chore(just): consolidate full-suite actions

.github/workflows/cache.yml

@@ -104,7 +104,7 @@ jobs:
       # and the test artifacts (codegen + linked test binaries). They share
       # little, and a PR usually needs both.
       - name: Check
-        run: nix develop --command just check-all
+        run: nix develop --command just check --all
         env:
           MOQ_STRICT: 1
 

.github/workflows/smoke.yml

@@ -81,7 +81,7 @@ jobs:
         shell: bash -leo pipefail {0}
 
       - name: Smoke (full matrix)
-        run: nix develop --command just test smoke-full
+        run: nix develop --command just test smoke --all
         shell: bash -leo pipefail {0}
 
       - name: Negative control

.gitignore

@@ -9,7 +9,7 @@
 # Claude Code scratch space. These are untracked working directories, and
 # `just _changed` counts untracked files, so leaving them visible pushes the
 # changed-file list past its cap and silently turns every scoped `just check`
-# into `check-all`. The worktrees also carry their own biome.jsonc, which the
+# into `check --all`. The worktrees also carry their own biome.jsonc, which the
 # unscoped run then rejects as a nested root config.
 /.claude/cache/
 /.claude/tmp/

CLAUDE.md

@@ -107,6 +107,6 @@ These diff the branch against its base and only run the affected packages.
 
 Any wire-format change updates its matching IETF draft in the same PR, including framing, message fields, enum values, and version negotiation. Use the feature-specific draft for extensions and validate with `just drafts check`. See `drafts/CLAUDE.md`.
 
-For wire, `moq-ffi`, or gateway changes, also run `just test smoke-full` for cross-language interop; plain `smoke` is Rust-only.
+For wire, `moq-ffi`, or gateway changes, also run `just test smoke --all` for cross-language interop; plain `smoke` is Rust-only.
 
 When a CLI interface changes, search the whole repo for the binary name and update every example invocation, including docs and demo recipes. Check examples against `--help`.

