbatim old algorithms where the rewrites are non-trivial. 6) PR #1707 - nextgen: probably the single largest win of the whole effort. Trait-score computation re-scanned all ~150k trait rows once per row (O(T^2), billions of ops per collection refresh); now linear via per-trait value-count maps. Rank calculation also groups categories once per call instead of re-filtering the full list per category. Property test proves persisted rows are identical INCLUDING array order (the six chained rank calls deliberately reorder, so order matters). https://github.com/6529-Collections/6529seize-backend/pull/1707 7) PR #1708 - consolidations: the per-address consolidation-key N+1 you'll recognize from #1701 existed in two more places (aggregatedActivityLoop, nftOwnersLoop). Now a shared, unit-tested db helper resolves all addresses in chunked batches and dedupes to one computation per unique consolidation. delegationsLoop calls both functions on every consolidation change, so it benefits too. https://github.com/6529-Collections/6529seize-backend/pull/1708 8) PR #1709 - subscriptions: consolidateSubscriptions (the key-migration path when consolidations change) queried the view per wallet part and TDH up to twice per part inside a nested loop; both lookups are now prefetched in batched queries with the decision logic untouched. Bonus hygiene: the LIKE filter is parameterized instead of string-concatenated. End-to-end test covers the migration decisions. Money-adjacent like #1703, so worth the closer look. https://github.com/6529-Collections/6529seize-backend/pull/1709 Status: all 3 green on CI (build, DCO, SonarCloud, Snyk), CodeRabbit reviewed clean (#1707/#1708 zero findings; #1709's cognitive-complexity suggestion applied). Total queue for you is now 8 PRs; suggested order for the new ones: #1707 -> #1708 -> #1709.
-
-
Backend deploy note - open PR backlog cleared (3 releases + housekeeping) All outstanding backend PRs were reviewed, updated, and either shipped or closed today. Each release went through 1a-staging deploy + validation before its production merge and deploy. Release 1 - PR #1693 "fix(api): use namespace import for Joi in drops-media routes" (merge SHA 2cc07e6a) - One-line fix: drops-media routes used a default Joi import that crashes the API at module load under ts-node local dev. Now matches the namespace import used by the other 58 route files. - Deployed: api to staging (run 28708636020) then prod (run 28708710996). Prod API validated healthy. - https://github.com/6529-Collections/6529seize-backend/pull/1693 Release 2 - PR #1658 "[codex] Fix safe backend npm audit updates" (merge SHA f1374310) - Dependency security refresh: testcontainers 11->12 (dev), esbuild 0.28, discord.js 14.26.4, overrides for ws 8.21.0 / tar 7.5.16 / uuid 11.1.1 / undici 6.x-patched, plus full lockfile regeneration and a fresh npm audit fix sweep on top of current main. - Result: root audit went from 46 vulnerabilities (13 high) to 7 low/moderate with 0 high/critical; api-serverless from 21 (3 critical) to 0. Full local build + 1770 tests passed before push; CI and all review bots green on the final head. - Deployed: api to staging (run 28709216407) then prod (run 28709288559). Prod API validated healthy. Other lambdas pick up the refreshed dependency tree on their next regular deploys. - https://github.com/6529-Collections/6529seize-backend/pull/1658 Release 3 - PR #1692 "[codex] Store Meme edition size floors for TDH rates" (merge SHA dd40047e) - While a Meme is still minting, TDH previously divided by in-flight supply, over-rewarding holders of the newest card. Now a per-NFT edition_size_floor (Manifold claim totalMax, capped at 310) is persisted and TDH hodl rates plus edition_size_rank use max(supply, floor). - Schema migration included; deployed in the documented order dbMigrationsLoop -> nftsLoop -> tdhLoop on staging (runs 28709339700, 28709413024, 28709468121) and then production (runs 28709534172, 28709607423, 28709667235). All six runs green; prod API and NFT endpoints validated healthy. Until nftsLoop's next refresh the floor equals current supply, so rollout is behavior-neutral by design. - https://github.com/6529-Collections/6529seize-backend/pull/1692 Housekeeping - PR #1698 "Enforce npm as the pinned package manager" (merge SHA bfe9b38f): new dependency-free checker script + PR CI step pins packageManager npm@10.9.8 across all 56 package.json files and fails CI on stray yarn/pnpm/bun lockfiles. CI-only; no service deploys. Iterated three review rounds with 6529bot + SonarCloud to a clean verdict. https://github.com/6529-Collections/6529seize-backend/pull/1698 - PR #1654 (Snyk ethers 6.17 bump) closed as superseded by #1658, which covers the ws advisory more completely and had no DCO issues. - PR #1511 (pnpm migration + CLI shim) closed: 640 commits behind with 16 conflicting files, frozen April lockfile, and 7 newer services never converted. #1698 delivers the enforcement goal with a fraction of the surface; workspace migration and Socket Firewall can return as separate scoped PRs if wanted. Validation and state - Every merged PR had all checks green (Build, SonarCloud, Snyk, DCO, CodeRabbit) and clean 6529bot verdicts on its final head before merge. - main -> 1a-staging synced after each production merge; staging is current with production. - Zero open PRs remain on the backend repo. Incidents: none. All staging and production deploy runs completed successfully on the first attempt.
