code-coordinator 0.5.294__tar.gz → 0.5.296__tar.gz

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/.github/coord-ci-acceptance.yml

@@ -73,21 +73,11 @@ acceptance:
           run: "pytest tests/acceptance"
           mock: "*.out"
           capability: python
-        - match: tui/**
-          kind: tui-tuidriver
-          # Unlike the two routes above, this one is identical to the
-          # fleet's real route: it never referenced `{ms}` even for
-          # `--issue` scoping, since `cargo test --test acceptance` always
-          # runs the whole tui/tests/acceptance.rs entrypoint (which
-          # `include!`s every ms-NN/*.rs slice) regardless of `{ms}`.
-          # RUSTC_BOOTSTRAP=1 + -Z unstable-options: --format json is
-          # nightly-gated; verified working on stable rustc (#1042).
-          # --features test-support: tui/Cargo.toml's `acceptance` test
-          # target has `required-features = ["test-support"]`, so a plain
-          # `cargo test` (no flag) silently skips building/running it
-          # altogether -- that gap is exactly what made this route
-          # unreachable from CI before #2180.
-          run: "cd tui && RUSTC_BOOTSTRAP=1 cargo test --test acceptance --features test-support -- -Z unstable-options --format json"
-          mock: "*.screen"
-          capability: rust
-          entrypoint: tui/tests/acceptance.rs
+        # #2899: the `tui/**` -> tui-tuidriver route is GONE from this file.
+        # The crate moved to the standalone `coord-tui` repo, taking its
+        # sealed entrypoint (`tui/tests/acceptance.rs` -> `tests/acceptance.rs`)
+        # and the `tui/tests/acceptance/ms-NN/` slices with it. coord-tui
+        # carries its own copy of this CI stand-in, keyed on its own repo
+        # name, so the route did not disappear -- it changed repos. This
+        # file is now single-route, and this repo's sealed set is exactly
+        # `tests/acceptance/**` again.

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/.github/workflows/test.yml

@@ -31,17 +31,15 @@ jobs:
       # testpaths=["tests"]) was the one accidental gap. (ms-51 — Playwright,
       # never .py either — left this repo entirely with the webapp, #2009.)
       - run: pytest --ignore=tests/acceptance
-      # #2897: the Rust drift gate `tests/test_generated_rust_fixture.py`
-      # no longer byte-compares the committed `tui/src/app/types/generated.rs`
-      # (see docs/ADR_COORD_TUI_CI.md and the header comment further down
-      # this file) — this explicit step stands in for that comparison while
-      # `tui/` still lives here. `$COORD_TUI_SRC=.` names this checkout as
-      # the coord-tui root (`RUST_OUTPUT_RELPATH` keeps the `tui/` prefix for
-      # exactly this reason). One matrix leg only: the generator's output
-      # depends on `coord.serve_app.openapi_spec()`, not the Python version.
-      - name: python scripts/codegen.py --rust --check
-        if: matrix.python-version == '3.12'
-        run: COORD_TUI_SRC=. python scripts/codegen.py --rust --check
+      # #2899: the `COORD_TUI_SRC=. python scripts/codegen.py --rust --check`
+      # step that used to live here is GONE. It was #2897's explicit stand-in
+      # for the byte-comparison, held here only for as long as the crate did
+      # — that ADR (docs/ADR_COORD_TUI_CI.md) says in terms: "the day
+      # coord-tui gets its own checkout + CI, delete the step below (its
+      # replacement lives in that CI, installing `code-coordinator[server]`
+      # from PyPI exactly as `coord-web`'s does)". This is that day. There is
+      # no `generated.rs` in this checkout to compare against, so the step
+      # could only fail or — worse — pass vacuously.
       # #2180: the cli-pytest sealed route (acceptance.drivers routes
       # "coord/**" -> cli-pytest -> tests/acceptance/ms-37) run through the
       # #2164 CI wrapper instead of a raw `pytest tests/acceptance` call, so
@@ -321,24 +319,21 @@ jobs:
   # for it to run against, and a job that installs Node to build nothing is
   # how the previous `acceptance`-on-every-push waste (#2389) started.
   #
-  # #1941/#2897: `scripts/codegen.py` also generates coord-tui's RUST wire
-  # types (`--rust`). #2897 gave that half the SAME cross-repo shape as the
-  # TS half above (`--out PATH` / `$COORD_TUI_SRC`, no fallback default,
-  # docs/ADR_COORD_TUI_CI.md) — but `tui/` has not actually moved out of
-  # this repo yet, so unlike the TS byte-comparison (which is now
-  # exclusively `coord-web`'s CI job), the Rust byte-comparison has nowhere
-  # else to run yet and stays HERE, as an explicit step below rather than a
-  # pytest test:
-  #
-  #     COORD_TUI_SRC=. python scripts/codegen.py --rust --check
+  # #1941/#2897/#2899: `scripts/codegen.py` also generates coord-tui's RUST
+  # wire types (`--rust`). #2897 gave that half the SAME cross-repo shape as
+  # the TS half above (`--out PATH` / `$COORD_TUI_SRC`, no fallback default,
+  # docs/ADR_COORD_TUI_CI.md), and #2899 moved the crate out — so the Rust
+  # byte-comparison is now, exactly like the TS one, EXCLUSIVELY the
+  # downstream repo's CI job. coord-tui's `.github/workflows/codegen-drift.yml`
+  # installs `code-coordinator[server]` from PyPI and runs
+  # `COORD_TUI_SRC=. python -m ... codegen --rust --check` against its own
+  # committed `src/app/types/generated.rs`.
   #
-  # `tests/test_generated_rust_fixture.py` itself is narrowed to what one
+  # Nothing about that gate can run here: this checkout has no `generated.rs`
+  # to compare against. What stays is
+  # `tests/test_generated_rust_fixture.py`, narrowed by #2897 to what one
   # checkout can prove — the generator runs and covers every schema — same
-  # as `tests/test_generated_types_fixture.py`; it no longer byte-compares
-  # the committed `tui/src/app/types/generated.rs`. The day coord-tui gets
-  # its own checkout + CI, delete the step below (its replacement lives in
-  # that CI, installing `code-coordinator[server]` from PyPI exactly as
-  # `coord-web`'s does) — no pytest file to un-narrow.
+  # as `tests/test_generated_types_fixture.py`.
   #
   # The consumer-side INTEGER-backed-bool text-scraping guard
   # (`coord/board_bool_guard.py`) that used to also read Rust source text

