sync: upstream and mobile background connection retention

.agents/skills/test-t3-app/references/sqlite-fixtures.md

@@ -4,7 +4,9 @@ Load this reference only when inspecting or seeding local T3 state directly.
 
 ## Select the correct database
 
-When `--base-dir` or `--home-dir` is explicit, runtime state lives under `<base-dir>/userdata` and the database path is `<base-dir>/userdata/state.sqlite`. The `<base-dir>/dev` state directory is only the fallback for an implicit development home, preventing an ordinary `vp run dev` from touching production state.
+When `--base-dir` or `--home-dir` is explicit, runtime state lives under `<base-dir>/userdata` and the database path is `<base-dir>/userdata/statev2.sqlite`. The `<base-dir>/dev` state directory is only the fallback for an implicit development home, preventing an ordinary `vp run dev` from touching production state.
+
+The server copies the V1 `state.sqlite` into `statev2.sqlite` only when `statev2.sqlite` is missing. After the first start, edits to `state.sqlite` change nothing.
 
 Start the target runtime once before seeding so all migrations have run. Use an isolated base directory. Stop the server before writes to avoid racing application state or an active projection.
 
@@ -23,7 +25,7 @@ Inspect current columns before writing a fixture:
 ```bash
 node apps/server/scripts/t3-sqlite-state.ts query \
   --base-dir <base-dir> \
-  --sql "PRAGMA table_info(projection_threads)"
+  --sql "PRAGMA table_info(orchestration_v2_projection_threads)"
 ```
 
 Apply a SQL fixture from a file:
@@ -38,21 +40,10 @@ Use one statement per invocation for both `query` and `exec`; the helper wraps w
 
 ## Seed projection data carefully
 
-The web UI primarily reads these projection tables:
-
-- `projection_projects`
-- `projection_threads`
-- `projection_thread_messages`
-- `projection_thread_activities`
-- `projection_thread_sessions`
-- `projection_turns`
-- `projection_pending_approvals`
-- `projection_thread_proposed_plans`
-
-Inspect `PRAGMA table_info(<table>)` and the current migrations under `apps/server/src/persistence/Migrations/` before constructing inserts. Keep identifiers unique, timestamps as ISO strings, JSON columns valid, and related project/thread/turn IDs consistent.
+Clients read projects from `projection_projects` and everything else from the `orchestration_v2_projection_*` tables: threads, runs, messages, turn items, runtime requests, plans, and provider sessions. The older `projection_thread*` tables hold V1 history, which the server imports once per thread at startup; later edits there do not reach the UI.
 
-For a substantial current example, inspect `seedDatabase` in `scripts/mobile-showcase-environment.ts`. Adapt its column set to the target database instead of assuming copied SQL remains current.
+Most V2 rows carry a `payload_json` that must decode against the schemas in `packages/contracts/src/orchestrationV2.ts`. The safest start is a row the app wrote itself: create a thread through the UI, copy its rows, and edit them. Keep identifiers unique, timestamps as ISO strings, and related project, thread, and run IDs consistent.
 
-Direct projection writes are appropriate for ephemeral visual states, edge-case counts, long titles, activity lists, and similar UI fixtures. They do not create a coherent orchestration event history. Do not modify `orchestration_events` unless the test specifically exercises projector internals, and do not use direct projection writes to claim backend business behavior works.
+Direct projection writes are appropriate for ephemeral visual states, edge-case counts, long titles, long timelines, and similar UI fixtures. They do not create a coherent event history. Leave the event log (`orchestration_events`) unchanged, and do not use direct projection writes to claim backend business behavior works.
 
 Use the app's commands or APIs for behavior tests. Use `node apps/server/src/bin.ts auth ...` for auth state rather than editing `auth_pairing_links` or `auth_sessions`.