-
Backend deploy note - GitHub Actions runtime maintenance What shipped - PR #1637 "[codex] Update GitHub Actions runtimes" merged to main (merge SHA fd4f7a6a3ee032273e107228d4e74269f3e555be): https://github.com/6529-Collections/6529seize-backend/pull/1637 - actions/checkout v3/v4 -> v6 and actions/setup-node v4 -> v6 (Node 24-ready action majors) - Deploy precondition checks moved from deprecated actions/github-script@v3 to plain bash failure steps - aws-actions/configure-aws-credentials bumped from a 2021 v1.5.10 pin to the v6.2.0 release commit SHA (e7f100cf) - Shellcheck cleanups: quoted "$GITHUB_OUTPUT", grouped "$GITHUB_ENV" env exports; deploy.yml regenerated from scripts/generate-deploy-config.mjs (regeneration idempotent) Scope - Workflow maintenance only. No application code, Lambda code, or runtime configuration changes. No production service redeploys required; the next regular deploys simply use the updated workflow. Validation - CI green on head 2c82fcb3: Build backend and API, SonarCloud (0 new issues), Snyk, DCO - Reviews: 6529bot general "Good to merge", crypto security "No security findings", follow-up "No new findings"; CodeRabbit review completed with no findings; actionlint fully clean - Staging: branch merged to 1a-staging (ee280a1a) and api deployed to staging through the NEW workflow - run https://github.com/6529-Collections/6529seize-backend/actions/runs/28705675471 (success); staging API responding normally - Post-merge: main synced back into 1a-staging (6d13faad) Incidents / follow-ups - None. Merged via admin path on maintainer direction with all status checks green.
-
https://github.com/6529-Collections/6529seize-frontend/pull/2979 @[simo] @[prxt0] need approval, removed react-bootstrap
-
i thought i can not automerge
-
-
Rank should be disabled until mint is over
-

-
yup, and tdh rate too - i’ll pick it up
-
 shows 'Unranked' until rank is finalized also TDH Rate will show 'Pending' while card doesn't have TDH cc @[gpebbles] i think you asked for this!
-
-
shouldn't this say somewhere what card in the szn it is
-

-
and then decide
-
i mean check it first, since its preview
-
Check this pls
-
https://x.com/i/status/2071647669467201657
-
i take it
-
-
@[simo] https://github.com/6529-Collections/6529seize-backend/pull/1670 need approval for this please
-
or @[punk6529] @[ragne] :pray:
-
don't you have access to merge on BE?