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/.gitignore

@@ -9,11 +9,6 @@ board_state.json
 agent_setup.json
 coordinator.yml
 
-# Rust build artifacts (coord-tui)
-tui/target/
-tui/*.err
-tui/*.out
-
 # scripts/coord-test-runner.sh scratch output, written into the worktree it
 # is pointed at. Harmless in the throwaway worktree the Test stage uses, but
 # a worker who runs the runner against its own checkout and then `git add -A`
@@ -33,11 +28,6 @@ tui/*.out
 .fallback.out
 .coord-ci-home/
 
-# #1973: opt-in local quadraui path override (see
-# tui/cargo-config-local-quadraui.toml.example) — never commit a
-# machine-local co-development override.
-tui/.cargo/config.toml
-
 # scratch $HOME created inside disposable worktrees by the CI-mode
 # acceptance runner (scripts/coord-test-runner.sh) — never meant to
 # leave the throwaway worktree, let alone land in git history (#2527).

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/CLAUDE.md

@@ -111,23 +111,25 @@ activation from an earlier command carried forward.
 The full suite exceeds Claude Code's 600s Bash ceiling on this repo and
 duplicates the Test stage + CI, which both run it against your pushed SHA
 regardless. Run just the file(s) that mirror what you changed — `pytest
-tests/test_<module>.py` for a `coord/<module>.py` change, `cargo test` from
-`tui/` scoped with a test-name filter for a `tui/**` change. To see what
+tests/test_<module>.py` for a `coord/<module>.py` change. To see what
 the Test stage itself would run for your diff (and confirm you're not
 missing a suite), `scripts/coord-test-runner.sh <worktree> --print-routing`
 computes the routing without actually building or testing anything.
 