README.md

@@ -132,9 +132,9 @@ just test
 just fix
 
 # Same as the above, over every package
-just check-all
+just check --all
 just test all
-just fix-all
+just fix --all
 ```
 
 CI runs these same two recipes, so they cover the same ground locally. It sets two things you don't: `MOQ_STRICT=1`, which turns a missing tool into an error instead of a skipped check, and `NEXTEST_PROFILE=ci`, which allows a longer hang timeout.

doc/setup/dev.md

@@ -16,7 +16,7 @@ match CI.
 | `just fix` | Format and lint the packages this branch changed. |
 | `just check` | Compile and lint the same scope. This is what CI runs. |
 | `just test` | Run tests for the same scope. |
-| `just fix-all`, `just check-all`, `just test all` | The same, over every package. |
+| `just fix --all`, `just check --all`, `just test all` | The same, over every package. |
 | `just pub bbb <url>` | Publish Big Buck Bunny (also `tos`, `clock`, `gst`, `hls`). |
 | `just sub gst bbb <url>` | Play a broadcast through GStreamer. |
 | `just relay` | Run a local relay on its own. |

justfile

@@ -364,7 +364,7 @@ _changed-test $LIMIT=changed_max:
     fail() { echo "changed: _changed-test: $1" >&2; exit 1; }
 
     # Synthetic sizes rather than whatever the working tree happens to hold: a
-    # clean checkout has no diff at all, and `check-all` runs there (cache.yml
+    # clean checkout has no diff at all, and `check --all` runs there (cache.yml
     # warms the cache from `main`), so a test keyed on the real list would take
     # down the one job allowed to write the shared Rust cache.
     [[ "$(printf 'aaaa' | just _changed-cap 3)" == ALL ]] || fail "over budget must print ALL"
@@ -424,7 +424,7 @@ _echo $VALUE:
 # Required per scope, mirroring what `check` actually dispatches for a given
 # diff: demanding gradle on a docs-only PR would fail a run that was never going
 # to invoke it. Takes the same file list as the dispatch, or `ALL` to require
-# everything (`check-all`).
+# everything (`check --all`).
 #
 # One deliberate absence: swift exists only on macOS, and `swift check` skips
 # off-macOS by design; swift.yml is its real gate.
@@ -480,36 +480,49 @@ _tools $FILES="":
 # Lints and compiles only the packages the branch changed plus everything
 # depending on them, so several worktrees can build at once. This is also what
 # CI runs (with MOQ_STRICT=1), so there is no second, drifting definition of
-# "checked". Tests are the sibling `just test`; `check-all` is the unscoped suite.
+# "checked". Tests are the sibling `just test`; `check --all` is the unscoped suite.
 
 # Lint and compile what the branch changed since BASE, plus its dependents.
-check $BASE="":
+check $BASE="" *args:
     #!/usr/bin/env bash
     set -euo pipefail
 
-    files=$(just _changed "$BASE")
+    if [[ "$BASE" == --all ]]; then
+        files=ALL
+    else
+        files=$(just _changed "$BASE")
+    fi
 
     # `_changed` says ALL when the list outgrew what argv can carry. The unscoped
     # suite is the path that passes no list at all, so it is the one that works.
-    if [[ "$files" == ALL ]]; then
-    	just check-all
-    	exit 0
-    fi
-
-    just _tools "$files"
-
     # The dispatch below lives in these two files, and neither matches any
     # language scope, so a PR that rewrites how CI dispatches would otherwise
-    # validate none of it. Hand off to the unscoped suite instead.
-    if grep -qE '^(justfile|test/justfile)$' <<< "$files"; then
-    	echo "check: root orchestration changed; checking everything." >&2
-    	just check-all
-    	exit 0
+    # validate none of it. Widen to the unscoped suite instead.
+    if [[ "$files" != ALL ]] && grep -qE '^(justfile|test/justfile)$' <<< "$files"; then
+        echo "check: root orchestration changed; checking everything." >&2
+        files=ALL
     fi
 
-    # An empty list means "force-run" to the per-lang recipes, which is the
-    # wrong semantic here, so don't dispatch at all.
-    if [[ -n "$files" ]]; then
+    just _tools "$files"
+
+    if [[ "$files" == ALL ]]; then
+        just js check
+        just drafts check
+        just rs check --workspace --exclude moq-net-fuzz {{ args }}
+        just rs tokio-features
+        just --justfile bench/justfile check
+        cargo run --quiet --locked --package quest -- check
+        # Not covered by the line above: moq-wasm only exists on the wasm32 target.
+        just rs wasm
+        just py check