-
i didnt when i sent this out, i do now :pray:
-
-
-
Frontend deploy update: production is live at 41976a02b7fdbcda2477d1259e0d63a3de6ec00b. Production advanced from the previous deployed SHA 3682d3fb327a9775677cb14a4f4743075d173ece to current main 41976a02b7fdbcda2477d1259e0d63a3de6ec00b. Main user-facing/operator-facing changes in this train: - PR #2673 Mobile floating bottom dock: https://github.com/6529-Collections/6529seize-frontend/pull/2673 Merge SHA: 41976a02b7fdbcda2477d1259e0d63a3de6ec00b - PR #2843 Restore wallet auth frontend changes: https://github.com/6529-Collections/6529seize-frontend/pull/2843 Merge SHA: b6568969b68f27bc5f544cf1396e7a00bfc774e9 - Current-main delta since the prior prod deploy also included PR #2752, #2411, #2861, #2860, #2680, #2662, #2660, #2658, #2661, #2763, #2666, #2836, and #2853. Deployment evidence: - Staging branch/deploy SHA: 13eba540dbcc6685799099e7fad36bcf1e84fecf - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28067710525 - Staging deploy, HTTP version check, durable version evidence upload, terminal deployment status, and release report all passed. - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/28068586559 - Production build, package, bundle-size check, main-advanced guard, S3 asset upload, EB update, EB health/readiness, deployment validation, /api/version check, durable version evidence upload, terminal deployment status, and release report all passed. Validation: - Local focused PR #2673 validation before merge: related Jest suites, lint:changed, typecheck:changed, codex-diff-check, and browser smoke for the routes flagged by the responsiveness runner. - Deployed staging Playwright pack against https://staging.6529.io: 24 passed, 22 skipped across desktop/mobile Chromium. - Production-safe Playwright pack against https://6529.io: 24 passed, 22 skipped across desktop/mobile Chromium. - Independent production /api/version verification matched 41976a02b7fdbcda2477d1259e0d63a3de6ec00b. Review notes and follow-ups: - PR #2673 had a non-blocking responsiveness-runner first-render flake on GitHub Actions; this was documented on the PR with local repro evidence, and staging/production deployed smoke is now green. - No backend deploy was involved. - Remaining PR queue audit: #2455 should be rescued into a fresh current-main PR rather than merged as-is; #2134 is conflicted and needs a CSV product-rule decision; #2022 is valuable but needs a fresh reimplementation on current main.
-
@[punk6529] @[simo] @[ragne] https://github.com/6529-Collections/6529seize-frontend/pull/2849 Test fixes on main this should be all of them 🤞
-
Testing hardening deploy update: PR #2848 is live in production. Shipped: - PR #2848: native and Electron simulated shell read-only E2E coverage https://github.com/6529-Collections/6529seize-frontend/pull/2848 What changed: - Added a focused read-only Playwright pack for native-adjacent shell behavior. - Covered Capacitor runtime signals, iOS non-US Open Data subscription hiding, iOS US and Android subscription visibility, Capacitor app-wallet empty state, Electron app-wallet unsupported behavior, and Electron share-modal handoff suppression. - Kept `test:e2e:native-sim` on an explicit native-surface allowlist so future specs join that budget intentionally. - Repaired relevant share/app-wallet Jest harness drift so those tests run cleanly again. Merge and deploy: - PR merge SHA: fe4af27e79e18dae2ca793e552c619fd3a6bf5de - Staging deploy SHA: 2b8799ad1b6fe418ab50367a42715c58132b5ada - Staging workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27985961939 - Production workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27986836935 - Production deployed SHA: fe4af27e79e18dae2ca793e552c619fd3a6bf5de Validation: - Local PR validation included native-shell readonly, native-sim, surface-matrix, Playwright typecheck, changed lint/typecheck, targeted Jest, risk/security checks, secret scan, workflow security scan, and diff whitespace check. - PR CI passed: DCO, CodeQL, App PR checks, Plan risk/security checks, SonarCloud, Snyk, dependency governance, and CodeRabbit. - 6529bot final head: Opus general Good to merge; WCAG/i18n/security no findings; GLM swarm advisory found no high-confidence regression; responsiveness passed final-head reruns. - One contradictory responsiveness run on the same head produced black desktop screenshots; two final-head reruns passed, local surface-matrix passed, and a verifier concluded it was reviewbot runner/dev-server flake rather than a frontend regression. - Staging `/api/version` matched 2b8799ad1b6fe418ab50367a42715c58132b5ada. - `test:e2e:staging`: 24 passed / 20 skipped after loading the stored staging access credential. - Production `/api/version` matched fe4af27e79e18dae2ca793e552c619fd3a6bf5de. - `test:e2e:production:readonly`: 65/65 passed. Known boundaries: - This is browser-simulation evidence for native/Electron branching. It is not real packaged iOS, Android, or Electron app evidence. - No backend deploy, data migration, auth change, wallet behavior change, or user-facing runtime migration was included.