.coderabbit.config.ts

@@ -0,0 +1,66 @@
+import { defineConfig, type CodeRabbitContext } from "@coderabbitai/config";
+
+const approvability = `Fail when a maintainer should read this pull request before CodeRabbit approves it, and name the rule and file. Fail if it:
+
+- Changes a product default: a setting's default value, or what users get without opting in. Making a feature do what it already promises is a bug fix, not a default change.
+- Adds or broadens a directive that disables or suppresses a lint, type-checker, LSP, or other static-analysis diagnostic, including file-level, line-level, and configuration-level overrides.
+- Adds a subsystem or user workflow, or is a large refactor across apps or packages.
+- Changes packages/contracts or persisted data in a way that existing clients or stored data might not accept.
+- Changes authentication, pairing, credentials, secrets, or remote connection trust.
+- Adds or changes an external side effect, such as acting on GitHub, publishing a release, or calling a webhook.
+- Adds, upgrades, or patches a dependency.
+- Changes CI or release configuration, agent or contributor instructions, or any review tool's configuration, including .github/, AGENTS.md, CONTRIBUTING.md, .agents/, and .coderabbit.config.ts.
+
+Otherwise pass. A focused bug fix, copy or layout fix, revert, or docs-only or test-only change passes unless a rule above applies. If you cannot decide, fail rather than report inconclusive. When failing, say that the pull request needs a maintainer's review.
+`;
+
+// Org members and collaborators merge their own pull requests. On anyone else's, CodeRabbit
+// requests changes until its comments are resolved and its checks pass, then approves.
+const UNGATED_AUTHORS = new Set(["OWNER", "MEMBER", "COLLABORATOR"]);
+const isGated = ({ pr }: CodeRabbitContext) => !UNGATED_AUTHORS.has(pr?.authorAssociation ?? "");
+
+export default defineConfig((ctx) => ({
+  reviews: {
+    high_level_summary: false,
+    review_status: false,
+    request_changes_workflow: isGated(ctx),
+    allow_author_approval: !isGated(ctx),
+    auto_review: {
+      enabled: true,
+    },
+    pre_merge_checks: {
+      docstrings: { mode: "off" },
+      override_requested_reviewers_only: isGated(ctx),
+      custom_checks: [
+        {
+          name: "Approvability",
+          mode: isGated(ctx) ? "error" : "off",
+          instructions: approvability,
+        },
+      ],
+    },
+    path_filters: [
+      // Vendored read-only reference checkouts of upstream Effect and Alchemy
+      // (see scripts/lib/reference-repos.ts). Nothing imports from them.
+      "!.repos/**",
+    ],
+    path_instructions: [
+      {
+        path: "{apps,packages,infra}/**/*.ts",
+        instructions: "Hold changed code to the rules in docs/internals/effect-services.md.",
+      },
+      {
+        path: "apps/web/src/**/*.{tsx,css}",
+        instructions: "Hold changed code to the rules in docs/internals/web-ui.md.",
+      },
+    ],
+  },
+  knowledge_base: {
+    code_guidelines: {
+      filePatterns: [
+        { files: "docs/internals/effect-services.md", applyTo: "{apps,packages,infra}/**/*.ts" },
+        { files: "docs/internals/web-ui.md", applyTo: "apps/web/src/**/*.{tsx,css}" },
+      ],
+    },
+  },
+}));

.cursor/rules/cursor-cloud.mdc

@@ -69,4 +69,4 @@ Gradle auto-downloads any additional pinned SDK packages on demand.
 ## Known flaky tests (unrelated to setup)
 
 - `apps/server/src/git/GitManager.test.ts` (cross-repo PR metadata, can time out at 12s).
-- `apps/server/src/provider/Layers/ProviderRegistry.test.ts` (codex binaryPath re-probe ordering).
+- `apps/server/src/provider/ProviderRegistry.test.ts` (codex binaryPath re-probe ordering).

.github/TRIAGE_EXEMPTIONS.td

@@ -1,6 +1,9 @@
 # Explicit contribution-triage exemptions; independent of VOUCHED.td.
 # Syntax: github:username (one entry per line; no denouncements).
 # Keep entries sorted alphabetically. Only the trusted upstream main copy applies.
+github:bmdavis419
 github:juliusmarminge
 github:maria-rcks
+github:markflorkowski
 github:t3dotgg
+github:Yash-Singh1

.github/VOUCHED.td

@@ -25,6 +25,7 @@ github:cursoragent
 github:D3OXY
 github:dbalders
 github:eggfriedrice24
+github:esthor
 github:extoci
 github:f-trycua
 github:flamboh