+        just kt check
+        just swift check
+        just go check
+        just dart check
+        just obs check
+        just obs compile
+        just _flake
+    elif [[ -n "$files" ]]; then
         just js check "$files"
         just rs check-changed "$files"
         if echo "$files" | grep -q '^bench/'; then
@@ -560,27 +573,6 @@ check $BASE="":
 
     just _check-common
 
-# Check every package in every language, plus moq-wasm.
-check-all *args:
-    just _tools ALL
-    just js check
-    just drafts check
-    just rs check --workspace --exclude moq-net-fuzz {{ args }}
-    just rs tokio-features
-    just --justfile bench/justfile check
-    cargo run --quiet --locked --package quest -- check
-    # Not covered by the line above: moq-wasm only exists on the wasm32 target.
-    just rs wasm
-    just py check
-    just kt check
-    just swift check
-    just go check
-    just dart check
-    just obs check
-    just obs compile
-    just _flake
-    just _check-common
-
 # Skips when nix is absent: the flake is not a precondition for working on the
 # repo, and `_tools` already makes it required under MOQ_STRICT.
 
@@ -589,7 +581,7 @@ check-all *args:
 _flake:
     @if command -v nix >/dev/null 2>&1; then nix flake check; fi
 
-# Repository-wide non-compiling checks shared by `check` and `check-all`.
+# Repository-wide non-compiling checks shared by scoped and unscoped `check`.
 # Optional shell, workflow, TOML, Nix, and justfile lints skip if missing.
 #
 # `bun install` because remark-cli lives in node_modules and `just js check` is
@@ -745,7 +737,7 @@ _markdown-test:
 
     echo "markdown: check/fix regression ok"
 
-# Repository-wide lints, shared by `check` and `check-all`.
+# Repository-wide lints, shared by scoped and unscoped `check`.
 [private]
 _check-common:
     just _changed-test
@@ -759,22 +751,27 @@ _check-common:
     just gh check
 
 # Scoped exactly like `check`, because `clippy --fix` compiles what it fixes.
-# `fix-all` is the unscoped version.
+# Pass `--all` for the unscoped version.
 
 # Auto-fix lint and formatting for what the branch changed since BASE.
 fix $BASE="":
     #!/usr/bin/env bash
     set -euo pipefail
 
-    files=$(just _changed "$BASE")
+    if [[ "$BASE" == --all ]]; then
+        files=ALL
+    else
+        files=$(just _changed "$BASE")
+    fi
 
     # Mirrors `check`: too long for argv means fix everything instead.
     if [[ "$files" == ALL ]]; then
-    	just fix-all
-    	exit 0
-    fi
-
-    if [[ -n "$files" ]]; then
+        just js fix
+        just rs fix --workspace --exclude moq-net-fuzz
+        just py fix
+        just dart fix
+        just obs fix
+    elif [[ -n "$files" ]]; then
     	just js fix "$files"
     	just rs fix-changed "$files"
     	just py fix "$files"
@@ -788,19 +785,10 @@ fix $BASE="":
 
     just _fix-common
 
-# Auto-fix every JavaScript workspace and every default Rust member.
-fix-all:
-    just js fix
-    just rs fix --workspace --exclude moq-net-fuzz
-    just py fix
-    just dart fix
-    just obs fix
-    just _fix-common
-
 # Optional tools skip if missing locally. `bun install` for the same reason as
 # `_check-common`.
 
-# Repository-wide fixes, shared by `fix` and `fix-all`.
+# Repository-wide fixes, shared by scoped and unscoped `fix`.
 [private]
 _fix-common:
     bun install

quest/dev/api-origin-scopes.md

@@ -19,7 +19,7 @@ cross-language binding mirror that PR explicitly left unfinished.
   with `captures` on `MoqAnnounceUpdate`; `rs/libmoq`
   (`moq_announce_update` gains the captures under the existing string-out
   convention), the `py`, `swift`, `kt`, `dart`, and `go` wrappers,
-  `doc/lib/*` for each, and `just test smoke-full`.
+  `doc/lib/*` for each, and `just test smoke --all`.
 - `announced_broadcast(path)` keeps its shape; it is the literal case.
 
 Public API: breaking on moq-ffi, libmoq, and every binding, so on dev.

quest/dev/merge-dev.md

@@ -11,8 +11,8 @@ adopts.
 ## Plan
 
 main is merged into dev as of 2026-09-12, so the merge PR opens from dev with
-the list below as closing keywords once the gates clear. Run `just check-all`,
-`just test all`, `just test smoke-full`, and `just bench origin/main` on the
+the list below as closing keywords once the gates clear. Run `just check --all`,
+`just test all`, `just test smoke --all`, and `just bench origin/main` on the
 merged tree, and record the revision the proof ran on.
 