-
Testing hardening deploy update: PR #2847 is live in production. Shipped: - PR #2847: production read-only E2E flake hardening https://github.com/6529-Collections/6529seize-frontend/pull/2847 - Main / production SHA: 0c55e0c628541fb2ac695d87f871568848e7c057 - Staging deploy SHA: 0fdc39ecddebc8f6730da82ff9838284924e9492 - Staging workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27976114194 - Production workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27977449591 Validation: - Required PR CI passed: CodeQL, DCO, app PR checks, SonarCloud, Snyk, and CodeRabbit status. - 6529bot on final head: general good to merge; WCAG, i18n, and security had no findings; responsiveness passed across web, mobile, native simulation, and Electron simulation. - GLM swarm found useful issues on the previous head; they were fixed in the final head and verified by follow-up review. Two final-head GLM reruns failed in the reviewbot worker with an empty-output error before producing review text, so this was treated as reviewbot infra noise rather than a frontend finding. - Staging version matched 0fdc39ecddebc8f6730da82ff9838284924e9492. - Staging smoke/surface E2E passed: 24 passed, 6 skipped. - Broader staging read-only aggregate: 127 passed, 2 skipped, with one mobile NextGen staging API network miss that passed on exact focused rerun. - Production version matched 0c55e0c628541fb2ac695d87f871568848e7c057. - Production read-only aggregate passed: 65/65. Notes: - This is a test-harness-only release; no intended runtime UX or API behavior changed. - The staging aggregate transient should be watched as staging/API noise, but it did not reproduce and production validation was clean.
-
Testing hardening deploy update: PR #2846 is live in production. Shipped: - PR #2846: create-wave local sandbox E2E coverage https://github.com/6529-Collections/6529seize-frontend/pull/2846 Refs: - PR head: c3465affa45f262ccafb1e73682b9aab5e14d537 - Main / production SHA: cf0503f787ea4f0b86696f6e0dacc75c01e1ed3d - Staging deploy SHA: e541f00b36c08a15bda963de61a7f458a0f7b41b - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27967654287 - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27968664221 Validation: - Local changed lint, changed typecheck, Prettier, node syntax check, auth sandbox E2E, composer sandbox E2E, and diff whitespace checks passed before merge. - PR CI passed: DCO, CodeQL, Snyk, SonarCloud, Dependency Governance, installed app checks, and CodeRabbit. - 6529bot Opus follow-up review on c3465affa reported no new findings. - Staging deploy verified its HTTP version and staging E2E passed: 24 passed, 6 skipped across desktop/mobile Chromium. - Production deploy verified EB health, EB version, and HTTP version for cf0503f787ea4f0b86696f6e0dacc75c01e1ed3d. - Production read-only aggregate covered 65 tests; two unrelated single-test flakes were rerun cleanly, and production smoke passed 6/6. Notes: - This was a test-infrastructure release, not a runtime product change. - The production read-only flakes were isolated to a ReMemes filter wait and a third-party Google CSP report POST from legacy content; both reran cleanly and are good candidates for future test hardening.
-
Testing hardening deploy update: PR #2844 is live in production. Shipped: - PR #2844: authenticated local sandbox E2E coverage https://github.com/6529-Collections/6529seize-frontend/pull/2844 - Added `test:e2e:auth-sandbox` for authenticated notifications and direct-message creation flows against deterministic local sandbox data. - Added shared sandbox guard support that blocks unexpected unsafe writes, allows only exact loopback mock API mutations, aborts known third-party SDK background writes, and records unsafe audit evidence for oversized or failed body reads. Review and CI: - 6529bot was explicitly run on general, WCAG, i18n, security, responsiveness, and GLM-swarm lanes; final verdict on the latest head: no new findings. - SonarCloud duplication was resolved after refactoring shared sandbox setup. - DCO, CodeQL, Snyk, SonarCloud, CodeRabbit status, dependency governance, plan risk/security checks, and installed-app checks were green before merge. - Merged PR #2844 at `d26393b40d2fec0e9a2bf557f911324b27bc7686`. Staging: - Staging deploy SHA: `a803ae9d61e396112153418fb8b03f4a11cce562` - Production candidate contained in staging: `d26393b40d2fec0e9a2bf557f911324b27bc7686` - Staging deploy run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27955479616 - Local release-captain validation: staging `/api/version` matched, staging smoke 12/12 passed, staging surface pack 24 passed / 6 expected skips, staging WCAG/i18n surface matrix 6/6 passed. Production: - Production deploy run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27956752856 - Production deployed SHA/version: `d26393b40d2fec0e9a2bf557f911324b27bc7686` - Workflow passed build, package, main-advance preflight, S3 upload, Elastic Beanstalk deployment, health/readiness, EB version validation, and HTTP `/api/version` verification. - Local release-captain validation: production `/api/version` matched and production readonly E2E passed 65/65. Notes: - No backend deploy was required. - No user-visible behavior was intended to change; this is release-safety and agentic-development test infrastructure.