.github/workflows/ci.yml

@@ -255,6 +255,9 @@ jobs:
       # No Electron setup here: `t3` (apps/server) has no Electron dependency
       # and none of its tests touch the runtime. Only the non-server `test`
       # job, which covers @t3tools/desktop, needs the download.
+      - name: Install Chromium for server browser integration tests
+        run: node apps/server/node_modules/playwright-core/cli.js install --with-deps --only-shell chromium
+
       - name: Test
         env:
           T3CODE_TRANSFER_BUDGET_REPORT_PATH: ${{ runner.temp }}/t3code-transfer-budget.md
@@ -296,7 +299,11 @@ jobs:
   transfer-report:
     name: Transfer report artifact
     needs: test_server
-    runs-on: ubuntu-24.04
+    # On Blacksmith like every other job here: the GitHub-hosted queue is shared
+    # with the PR bots and can leave this job waiting until it is abandoned,
+    # which fails `Check` on an otherwise green run.
+    runs-on: blacksmith-2vcpu-ubuntu-2404
+    timeout-minutes: 5
     steps:
       - name: Require transfer measurements
         uses: actions/download-artifact@v8
@@ -487,6 +494,7 @@ jobs:
         test,
         test_web,
         test_server,
+        transfer-report,
         rust,
         mobile_native_changes,
         mobile_native_static_analysis,

.github/workflows/deploy-relay.yml

@@ -4,15 +4,15 @@ on:
   push:
     branches:
       - main
-  # Alchemy does not redeploy the Worker when only a Config value read in its
-  # Init changes (alchemy-run/alchemy#1831), so a changed repository variable
-  # needs a forced deploy to reach production.
+  # Redeploys main after a variable or secret change. Alchemy redeploys the
+  # Worker when a value it reads changes, so force is rarely needed, and it
+  # replaces the Postgres runtime role (alchemy-run/alchemy#1832).
   workflow_dispatch:
     inputs:
       force:
-        description: Redeploy every resource, including ones with no detected changes
+        description: Redeploy every resource, including ones with no detected changes. Replaces the Postgres runtime role.
         type: boolean
-        default: true
+        default: false
 
 permissions:
   contents: read
@@ -40,6 +40,7 @@ jobs:
       RELAY_API_ZONE_NAME: ${{ vars.RELAY_API_ZONE_NAME }}
       RELAY_TUNNEL_ZONE_NAME: ${{ vars.RELAY_TUNNEL_ZONE_NAME }}
       RELAY_TUNNEL_CLEANUP_MODE: ${{ vars.RELAY_TUNNEL_CLEANUP_MODE }}
+      RELAY_LEGACY_TUNNEL_CLEANUP_MODE: ${{ vars.RELAY_LEGACY_TUNNEL_CLEANUP_MODE }}
       CLERK_PUBLISHABLE_KEY: ${{ vars.CLERK_PUBLISHABLE_KEY }}
       CLERK_JWT_AUDIENCE: ${{ vars.CLERK_JWT_AUDIENCE }}
       APNS_ENVIRONMENT: ${{ vars.APNS_ENVIRONMENT }}

.github/workflows/mobile-eas-production.yml

@@ -18,8 +18,11 @@ name: Mobile EAS Production
 #      Store stays a manual App Store Connect step.
 #   2. OTA: publish a production-channel update for each platform where at
 #      least one finished production build matches the current native
-#      fingerprint. Old-version binaries with a matching fingerprint receive
-#      it too. When native drift means no binary could install the update,
+#      fingerprint. Older binaries of the same major version with a matching
+#      fingerprint receive it too; the major version is part of the
+#      fingerprint (apps/mobile/fingerprint.config.js), so a new major never
+#      reaches the previous major's store binaries over the air. When native
+#      drift means no binary could install the update,
 #      it is skipped and flagged in the job summary instead of published
 #      into the void.
 # workflow_dispatch remains as a manual override for both modes (e.g. to

.github/workflows/release.yml

@@ -330,6 +330,9 @@ jobs:
       # No Electron setup here: `t3` (apps/server) has no Electron dependency
       # and none of its tests touch the runtime. Only the non-server `test`
       # job, which covers @t3tools/desktop, needs the download.
+      - name: Install Chromium for server browser integration tests
+        run: node apps/server/node_modules/playwright-core/cli.js install --with-deps --only-shell chromium
+
       - name: Test
         run: vp run --filter t3 test --shard ${{ matrix.shard }}/${{ strategy.job-total }}
 
@@ -1347,6 +1350,24 @@ jobs:
             args:
               - --filter=@t3tools/scripts...
 