-## Working on `tui/` — the `quadraui` pin
-
-**`coord-tui` pins `quadraui` to a git rev in `tui/Cargo.toml`**
-(`quadraui = { git = "https://github.com/JDonaghy/quadraui", rev = "<sha>" }`, #1973) —
-never edit that `rev` as a shortcut for co-development, and never build `tui/` against
-whatever happens to be checked out in `~/src/quadraui`; a quadraui merge broke coord-tui's
-build with zero coord-tui commits and no warning once already, which is exactly what the pin
-prevents. Bumping the pin deliberately, and building against an unmerged quadraui branch/PR
-without touching the pin, are both procedures — see the `tui-quadraui-workflow` skill
-(`coord/skills/tui-quadraui-workflow/SKILL.md`) for the steps; install it with `coord
-install-skills` if it isn't already on this machine.
+> **#2899: the TUI is not in this repo.** The `coord-tui` crate that used to live
+> under `tui/` is now the standalone [`JDonaghy/coord-tui`](https://github.com/JDonaghy/coord-tui)
+> repo, checked out at `~/src/coord-tui`. Its rules — the `quadraui` git-rev pin, the
+> `TuiDriver` harness, its sealed `tests/acceptance.rs` — live in **its** CLAUDE.md,
+> per this file's own scope rule above: a rule that does not change what someone
+> editing *this* repo does, does not belong here. If your briefing asks you to change
+> Rust, you are in the wrong checkout — say so rather than recreating `tui/`.
+>
+> Two seams still cross the split, and both run **from coord-tui's CI, pulling
+> `code-coordinator` from PyPI** — never the other way round: `scripts/codegen.py --rust`
+> (its `src/app/types/generated.rs` wire types) and `scripts/gen_board_fixture.py`
+> (its `tests/fixtures/board_sample.json`). Both take `--out PATH` or `$COORD_TUI_SRC`
+> and deliberately have **no fallback default**, so a missing checkout is an error
+> rather than a gate that passes against nothing.
 
 ## Key Design Decisions
 
@@ -163,7 +165,7 @@ The reviewer reads the rules and enforces them against the diff. It does not hav
 ## Rules for workers
 
 - **Only the coordinator writes docs.** Workers must **not** update README, CHANGELOG, or shared documentation files — parallel doc edits cause merge conflicts. If a briefing lists docs in `files_forbidden`, respect it. An issue whose *entire* deliverable is a doc edit is coordinator work and should never have been dispatched; say so and stop rather than editing the doc.
-- **Never edit the sealed suite — which is `tests/acceptance/**` *plus every declared driver entrypoint*.** Those suites are delivered read-only / run-only; write your own unit and internal tests instead. **In this repo the sealed set includes `tui/tests/acceptance.rs`** — it is the `tui-tuidriver` driver's `entrypoint:`, and an entrypoint is sealed as a whole file, so it is sealed despite living nowhere near `tests/acceptance/`. Do not be reassured by its prologue calling itself "a seam smoke test": *any* `type="work"` diff touching a sealed path is an **unconditional, mandatory `request-changes`** (`coord/review.py`), and the additive-only carve-out applies only to `test-author` / `mock-author` dispatches. TuiDriver tests belong **in-crate** under `#[cfg(test)]` anyway — see the coord-tui section below. The authoritative list is `AcceptanceConfig.sealed_paths()` in `coord/config.py`; if a briefing's `## Files` names a sealed path, that briefing is wrong — say so and put the test somewhere else rather than following it.
+- **Never edit the sealed suite — which is `tests/acceptance/**` *plus every declared driver entrypoint*.** Those suites are delivered read-only / run-only; write your own unit and internal tests instead. **In this repo, since #2899 moved `tui/` out, the sealed set is exactly `tests/acceptance/**`** — the `tui-tuidriver` `entrypoint:` that used to widen it (`tui/tests/acceptance.rs`, sealed as a whole file despite living nowhere near `tests/acceptance/`) went to the coord-tui repo with the crate, and is sealed *there*. The entrypoint rule itself has not changed, so do not assume a one-directory sealed set in any other repo. *Any* `type="work"` diff touching a sealed path is an **unconditional, mandatory `request-changes`** (`coord/review.py`), and the additive-only carve-out applies only to `test-author` / `mock-author` dispatches. The authoritative list is `AcceptanceConfig.sealed_paths()` in `coord/config.py` — it is derived from the configured driver `entrypoint:`, never hardcoded, which is why it followed the crate automatically; if a briefing's `## Files` names a sealed path, that briefing is wrong — say so and put the test somewhere else rather than following it.
 - **Stay in file scope.** If you must touch a file outside your briefing, note it in your final message.
 - **Commit and push before your final message** — even if the build is broken or you ran out of time. Uncommitted work is destroyed when the session ends.
 - **`gh` is on the deny-list.** The coordinator owns all GitHub interaction; use plain `git`.
@@ -189,12 +191,6 @@ versus keep in the coordinator session, and the coordinator session's own
 
 **How it runs:** black-box tests are part of the repo's normal test command, so the **Test stage** executes them on a capability-matched machine — `smoke_tests.capability_rules` route platform-specific suites to capable hardware (a GTK box; a machine with a browser). Favor the automated pre-review gate; the point is to trust the suite so manual/interactive smoke (incl. driving from a phone) is rarely needed.
 