-
https://github.com/6529-Collections/6529seize-frontend/pull/2828 @[ragne]
-
https://github.com/6529-Collections/6529seize-frontend/pull/2825 @[ragne]
-
https://github.com/6529-Collections/6529seize-frontend/pull/2821 @[simo] @[prxt0] added initial ui/ux guidelines
-
@[punk6529] we ok to contribute to ops folder already?
-
-
Testing hardening deploy update: PRs #2820 and #2822 are live in production. Shipped: - PR #2820: local authenticated Waves composer sandbox E2E coverage https://github.com/6529-Collections/6529seize-frontend/pull/2820 Merge SHA: fa048794f72898c9604a7063e25192eaf2731c1c Adds a local mock API/Next runtime for composer-adjacent coverage, including attachment queue/remove and deterministic link-preview behavior, with diagnostics that fail if dangerous composer/upload mutation endpoints are touched. - PR #2822: production read-only E2E validation hardening https://github.com/6529-Collections/6529seize-frontend/pull/2822 Merge SHA: 7693d1138987175e0ccd6c54841d7547d99ce322 Aligns the 6529 Gradient title assertion with production's hydrated page title and adds `test:e2e:production:readonly`, a single aggregate production-safe read-only suite command. Deployment evidence: - PR #2820 staging deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27933627091 - PR #2820 production deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27934426448 - PR #2822 staging deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27937366218 - PR #2822 production deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27938016895 Validation: - PR #2820 local composer sandbox: 4/4 passed across desktop and mobile Chromium. - PR #2820 staging smoke: 24 passed, 6 expected skips. - PR #2820 production read-only packs: all current production-safe packs passed except one stale Gradient title assertion, which became PR #2822. - PR #2822 local collections read-only: 20/20 passed across desktop and mobile Chromium. - PR #2822 production collections read-only: 10/10 passed. - PR #2822 aggregate production read-only before deploy: 65/65 passed. - PR #2822 staging smoke after deploy: 24 passed, 6 expected skips. - PR #2822 aggregate production read-only after deploy: 65/65 passed. Review/CI: - DCO, App PR CI, dependency governance, CodeQL, SonarCloud, Snyk all green on #2822. - 6529bot Opus lane reported no findings on #2822. CodeRabbit was rate-limited on the final head and posted no actionable findings. Notes: - Staging collections read-only showed intermittent staging API fetch failures around a separate staging deployment. The focused failing mobile case passed on rerun, and production/local collections evidence is green, so this was recorded as staging environment/API stability signal rather than a release regression from #2822. - No user-facing product behavior changed here; this was internal testing/release-validation hardening, so I am not posting a public 6529 Releases note.
-
Deployment update: sidebar score polish and current main train are live. What shipped: - Frontend PR #2807: https://github.com/6529-Collections/6529seize-frontend/pull/2807 - Moved wave sidebar pin controls beside wave names. - Kept wave score aligned in the right-side score area. - Added a compact hover/focus explanation card for wave scores. - Frontend PR #2795: https://github.com/6529-Collections/6529seize-frontend/pull/2795 - Increased connected profile account limit from 3 to 5. - Release-safety and read-only E2E train now on main: PRs #2808, #2809, #2810, #2811, #2812, #2813, #2814, #2815, #2816, #2817, #2818, #2819, and #2820. Key SHAs: - PR #2807 squash merge SHA: 16ee2a4dec0e5f99e34263a33014272d105305d2 - Current production main SHA: fa048794f72898c9604a7063e25192eaf2731c1c - Staging deploy SHA containing current main: 1612c81ca6720550a105355b18f2a71d3dd8cedf Deployment evidence: - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27934791098 - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27934426448 - Production workflow confirmed main ref, build, bundle, S3 upload, Elastic Beanstalk update, EB health/readiness, and deployed version validation. Validation: - Staging / and /waves returned HTTP 200 after deploy. - Production / and /waves returned HTTP 200 after deploy. - PR #2807 also had local focused tests/lint/typecheck plus browser screenshots and a second visual review from Claude Opus 4.8 before merge. Public release note: - 4.41.8 posted to 6529 Releases: https://6529.io/waves/05b14183-e153-4e47-bc66-42a0f49102d4?drop=f9af3358-7443-406d-a45f-bcb43f6c5e06 Final state: production is live at fa048794f72898c9604a7063e25192eaf2731c1c. No rollback or fix-forward needed.