 Resolve by behavior, never by side. The traps the last main-into-dev merge

quest/next/3021-moq-gst-anchor-generated-media-timelines-to-wall-clock.md

@@ -30,7 +30,7 @@ delayed first buffers, multiple pads, timescale conversion, numeric limits,
 source restarts, idle gaps, and a system-clock adjustment. Existing timeline
 records remain unchanged. Consume the prerequisite's catalog format; no new
 transport TIMESTAMP/TIMESCALE semantics, synchronization protocol, or drift
-correction is introduced here. Run the GStreamer CI and smoke-full lanes.
+correction is introduced here. Run the GStreamer CI and `smoke --all` lanes.
 
 ## Required
 

quest/next/README.md

@@ -34,7 +34,7 @@ Implementation that depends on a dev contract requires
 - [@moq/net additive](/quest/next/js-net-additive.md) - a live-broadcasts getter, `Table.dynamic`, credential refresh before redial, inferred `share`
 - [Last frame duration](/quest/next/mux-last-frame-duration.md) - a group's final frame keeps its duration on 90 kHz and nanosecond imports
 - [Snapshot clobber](/quest/next/json-modify-clobber.md) - a snapshot edit fails on a shape mismatch instead of seeding a default
-- [Binding audio tests](/quest/next/binding-audio-tests.md) - every binding proves the Opus frame duration and throwing setters it exposes, and smoke-full publishes audio with an explicit config
+- [Binding audio tests](/quest/next/binding-audio-tests.md) - every binding proves the Opus frame duration and throwing setters it exposes, and `smoke --all` publishes audio with an explicit config
 - [Decode format](/quest/next/ffi-decode-format.md) - the C-only decode pixel format knob reaches every uniffi binding
 - [JSON mutate](/quest/next/json-mutate.md) - Rust gains the closure edit JS already has, beside the guard
 - [Publisher clocks](/quest/next/publisher-clock.md) - wire the shared clock through native and browser publisher restarts

quest/next/auth/moq-transport.md

@@ -96,7 +96,7 @@ union without disconnecting, cancelling only work that loses authorization;
 an out-of-scope new publication aborts before any PUBLISH_NAMESPACE is written; a
 peer without the option (the relay built without it, and the interop runner's
 reference relay) sees no AUTH stream and keeps working. Run
-`just test smoke-full`.
+`just test smoke --all`.
 
 On main, additive.
 

quest/next/binding-audio-tests.md

@@ -7,7 +7,7 @@ consumer's. Today only Python asserts a sub-millisecond Opus frame duration
 (`py/moq-rs/tests/test_local.py`), Go pins only the 20 ms default, Kotlin and
 Swift never set one, no Kotlin test proves the fallible configuration setters
 throw (#3642 added the Python and Swift cases and skipped `kt/`), and
-`smoke-full` publishes audio from no binding with an explicit codec config.
+`smoke --all` publishes audio from no binding with an explicit codec config.
 
 ## Plan
 
@@ -18,7 +18,7 @@ refuses a value outside the Opus set. Add a Kotlin test that a setter called
 during an in-flight connect throws and the client is cancelled, mirroring
 `py/moq-rs/tests/test_server.py`. Give the smoke publishers under
 `test/smoke/` an Opus config with a non-default frame duration so
-`just test smoke-full` exercises the audio path across bindings. Dart omits
+`just test smoke --all` exercises the audio path across bindings. Dart omits
 audio by design ([Dart codecs](/quest/next/dart-codecs.md)).
 
 Public API: none. Wire: none.

quest/next/check-uring-feature.md

@@ -13,7 +13,7 @@ feature for days (fixed in #3749) with every PR green.
 
 In `rs/justfile`, `check-changed` runs the feature clippy from the nightly
 `uring` recipe (`--features io-uring`, `-D warnings`, no tests) when
-`_select` includes moq-relay; `check-all` already covers it through `bench
+`_select` includes moq-relay; `check --all` already covers it through `bench
 check`. Keep the nightly `uring` matrix for the tests. Verify by breaking the
 feature on a scratch branch and watching `just check` fail on an
 `rs/moq-relay` diff, and by timing the added compile on a warm cache.

quest/next/cpp/package.md

@@ -5,7 +5,7 @@
 `find_package(moq)` or the tarball's `moq.pc` gives a C++17 program
 `<moq/moq.hpp>`, and the program publishes and subscribes through the full
 moq-ffi surface with futures and `moq::expected`, `co_await`-able on C++20.
-A `test/smoke/clients/cpp` client joins `just test smoke-full`, and
+A `test/smoke/clients/cpp` client joins `just test smoke --all`, and