-### coord-tui — quadraui `TuiDriver` (harness shipped: #690 / #691)
-- Drives the whole app through the real `event → handle → render` path against ratatui's headless `TestBackend` and asserts on the screen grid. `cargo test`-native, deterministic, no TTY.
-- `CoordApp` implements quadraui's `ShellApp`, so use `quadraui::tui::testing::driver_with_shell(app, CoordApp::shell_config(), w, h)`. API: `find("text")` → coords, `click(x, y)`, `press`, `type_char`, `screen()`, `screen_contains(needle)`. **Locate targets with `find` — never hardcode coordinates.**
-- **Reuse the existing fixtures** — `make_test_app(data: BoardData) -> CoordApp` (and `make_app_with_assignments`, `make_app_with_one_completed_issue`, …) in `tui/src/app.rs` build a full app from in-memory `BoardData`, no live daemon. Put the tests **in-crate** (`#[cfg(test)]`), **not** in `tui/tests/` — the fixtures are `#[cfg(test)]`/private and an integration-test crate can't see them.
-- Limit: `TuiDriver` renders to `TestBackend`, so it does **not** parse real ANSI — terminal-protocol bugs (raw-mode, SGR mouse, the embedded `claude` PTY pane) are out of reach and still need a live smoke. A native pty + vt100 tier is tracked in quadraui#302 (unbuilt).
-
 ### coord web (Phone Control Center)
 - The phone web app lives in `coord/dashboard/webapp/` (React / Vite / TS PWA, served by `coord/dashboard/server.py`). **Build the bundle before first use** — `dist/` is gitignored: `npm install && npm run build` from `coord/dashboard/webapp/`, re-run after pulling changes to `src/`. The server falls back to the legacy `index.html` when `dist/` is absent.
 - **Vitest unit tests** live in `coord/dashboard/webapp/src/components/__tests__/` (`npm test`). **Playwright E2E** specs live in `coord/dashboard/webapp/e2e/` (`npm run test:e2e`) and are the acceptance bar — browsers headless-test more cleanly than terminals, so lean on this gate rather than interactive smoke.

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-coordinator
-Version: 0.5.294
+Version: 0.5.296
 Summary: Coordinate Claude Code workers across multiple machines and repos.
 Author-email: John Donaghy <johnfdonaghy@gmail.com>
 License-Expression: LicenseRef-FSL-1.1-MIT
@@ -282,7 +282,8 @@ coord tui update      # downloads the release binary matching `coord --version`
 coord tui status       # check the installed version against the coordinator's
 ```
 
-(Building from source is only for developing the TUI itself: `cd tui && cargo build`.)
+(Building from source is only for developing the TUI itself, and it lives in its own
+repo since #2899: `git clone https://github.com/JDonaghy/coord-tui && cd coord-tui && cargo build`.)
 
 The whole lifecycle can also be driven as **human-attended interactive sessions** launched from
 a pipeline row's right-click menu — a testing agent, a reviewer, a merge agent, or a fix worker,

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/README.md

@@ -240,7 +240,8 @@ coord tui update      # downloads the release binary matching `coord --version`
 coord tui status       # check the installed version against the coordinator's
 ```
 
-(Building from source is only for developing the TUI itself: `cd tui && cargo build`.)
+(Building from source is only for developing the TUI itself, and it lives in its own
+repo since #2899: `git clone https://github.com/JDonaghy/coord-tui && cd coord-tui && cargo build`.)
 
 The whole lifecycle can also be driven as **human-attended interactive sessions** launched from
 a pipeline row's right-click menu — a testing agent, a reviewer, a merge agent, or a fix worker,