-
@[simo] @[ragne] increasing connected account limit from 3 to 5 https://github.com/6529-Collections/6529seize-frontend/pull/2795
-
Deployment update: broader release testing coverage is live. What shipped: - Frontend PR #2806: https://github.com/6529-Collections/6529seize-frontend/pull/2806 - Added read-only critical route-shell E2E guards for wallet/auth gates, messaging, notifications, public tools, and operator-only pages. - Merge SHA: 745130a19785fdc844410a2798ba63a6db8256e8 - Frontend PR #2805: https://github.com/6529-Collections/6529seize-frontend/pull/2805 - Added desktop/mobile Chromium surface-matrix E2E, browser-diversity lanes, native/Electron simulation lanes, required deployed pack metadata, and stricter release-report evidence handling. - Merge SHA and production version: cd8635e004191fc0509f253bc8d9a66c8ff51805 Deployment: - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27906005657 - Staging branch deploy SHA: ab1077c2d8949d421cf755872efc0868fe04057b - Production candidate SHA: cd8635e004191fc0509f253bc8d9a66c8ff51805 - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27906579357 - Production deployed SHA/version: cd8635e004191fc0509f253bc8d9a66c8ff51805 Validation: - Staging: surface matrix 24 passed / 6 skipped; WCAG/i18n surface matrix 6 passed; staging smoke 12 passed. - Production: API version matched cd8635e004191fc0509f253bc8d9a66c8ff51805; smoke surface matrix 12 passed; full surface matrix 24 passed / 6 skipped; WCAG/i18n surface matrix 6 passed. - Reviewbot lanes stayed additive: the existing general, wcag, i18n, security, and responsiveness reviews remain the floor, and GLM swarm review is live separately. Known follow-up: - Durable deployment artifact storage still needs operational wiring on 6529-controlled infrastructure. The deployment-bus accepts approved S3, artifact-service, or IPFS pointers, but the first expected S3 bucket was not present, so this train recorded validation evidence in workflow reports and release notes while leaving artifact storage as the next testing-roadmap task.
-
Deployment update: WCAG/i18n Playwright evidence pack is live. What shipped: - Frontend PR #2803: https://github.com/6529-Collections/6529seize-frontend/pull/2803 - Added the first browser WCAG/i18n evidence pack for The Memes route. - Added shared Playwright helpers for axe checks, expiring WCAG allowlists, visible keyboard focus, and locale-stress URLs. - Added Bootstrap Sass/Turbopack guard hardening needed for stable local/CI browser runs. - Existing 6529bot review lanes remain unchanged; this is additive local/CI evidence, not a replacement for reviewbot feedback. Merge and deploy: - PR head SHA: dc72bcf165cbd26931ec35837c152922482afaae - Main merge SHA: 952326d92c72d05230bbc35b6d21969de6b0832a - Staging deploy SHA: 85cd6d9df211c8a2ff681c20860ee24720ed8986 - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27889212939 - Production deployed SHA: 952326d92c72d05230bbc35b6d21969de6b0832a - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27889504543 Validation: - PR checks passed on latest head: App PR CI, dependency-governance, CodeQL, CodeRabbit, DCO, SonarCloud with 0 new issues, and Snyk. - CodeRabbit findings were fixed or withdrawn; no unresolved review threads remained. - Staging deployed and verified the expected checkout. - Staging smoke: `seize run test:e2e:staging` passed 6/6. - Staging WCAG/i18n pack: `seize run test:e2e:wcag-i18n` against staging passed 3/3. - Production workflow built, uploaded, deployed, reached healthy EB status, and verified the deployed SHA. - Production smoke: `seize run test:e2e:smoke` against https://6529.io passed 6/6. - Production WCAG/i18n pack: `seize run test:e2e:wcag-i18n` against https://6529.io passed 3/3. Operational notes: - The normal merge path was blocked by branch policy review state after the PR was authored by the same account. I used the admin merge path under the explicit autonomous rollout instruction after CI, CodeRabbit, 6529bot-triggered review cycle, Sonar, Snyk, and review-thread checks were green. - Public release note posted as 4.41.6: https://6529.io/waves/05b14183-e153-4e47-bc66-42a0f49102d4?drop=248c93e5-0f0a-4cbe-bd0e-bd92d17b344e - Next testing-roadmap step is to broaden these packs across more page clusters and wire the GLM/OpenRouter swarm review lane in the reviewbot repo without removing existing Opus/6529bot reviews.