+      - name: Read published release notes
+        if: needs.preflight.outputs.is_prerelease == 'true'
+        continue-on-error: true
+        env:
+          GH_TOKEN: ${{ github.token }}
+          GH_REPO: ${{ github.repository }}
+          RELEASE_TAG: ${{ needs.preflight.outputs.tag }}
+          RELEASE_CHANNEL: ${{ needs.preflight.outputs.release_channel }}
+        run: |
+          if [ "$RELEASE_CHANNEL" = 'nightly' ]; then
+            if ! gh release view "$RELEASE_TAG" --json body --jq '.body // ""' > "$RUNNER_TEMP/discord-release-notes.md"; then
+              : > "$RUNNER_TEMP/discord-release-notes.md"
+              echo '::warning::Could not read published release notes; announcing without a changelog.'
+            fi
+          else
+            : > "$RUNNER_TEMP/discord-release-notes.md"
+          fi
+
       - name: Announce prerelease on Discord
         if: needs.preflight.outputs.is_prerelease == 'true'
         continue-on-error: true
@@ -1359,7 +1380,8 @@ jobs:
             --release-name "${{ needs.preflight.outputs.release_name }}" \
             --release-version "${{ needs.preflight.outputs.version }}" \
             --tag "${{ needs.preflight.outputs.tag }}" \
-            --release-url "https://github.com/${{ github.repository }}/releases/tag/${{ needs.preflight.outputs.tag }}"
+            --release-url "https://github.com/${{ github.repository }}/releases/tag/${{ needs.preflight.outputs.tag }}" \
+            --release-notes-file "$RUNNER_TEMP/discord-release-notes.md"
 
       - name: Announce latest release on Discord
         if: needs.preflight.outputs.make_latest == 'true'

.repos/alchemy-effect/.github/workflows/check.yml

@@ -90,6 +90,14 @@ jobs:
       - name: Run Typecheck
         run: pnpm tsc -b
 
+      # Must run after `tsc -b`: it reads distilled's emitted `lib/*.d.ts`,
+      # and stale/missing declarations resolve to `any`.
+      - name: Lint providers
+        run: pnpm lint:providers
+
+      - name: Test CLI launcher
+        run: timeout 240 pnpm test test/Cli/launcher.test.ts test/Cli/exitCodes.test.ts --retry 0 --concurrency 1
+
       - name: Check API JSDocs
         run: pnpm docs:check-jsdoc
 

.repos/alchemy-effect/.github/workflows/pkg.yml

@@ -22,6 +22,7 @@ on:
       - "tsconfig.json"
       - ".github/workflows/pkg.yml"
   pull_request:
+    types: [opened, synchronize, reopened, labeled]
     paths:
       - "submodules/distilled"
       - "packages/**"
@@ -54,6 +55,7 @@ jobs:
           # Pack the pull request head rather than the synthetic merge commit,
           # so every tarball is addressed by a commit that exists on the PR.
           ref: ${{ github.event.pull_request.head.sha || github.sha }}
+          fetch-depth: 0
 
       - name: Checkout distilled
         run: git submodule update --init --depth=1 --checkout -- submodules/distilled
@@ -69,19 +71,29 @@ jobs:
           install: true
           require-lockfile: true
 
+      # Cold-building the alchemy package peaks near 17 GB of heap, which
+      # starves the runner agent on a 32 GB machine and drops the job. A soft
+      # limit holds the peak near 11 GB for roughly twice the compile time.
       - name: Build packages
+        env:
+          GOMEMLIMIT: 6GiB
+          GOGC: "40"
         run: pnpm build:pkg
 