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/code_coordinator.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: code-coordinator
-Version: 0.5.294
+Version: 0.5.296
 Summary: Coordinate Claude Code workers across multiple machines and repos.
 Author-email: John Donaghy <johnfdonaghy@gmail.com>
 License-Expression: LicenseRef-FSL-1.1-MIT
@@ -282,7 +282,8 @@ coord tui update      # downloads the release binary matching `coord --version`
 coord tui status       # check the installed version against the coordinator's
 ```
 
-(Building from source is only for developing the TUI itself: `cd tui && cargo build`.)
+(Building from source is only for developing the TUI itself, and it lives in its own
+repo since #2899: `git clone https://github.com/JDonaghy/coord-tui && cd coord-tui && cargo build`.)
 
 The whole lifecycle can also be driven as **human-attended interactive sessions** launched from
 a pipeline row's right-click menu — a testing agent, a reviewer, a merge agent, or a fix worker,

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/code_coordinator.egg-info/SOURCES.txt

@@ -17,7 +17,6 @@ pyproject.toml
 .github/coord-ci-acceptance.yml
 .github/ISSUE_TEMPLATE/bug_report.md
 .github/workflows/auto-release.yml
-.github/workflows/cargo-test.yml
 .github/workflows/publish.yml
 .github/workflows/test.yml
 code_coordinator.egg-info/PKG-INFO
@@ -341,6 +340,7 @@ scripts/codegen.py
 scripts/coord-test-runner.sh
 scripts/coord-venv-rollback.sh
 scripts/drive-batch.sh
+scripts/extract-coord-tui.sh
 scripts/fleet_watchdog.py
 scripts/gen_board_fixture.py
 scripts/gen_mock_index.py
@@ -361,6 +361,10 @@ scripts/azure-workers/provision-worker.sh
 scripts/azure-workers/scrub-and-generalize.sh
 scripts/azure-workers/tailnet-acl.hujson
 scripts/azure-workers/verify-github-token.sh
+scripts/coord-tui-scaffold/CLAUDE.md
+scripts/coord-tui-scaffold/.github/coord-ci-acceptance.yml
+scripts/coord-tui-scaffold/.github/workflows/cargo-test.yml
+scripts/coord-tui-scaffold/.github/workflows/codegen-drift.yml
 tests/SMOKE_SET.md
 tests/__init__.py
 tests/backends.py
@@ -799,47 +803,7 @@ tests/fixtures/test_report/pytest_error_string_in_body.txt
 tests/fixtures/test_report/pytest_internal_error.txt
 tests/fixtures/test_report/pytest_mixed_failures.txt
 tests/fixtures/test_report/pytest_no_parseable_failures.txt
-tui/Cargo.lock
-tui/Cargo.toml
-tui/cargo-config-local-quadraui.toml.example
-tui/.github/workflows/release-tui.yml
-tui/src/commands.rs
-tui/src/lib.rs
-tui/src/main.rs
-tui/src/settings.rs
-tui/src/app/approved.rs
-tui/src/app/audit.rs
-tui/src/app/chat.rs
-tui/src/app/data.rs
-tui/src/app/dialogs.rs
-tui/src/app/doc_tabs.rs
-tui/src/app/drive.rs
-tui/src/app/drive_queue.rs
-tui/src/app/escalation.rs
-tui/src/app/events.rs
-tui/src/app/fixtures.rs
-tui/src/app/fleet_health.rs
-tui/src/app/fleet_sessions.rs
-tui/src/app/fleet_terminals.rs
-tui/src/app/format.rs
-tui/src/app/milestone_dag.rs
-tui/src/app/mod.rs
-tui/src/app/msv.rs
-tui/src/app/pipeline.rs
-tui/src/app/plans.rs
-tui/src/app/render.rs
-tui/src/app/reports.rs
-tui/src/app/sessions.rs
-tui/src/app/settings_ui.rs
-tui/src/app/sidebar.rs
-tui/src/app/terminal.rs
-tui/src/app/tests.rs
-tui/src/app/types.rs
-tui/src/app/workspace.rs
-tui/src/app/types/generated.rs
-tui/src/bin/gtk.rs
 tui/tests/acceptance.rs
-tui/tests/cli.rs
 tui/tests/acceptance/ms-33/audit_1039.rs
 tui/tests/acceptance/ms-33/contract.md
 tui/tests/acceptance/ms-33/manifest.yml
@@ -885,5 +849,4 @@ tui/tests/acceptance/ms-67/mocks/approved-items-chat-opened.screen
 tui/tests/acceptance/ms-67/mocks/approved-items-context-menu.screen
 tui/tests/acceptance/ms-67/mocks/approved-items-detail.screen
 tui/tests/acceptance/ms-67/mocks/approved-items-empty.screen
-tui/tests/acceptance/ms-67/mocks/approved-items-populated.screen
-tui/tests/fixtures/board_sample.json
+tui/tests/acceptance/ms-67/mocks/approved-items-populated.screen

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/code_coordinator.egg-info/scm_file_list.json

@@ -10,7 +10,6 @@
     ".github/ISSUE_TEMPLATE/bug_report.md",
     ".github/coord-ci-acceptance.yml",