-
Deployment update: staging smoke harness is live. What shipped: - Frontend PR #2793: https://github.com/6529-Collections/6529seize-frontend/pull/2793 - Production deployed SHA: `0c13aef47aee46316772bcabc02286d4fdad5032` - Staging merge SHA: `6424d8e253c70d840e729256b70ad338c022ba35` Reviewbot hardening completed during the loop: - Reviewbot #404: https://github.com/6529-Collections/6529reviewbot/pull/404 - Reviewbot #405: https://github.com/6529-Collections/6529reviewbot/pull/405 - Existing reviewbot lanes/configs were preserved; these were additive responsiveness-harness fixes. Deploy runs: - Staging: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27875512708 - Production: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27875842447 Validation: - Local focused checks passed before merge: risk floor, staging E2E, eslint on touched files, changed-file typecheck/lint, and diff check. - 6529bot general, WCAG, i18n, security, and responsiveness reviews were preserved and iterated until clean. - Latest responsiveness review passed 32/32 across web desktop, web mobile, native-mobile simulation, and electron-desktop simulation: https://github.com/6529-Collections/6529reviewbot/actions/runs/27875063485 - Deployed staging smoke: `seize run test:e2e:staging` passed 6/6 against `https://staging.6529.io`. - Production deploy workflow passed build, EB health/readiness, and validation. - Production smoke passed 6/6 read-only Playwright checks against `https://6529.io`. User-facing impact: - No visible site behavior changed. - The important change is release safety: deployed staging browser smoke now works with the staging access gate, and staging traces are disabled to avoid retaining the access code.
-
Deployment update: testing strategy foundation is live. What shipped: - Frontend PR #2792: https://github.com/6529-Collections/6529seize-frontend/pull/2792 - Added `ops/scripts/testing-strategy.cjs` with risk-floor computation, validation manifest checks, artifact pointer validation, and mutation-registry validation. - Added schema/example contracts under `ops/testing-strategy/`. - Added focused Jest coverage for risk classification, manifest validation, artifact safety, mutation registry validation, and preservation of the existing reviewbot lanes. - Updated the WCAG/i18n mega-run strategy docs and playbook. Existing reviewbot lanes remain the minimum PR review floor: general, wcag, i18n, security, responsiveness. Merge/deploy evidence: - PR head: f701d089740237500a571ab912a5befb945c5495 - Main merge SHA deployed to production: 8cad7422cad26e5941578f2971d23f63772e082d - Staging branch deploy SHA: d23f0b8572048730d90ad47ce33c08c3adf7b859 - Staging workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27870207986 - Production workflow: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27870575812 Validation: - PR checks passed: CodeQL, CodeRabbit, DCO, SonarCloud, Dependency Governance, Snyk. - 6529bot follow-up on the latest head reported no new findings. - CodeRabbit review threads were resolved and confirmed addressed. - Local validation before merge included focused Jest tests, changed lint/typecheck, manifest validation, mutation registry validation, risk-floor computation, and whitespace checks. - Staging deploy passed for the production-eligible candidate. - Staging browser smoke passed for `/`, `/about/the-memes`, and `/the-memes` with HTTP 200 and zero page errors. - Production deploy passed and Elastic Beanstalk verified the deployed SHA. - Production browser smoke passed for `/`, `/about/the-memes`, and `/the-memes` with HTTP 200 and zero page errors. Known follow-up: - The existing `test:e2e:staging` harness still fails before browser execution because `tests/testHelpers.ts` is missing on current main. This was documented in the testing strategy and is the next testing-foundation PR to repair. Manual staging and production browser smoke covered this release. - Production smoke saw three CloudFront 403 media rendition responses on public feed content; routes still loaded successfully with no page errors, and this PR did not touch runtime media behavior.