-      # Keep the directory lists in sync with the `build:pkg` script.
+      # The @distilled.cloud list is alchemy's dependency set, which `build:pkg`
+      # derives with `--filter=alchemy...`; keep the @alchemy.run list in sync.
       - name: Pack packages
         id: pack
         run: >-
           pnpm exec pkg pack
+          ${{ contains(github.event.pull_request.labels.*.name, 'force-ci') && '--all' || '' }}
+          --rebuild-all-path 'scripts/copy-package-files.ts'
           --group 'alchemy=./packages/alchemy'
           --group '@alchemy.run[Collapsed]=./packages/{better-auth,cloudflare-runtime,frontend-frameworks,node-utils,floci,pkg}'
-          --group '@distilled.cloud[Collapsed]=./submodules/distilled/packages/{core,aws,axiom,cloudflare,fly-io,github,hetzner,neon,planetscale,railway,stripe}'
+          --group '@distilled.cloud[Collapsed]=./submodules/distilled/packages/{core,acme,aws,axiom,cloudflare,doppler,fly-io,github,hetzner,infisical,neon,prisma,planetscale,railway,stripe,zerossl}'
 
       - name: Vouch for the manifest
+        if: steps.pack.outputs.package-count != '0'
         uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7.0.1
         with:
           name: ${{ steps.pack.outputs.artifact-name }}
@@ -91,4 +103,5 @@ jobs:
           retention-days: 1
 
       - name: Publish packages
+        if: steps.pack.outputs.package-count != '0'
         run: pnpm exec pkg publish

.repos/alchemy-effect/.github/workflows/release.yml

@@ -81,6 +81,11 @@ jobs:
           bun scripts/release/prepare.ts '${{ inputs.version }}' | tee -a "$GITHUB_OUTPUT"
 
       - name: Build release packages
+        # Match preview CI: the cold tsgo build can starve the runner
+        # without Go heap limits on this 32 GB machine.
+        env:
+          GOMEMLIMIT: 6GiB
+          GOGC: "40"
         run: >-
           pnpm turbo run build
           $(jq -r '.[] | "--filter=\(.name)"' release-packages.json)

.repos/alchemy-effect/.gitignore

@@ -32,6 +32,11 @@ smoke
 *.d.ts
 !examples/prisma-tanstack-start/migrations/**/start-contract.d.ts
 !examples/prisma-tanstack-start/migrations/**/end-contract.d.ts
+# Emitted Prisma ORM v8 contract types: imported by the fixtures/example that
+# consume them, so a clean checkout must typecheck without running a deploy.
+!examples/cloudflare-neon-prisma/src/prisma/generated/contract.d.ts
+!packages/alchemy/test/Cloudflare/Workers/fixtures/prisma-orm/generated/contract.d.ts
+!packages/alchemy/test/Prisma/ORM/fixtures/client/generated/contract.d.ts
 # drizzle-kit's generated durable-sqlite migrations bundle (checked-in test fixture)
 !packages/alchemy/test/Cloudflare/Workers/fixtures/drizzle-do/drizzle/migrations.js
 # Container image contexts COPY these into the image; without the negation
@@ -67,6 +72,8 @@ examples/cloudflare-worker/src/scratch.ts
 /packages/alchemy/README.md
 
 !/packages/alchemy/bin/cli.js
+!/packages/alchemy/bin/alchemy.js
+!/packages/alchemy/bin/exec.js
 !/packages/alchemy/bin/register-dev-mode.js
 !/packages/alchemy/bin/register-oxc.js
 .tmp

.repos/alchemy-effect/AGENTS.md

@@ -857,8 +857,8 @@ Resolve the bindings, expose one route per behavior, default-export the class so
 // fixtures/worker.ts
 import * as Cloudflare from "@/Cloudflare/index.ts";
 import * as Effect from "effect/Effect";
-import { HttpServerRequest } from "effect/unstable/http/HttpServerRequest";
-import * as HttpServerResponse from "effect/unstable/http/HttpServerResponse";
+import { HttpServerRequest } from "effect/http/HttpServerRequest";
+import * as HttpServerResponse from "effect/http/HttpServerResponse";
 import { Gateway } from "./gateway.ts";
 
 export default class TestWorker extends Cloudflare.Worker<TestWorker>()(
@@ -895,7 +895,7 @@ import * as Test from "@/Test/Alchemy";
 import { expect } from "alchemy-test";
 import * as Effect from "effect/Effect";
 import * as Schedule from "effect/Schedule";
-import * as HttpClient from "effect/unstable/http/HttpClient";
+import * as HttpClient from "effect/http/HttpClient";
 import TestWorker from "./fixtures/worker.ts";
 
 const { test, beforeAll, afterAll, deploy, destroy } = Test.make({