     ".github/workflows/auto-release.yml",
-    ".github/workflows/cargo-test.yml",
     ".github/workflows/publish.yml",
     ".github/workflows/test.yml",
     ".gitignore",
@@ -343,8 +342,13 @@
     "scripts/check_gh_argv_containment.py",
     "scripts/codegen.py",
     "scripts/coord-test-runner.sh",
+    "scripts/coord-tui-scaffold/.github/coord-ci-acceptance.yml",
+    "scripts/coord-tui-scaffold/.github/workflows/cargo-test.yml",
+    "scripts/coord-tui-scaffold/.github/workflows/codegen-drift.yml",
+    "scripts/coord-tui-scaffold/CLAUDE.md",
     "scripts/coord-venv-rollback.sh",
     "scripts/drive-batch.sh",
+    "scripts/extract-coord-tui.sh",
     "scripts/fleet_watchdog.py",
     "scripts/gen_board_fixture.py",
     "scripts/gen_mock_index.py",
@@ -793,45 +797,6 @@
     "tests/test_worktree_rmtree_chokepoint.py",
     "tests/test_write_parity.py",
     "tests/write_parity.py",
-    "tui/.github/workflows/release-tui.yml",
-    "tui/Cargo.lock",
-    "tui/Cargo.toml",
-    "tui/cargo-config-local-quadraui.toml.example",
-    "tui/src/app/approved.rs",
-    "tui/src/app/audit.rs",
-    "tui/src/app/chat.rs",
-    "tui/src/app/data.rs",
-    "tui/src/app/dialogs.rs",
-    "tui/src/app/doc_tabs.rs",
-    "tui/src/app/drive.rs",
-    "tui/src/app/drive_queue.rs",
-    "tui/src/app/escalation.rs",
-    "tui/src/app/events.rs",
-    "tui/src/app/fixtures.rs",
-    "tui/src/app/fleet_health.rs",
-    "tui/src/app/fleet_sessions.rs",
-    "tui/src/app/fleet_terminals.rs",
-    "tui/src/app/format.rs",
-    "tui/src/app/milestone_dag.rs",
-    "tui/src/app/mod.rs",
-    "tui/src/app/msv.rs",
-    "tui/src/app/pipeline.rs",
-    "tui/src/app/plans.rs",
-    "tui/src/app/render.rs",
-    "tui/src/app/reports.rs",
-    "tui/src/app/sessions.rs",
-    "tui/src/app/settings_ui.rs",
-    "tui/src/app/sidebar.rs",
-    "tui/src/app/terminal.rs",
-    "tui/src/app/tests.rs",
-    "tui/src/app/types.rs",
-    "tui/src/app/types/generated.rs",
-    "tui/src/app/workspace.rs",
-    "tui/src/bin/gtk.rs",
-    "tui/src/commands.rs",
-    "tui/src/lib.rs",
-    "tui/src/main.rs",
-    "tui/src/settings.rs",
     "tui/tests/acceptance.rs",
     "tui/tests/acceptance/ms-33/audit_1039.rs",
     "tui/tests/acceptance/ms-33/contract.md",
@@ -878,8 +843,6 @@
     "tui/tests/acceptance/ms-67/mocks/approved-items-detail.screen",
     "tui/tests/acceptance/ms-67/mocks/approved-items-empty.screen",
     "tui/tests/acceptance/ms-67/mocks/approved-items-populated.screen",
-    "tui/tests/acceptance/ms-67/pull_decomposition_2533.rs",
-    "tui/tests/cli.rs",
-    "tui/tests/fixtures/board_sample.json"
+    "tui/tests/acceptance/ms-67/pull_decomposition_2533.rs"
   ]
 }

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/code_coordinator.egg-info/scm_version.json

@@ -1,7 +1,7 @@
 {
-  "tag": "0.5.294",
+  "tag": "0.5.296",
   "distance": 0,
-  "node": "gad3ed43a4b2310f6ef6968a7676f11e2deaca61f",
+  "node": "ga711fc248f5a8207d97fdc8ec5709d04beb8f505",
   "dirty": false,
   "branch": "HEAD",
   "node_date": "2026-08-29"

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/coord/commands/tui.py

@@ -325,12 +325,13 @@ def tui_update(
             click.echo(
                 f"refusing to overwrite {dest_path}: it reports version "
                 f"{DEV_BUILD_SENTINEL_VERSION!r}, the sentinel a locally "
-                "`cargo build`'d coord-tui always carries (tui/Cargo.toml's "
+                "`cargo build`'d coord-tui always carries (coord-tui's "
+                "Cargo.toml "
                 "committed version is only stamped for real in CI release "
                 "builds). This looks like a dev build someone is iterating "
                 "on -- pass --force to overwrite it anyway, or keep "
-                "building it yourself: cd tui && cargo build && cp "
-                f"target/debug/coord-tui {dest_path}",
+                "building it yourself: from a coord-tui checkout, "
+                f"cargo build && cp target/debug/coord-tui {dest_path}",
                 err=True,
             )
             sys.exit(3)