-
Deployment update: cleaner boosted image posts are live. What shipped: - Frontend PR #2791: https://github.com/6529-Collections/6529seize-frontend/pull/2791 - Boosted posts whose content is a standalone Markdown image now render that image through the normal media surface instead of showing a generic link-preview card and leftover Markdown text. - No backend changes. Deployment evidence: - PR merged to main at 42c59908226396a977fdee31afdca0be34233158. - Staging deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27867347351, staging SHA 530593bba2fb8638dd81ee5eb914d455d00c0a7e, production candidate 42c59908226396a977fdee31afdca0be34233158. - Production deploy: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27867643314, deployed SHA 42c59908226396a977fdee31afdca0be34233158. - GitHub production deployment 5132374031 ended success with Production deployment verified. Validation: - Review bots and required PR checks passed before merge; CodeRabbit and 6529bot had no remaining findings. - Focused unit tests, changed-file lint/typecheck, and whitespace checks passed before merge. - Staging browser smoke loaded the deployed app; the exact maybe's dive bar production wave is not present in staging data, so the release-specific visual check was completed on production after deploy. - Production browser validation on maybe's dive bar confirmed the gpebbles boosted post renders the transient image as media on desktop and mobile, with no raw img.transient.xyz text and no leftover !Seize Markdown text. - Console check only showed existing Permissions-Policy warnings unrelated to boosted media. Known note: - The existing staging Playwright command is currently blocked by missing ../testHelpers imports in several specs, so production promotion used workflow health plus targeted browser smoke/visual validation rather than that broken harness.
-
Deployment update: GitHub preview repo labels are live. What shipped: - Frontend PR #2790: https://github.com/6529-Collections/6529seize-frontend/pull/2790 - GitHub issue/PR preview cards now show compact repo labels for known 6529 repos, making active cross-repo work easier to scan. - The accessible label for GitHub preview cards now includes the full `owner/repo` context. - No backend changes. Code + merge: - PR head commit: 24b51ac2f4147f41e30320292d55a15884e74350 - Squash merge / production candidate: 532233bdd7a1a52fddca42fa77e543ca8bd31edb Validation before merge: - Focused component tests passed: `GithubPreviewStatusBadge.test.tsx` - Focused ESLint passed for touched files - `typecheck:changed` passed - DCO, CodeQL, Snyk, SonarCloud, and CodeRabbit passed; CodeRabbit had no actionable comments. Staging: - `1a-staging` deploy merge: 6f451daa8d7b2b44a46be5a9c0b9f2672d79183e, containing production candidate 532233bdd7a1a52fddca42fa77e543ca8bd31edb - Staging run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27864241224 - Staging deployment id: 5131636779, terminal status `success` - Staging browser smoke passed for `/`, `/about/the-memes`, and `/the-memes`: all 200s, expected titles/H1s, no access prompt, no relevant console errors. Production: - Production run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27864660143 - Production deployment id: 5131729798, terminal status `success` - Deployment manifest artifact id: 7763458792 - Elastic Beanstalk reached Green/Ready and validated deployed version 532233bdd7a1a52fddca42fa77e543ca8bd31edb. - Production browser smoke passed for `/`, `/about/the-memes`, and `/the-memes`: all 200s, expected titles/H1s, no relevant console errors. Notes: - The standard staging E2E script currently fails before browser launch because the checked-in specs import a missing `tests/testHelpers` module, so validation used an equivalent browser smoke for the release-critical pages. - No incidents, rollback, or follow-up deploy required. - Public release note posted as 4.41.2 in 6529 Releases: drop #1116074.
-
Deployment update: reviewbot WCAG + i18n second reviewer is live. Merged/deployed: - Reviewbot #399: https://github.com/6529-Collections/6529reviewbot/pull/399 - Frontend #2788: https://github.com/6529-Collections/6529seize-frontend/pull/2788 - Frontend #2789: https://github.com/6529-Collections/6529seize-frontend/pull/2789 Frontend PRs now automatically run general, WCAG, i18n, security, and responsiveness review lanes. Staging deploy passed, production deploy passed, and live browser smoke passed on 6529.io for Home, About The Memes, and The Memes. Prod run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27851738857
-
https://github.com/6529-Collections/6529seize-frontend/pull/2784 @[simo]
-
@[simo] @[ragne] responsive update for network nerd https://github.com/6529-Collections/6529seize-frontend/pull/2775
-
Deployment update: followed-subwave sidebar support is live. Backend #1656 is merged and deployed to staging + production on API commit 82f4f4f; prod Lambda reports 82f4f4f and the prod wave API returns score/REP data. Frontend #2774 is merged; current frontend main 523284a9 is deployed to production. E2E passed on staging and prod across desktop, Android S25-size, and tall mobile for /waves, /discover, /network/wave-score, and a live wave page. Sidebar checks passed: Highly Rated + All are visible and score values no longer show the word Score. FE prod run: https://github.com/6529-Collections/6529seize-frontend/actions/runs/27816708853 BE prod run: https://github.com/6529-Collections/6529seize-backend/actions/runs/27815639933
-
https://github.com/6529-Collections/6529seize-frontend/pull/2781 @[simo]
Loading more drops...