 `doc/lib/cpp` documents the package the way `doc/lib/go` does Go.
 
 ## Plan

quest/next/cs/package.md

@@ -6,7 +6,7 @@
 with the mocking library on NuGet) gives a .NET 8 program the full moq-ffi
 surface as `await`-able classes, with `runtimes/<rid>/native` libraries for
 the release matrix. A `test/smoke/clients/cs` client joins `just test
-smoke-full` and `doc/lib/cs` documents the package.
+smoke --all` and `doc/lib/cs` documents the package.
 
 ## Plan
 

quest/next/decoded-frames.md

@@ -35,7 +35,7 @@ fields to a published C struct is not automatically additive.
 Test handle release, conversion failures, cancellation, delayed consumption,
 and retained ownership through the existing libmoq/FFI test lanes. Platform
 adapters own their hardware import proof. Update `moq.h`, affected wrappers,
-and C/binding documentation; run `just test smoke-full` in CI.
+and C/binding documentation; run `just test smoke --all` in CI.
 
 Public API: owned frame access and conversion at the binding boundary. Wire:
 none. Consume the settled main frame/output contracts without replacing them.

quest/next/ffi-websocket-fallback.md

@@ -49,4 +49,4 @@ Cross-package sync from the root CLAUDE.md table, with the lines that mirror
 
 Tests: a moq-ffi unit test that each setter lands in `config.websocket`, and a
 wrapper test where the tls setter is covered today
-(`py/moq-rs/tests/test_server.py`). `just test smoke-full` for the bindings.
+(`py/moq-rs/tests/test_server.py`). `just test smoke --all` for the bindings.

quest/next/flate/bindings.md

@@ -46,7 +46,7 @@ sibling. Document in `doc/lib/{c,py,swift,kt,go,dart}` beside the JSON entry.
 Tests: a moq-ffi round trip next to `json_snapshot_roundtrip`, a libmoq C
 round trip in `src/test.rs`, and one cross-language check that a C-published
 group decodes with the shared vector from the track quest. Run
-`just test smoke-full`.
+`just test smoke --all`.
 
 Public API impact: additive on moq-ffi, libmoq, and every wrapper; `main`.
 Wire impact: none.

quest/next/intra-refresh/bindings.md

@@ -20,7 +20,7 @@ and changes the libmoq C struct layout, so it targets `dev`.
 - Hand-written wrappers and docs per the cross-package table: `py/moq-rs`,
   `swift/`, `kt/`, `dart/moq`, `go/wrapper/moq`, and `doc/lib/{py,swift,kt,go,dart,c}`.
   Go gets no uniffi default, so its zero value must read as keyframe mode.
-- Run `just test smoke-full` for the cross-language check.
+- Run `just test smoke --all` for the cross-language check.
 
 ## Required
 

quest/next/js-fetch.md

@@ -30,7 +30,7 @@ Use an in-memory application responder for verification, without OPFS or an
 archive writer. A browser publisher serves a native subscriber after a group
 is evicted or was never cached; verify exact group/frame replay, empty groups,
 missing groups, concurrent requests, and cancellation while awaiting a reply.
-Run the supported-draft matrix and `just test smoke-full` through CI.
+Run the supported-draft matrix and `just test smoke --all` through CI.
 
 Public API: a producer-side on-demand group request surface in `@moq/net`,
 matching Rust's lifecycle. Wire: implement the existing supported IETF FETCH

quest/next/js-hotpath/publish-audio.md

@@ -30,7 +30,7 @@ Choose a group duration only from those comparisons; retain one-frame groups
 if the overhead saving does not justify the latency or loss tradeoff.
 
 Add CI regressions for final-group completion, discontinuity, late join,
-reconnect, and audio delivery on stream-only sessions. Run smoke-full for the
+reconnect, and audio delivery on stream-only sessions. Run `smoke --all` for the
 cross-language delivery paths. Reuse the audio quality harness's metrics and
 impairment fixtures when available rather than creating a second definition.
 

quest/next/mobile/ffi-video-consumer.md

@@ -28,7 +28,7 @@ owns mobile capture; native mobile surface views remain deferred.
   its usable wrapper, codec-enabled artifacts, and decode proof to Dart codec
   parity. Document the temporary omission rather than shipping a stub.
 - Verify on an iOS simulator and an Android emulator with the smoke media,
-  and on macOS through libmoq so `just test smoke-full` covers it.
+  and on macOS through libmoq so `just test smoke --all` covers it.
 
 Dart codec parity requires this quest and owns the remaining Dart integration.
 This quest must not require Dart artifact enablement, which would create a cycle.