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/coord/config.py

@@ -1984,14 +1984,23 @@ class HealthConfig:
     # exists because it was found three releases stale on 2026-07-29.
     cli_venv_python: str | None = None
     # Absolute path to the built coord-tui binary.  None → ~/.local/bin/coord-tui
-    # (README: `cd tui && cargo build && cp target/debug/coord-tui
+    # (coord-tui's README: `cargo build && cp target/debug/coord-tui
     # ~/.local/bin/coord-tui`).
     tui_binary_path: str | None = None
-    # Directory holding the tui/ Rust sources the binary was built from.  None →
-    # `<checkout>/tui/src` for the first configured local checkout that has one.
+    # Directory holding the coord-tui Rust sources the binary was built from
+    # (#2899; before that split the crate lived at `tui/` inside THIS repo).
+    # None → `<checkout>/src` for this machine's `coord-tui` checkout,
+    # discovered the same way `coord_tui_checkout` is (see
+    # `coord.health.checks.deploy_lane_facts.resolve_coord_tui_checkout`),
+    # falling back to the pre-split `<checkout>/tui/src`.
     # Deliberately points at `src/`, not the crate root: rooting the mtime walk
     # above `target/` would sweep a multi-GB build dir on every refresh.
     tui_source_dir: str | None = None
+    # Root of a local `coord-tui` checkout (#2899). None → discover it from
+    # `repo_paths`/`ctx.checkouts` (a checkout named `coord-tui`, else one
+    # carrying `src/app/data.rs`).  Same convention as the lanes above: None
+    # means "discover it", never "disable the lane".
+    coord_tui_checkout: str | None = None
     # Absolute path to the live `coord web --dist` bundle (#1834 lane 5).
     # None → ~/coord-web-dist — the symlink `deploy/coord-web-dist-build.timer`
     # atomically repoints at each new release (#1543).
@@ -3601,6 +3610,7 @@ _HEALTH_OPT_STR_FIELDS: tuple[str, ...] = (
     "cli_venv_python",
     "tui_binary_path",
     "tui_source_dir",
+    "coord_tui_checkout",
     "deploy_dir",
     "systemd_user_dir",
     "webapp_dist_path",

{code_coordinator-0.5.294 → code_coordinator-0.5.296}/coord/drive_queue.py

@@ -3759,18 +3759,29 @@ def _reconcile_blocked(
     except a CONFIRMED-clear reading OR a probe that came back unreadable
     (#2806), which is deliberate: a `blocked` entry this sweep cannot say
     anything new about must render EXACTLY as it did before #2230 existed.
-    Three ways to land there:
+    Four ways to land there:
 
     * the block is PERMANENT (:func:`is_permanent_block_reason`) — #1844's
       guard refusal or #2019's dead end — neither of which any amount of
       re-checking can ever change;
+    * the block is `_resolve_prereqs`'s own unsatisfiable ``after=`` verdict
+      (:func:`_is_unsatisfiable_prereq_reason`, #2935) — a `waiting` entry
+      that never reached `running` has no assignment and no branch, so
+      there is categorically no merge-queue row for THIS sweep to have any
+      opinion about; the entry's fate is `_reconcile_blocked_after`'s alone
+      to decide, from the ``after=`` graph, on this or a later tick. Checked
+      — and returned from — before :func:`_blocked_gate_reading` even runs,
+      specifically so :func:`_reconcile_blocked_unreadable` never gets the
+      chance to overwrite `last_reason` with a merge-gate "could not read"
+      sentence and clobber the one marker `_reconcile_blocked_after` needs
+      to recognise this row again next tick — see #2935 for the incident a
+      missing version of this branch caused;
     * there is no evidence either way (:func:`_blocked_gate_reading` returns
       ``None``) AND :func:`_reconcile_blocked_unreadable` also has nothing
-      to add — an entry that never reached the merge queue at all (a
-      dispatch-time failure, an unsatisfiable ``after=``) has nothing this
-      sweep can cheaply re-check, and guessing would be exactly the "worse
-      than nothing" sweep the issue warns a naive "retry everything" pass
-      would be;
+      to add — a dispatch-time failure (#2589's two per-run shapes) has
+      nothing this sweep can cheaply re-check, and guessing would be
+      exactly the "worse than nothing" sweep the issue warns a naive
+      "retry everything" pass would be;
     * the gate is CONFIRMED still shut — the common, honest outcome for a
       `blocked` entry that has not in fact recovered yet.
 
@@ -3801,6 +3812,58 @@ def _reconcile_blocked(
     """
     if is_permanent_block_reason(entry.last_reason):
         return None
+    if _is_unsatisfiable_prereq_reason(entry.last_reason):
+        # #2935: this entry's ONLY `blocked` cause is `_resolve_prereqs`'s
+        # own after= verdict (step 4's launch walk, or its #2362/#2756
+        # sibling `_reconcile_blocked_after`, called immediately before this
+        # function in the same per-tick loop) — a `waiting` entry that never
+        # reached `running` at all (see `_resolve_prereqs`'s own comment:
+        # "attempts is deliberately NOT incremented: nothing was ever
+        # launched for this entry"). A never-dispatched entry has no
+        # assignment and no branch, so there is no merge-queue row for
+        # #2230's gate sweep to have ANY opinion about — a merge gate cannot
+        # exist for something that was never built, and probing for one
+        # anyway is a category error, not a failed probe.
+        #
+        # This must be checked BEFORE `_blocked_gate_reading`/
+        # `_reconcile_blocked_unreadable` below, not folded into their "no
+        # evidence" handling: `_reconcile_blocked_unreadable` WRITES
+        # `last_reason` the instant the shell's live probe was attempted and
+        # came back with no key (`live_blocked_unreadable`) — and doing that
+        # here would CLOBBER this exact marker text, which is the ONLY
+        # signal `_reconcile_blocked_after` uses (both here and on every
+        # future tick) to recognise this row as an after=-caused block worth
+        # re-diagnosing at all. The claude-coordinator#2935 incident: five
+        # dependents whose prerequisite had already landed stayed `blocked`
+        # for 178-207 consecutive ticks apiece, because the FIRST tick after
+        # they became re-evaluable this way overwrote their own "will never
+        # satisfy"/"not queued, not merged and not open" text with this
+        # sweep's "no merge-queue row ... could not be read" sentence —
+        # after which `_reconcile_blocked_after`'s guard could never
+        # recognise the row again, even once the prerequisite genuinely
+        # landed. Bailing out here, before any write, keeps the marker
+        # intact for as many ticks as it takes the after= graph to actually
+        # clear — the entry's fate stays entirely `_reconcile_blocked_
+        # after`'s to decide, exactly as the issue's "Expected behaviour"
+        # asks: answered from the after= graph, never from a merge-gate
+        # probe.
+        #
+        # Deliberately a text match on the SAME two verdict shapes
+        # `_reconcile_blocked_after` itself already trusts to gate its own
+        # resume — not the #2589/#2635 per-run dispatch markers
+        # (`is_dispatch_failure_reason`/`is_empty_branch_death_reason`),
+        # which stay OUT of this function on purpose (see
+        # `test_a_pre_dispatch_reason_text_does_not_suppress_a_real_
+        # unreadable_note`): those two describe what ONE launch attempt
+        # dispatched, which can go stale the moment an earlier attempt on
+        # the SAME entry left real evidence behind, so only the shell's live
+        # re-check (`live_blocked_unreadable`) is trusted to suppress them.
+        # `_resolve_prereqs`'s unsatisfiable verdict carries no such
+        # per-run ambiguity: it is only ever written for a `waiting` entry
+        # that has NEVER reached `running`, so the "nothing was ever
+        # dispatched" reading it encodes cannot be contradicted by an
+        # earlier attempt the way a per-run marker can.
+        return None
     reading = _blocked_gate_reading(entry, facts, live_blocked_gate)
     if reading is None:
         return _reconcile_blocked_unreadable(entry, live_blocked_unreadable)