@devrouter/cli 0.0.58 → 0.0.60

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@devrouter/cli",
3
- "version": "0.0.58",
3
+ "version": "0.0.60",
4
4
  "description": "Local dev routing CLI with shared Traefik reverse proxy",
5
5
  "author": "Roland Schlaefli",
6
6
  "homepage": "https://github.com/rschlaefli/devrouter#readme",
@@ -48,6 +48,7 @@
48
48
  "test": "vitest run && pnpm test:process",
49
49
  "test:package": "./scripts/package-smoke.sh",
50
50
  "qualify:lifecycle": "node --import tsx scripts/qualify-lifecycle.ts",
51
+ "qualify:controller": "node --import tsx scripts/qualify-controller.ts",
51
52
  "test:process": "bash scripts/test-devrouter-process.sh",
52
53
  "check": "biome check .",
53
54
  "check:fix": "biome check --write .",
@@ -0,0 +1,34 @@
1
+ # Upgrade to devrouter 0.0.59
2
+
3
+ This patch lets short-lived preparation children finish naturally after the
4
+ foreground command exits successfully. The helper holds its lifecycle lock and
5
+ checks cancellation during a bounded two-second drain window. Persistent children
6
+ still prevent application startup and use existing owned-group cleanup.
7
+
8
+ ## Task
9
+
10
+ Install the published patch and update `.devrouter.yml` devrouter.version to
11
+ 0.0.59 through the consumer workflow. No schema or dependency migration is needed.
12
+ The next authorized ensure delivers the matching runtime helper automatically.
13
+ Keep preparation synchronous and avoid detached sessions or process groups.
14
+
15
+ ## Validation
16
+
17
+ On an authorized exact checkout, verify preparation finishes and the intended
18
+ profile starts. Test application readiness independently from successful build
19
+ output. Preserve data, stop task-owned runtimes and verify stopped provider status
20
+ and zero exact routes. Do not start another task's runtime to repeat qualification.
21
+
22
+ ## Limits
23
+
24
+ An isolated Linux fixture proves natural drainage of a short-lived child and
25
+ continued rejection of persistent children. The reported Turbo descendant has
26
+ not been identified, so this patch alone does not prove that consumer startup
27
+ succeeds. Warm mount reconciliation, capacity admission and complete OOM protection
28
+ remain outside this patch.
29
+
30
+ ## Report
31
+
32
+ Report installed version, exact checkout/profile, preparation outcome, application
33
+ readiness and final runtime state. Include remaining blockers without treating
34
+ cached build success as successful environment startup.
@@ -0,0 +1,40 @@
1
+ # Upgrade to devrouter 0.0.60
2
+
3
+ This patch retains a coherent degraded candidate after an external reset and
4
+ failed recreation when exact ownership, complete service population and unchanged
5
+ configuration are proven. Existing ordinary ensure can recover that record.
6
+ It also includes foreground controller observation of independent infrastructure
7
+ and application readiness, with bounded subscriptions and durable session fencing.
8
+
9
+ ## Task
10
+
11
+ Install the published patch through the existing consumer workflow and update
12
+ `.devrouter.yml` devrouter.version to 0.0.60. No schema migration is required.
13
+ Refresh distributed agent guidance with the repository's established workflow.
14
+ Use ordinary ensure on the authorized exact checkout; preserve data and recorded
15
+ ownership. Never manufacture fingerprints or adopt another checkout's containers.
16
+
17
+ ## Validation
18
+
19
+ Verify the requested profile, exact provider/container ownership, preparation and
20
+ application readiness separately. If recreation fails, inspect whether the exact
21
+ candidate was proven and retained as degraded. Retry ordinary ensure through its
22
+ supported recovery path. Incomplete proof remains an explicit diagnostic.
23
+
24
+ For ongoing observation, use the documented foreground controller session flow;
25
+ verify infrastructure readiness remains distinct from an application error.
26
+ Stop the exact task runtime after qualification and verify stopped provider state
27
+ and zero exact routes, retaining data.
28
+
29
+ ## Limits
30
+
31
+ Source and installed synthetic-provider tests cover these contracts. They do not
32
+ prove the original other-host reset incident, complete OOM protection, capacity
33
+ admission or autonomous agent recovery. Controller observation does not repair or
34
+ schedule workloads. Live acceptance requires the existing runtime owner's evidence.
35
+
36
+ ## Report
37
+
38
+ Report installed version, exact checkout/profile, ownership and recovery results,
39
+ application versus tooling readiness, retained data and final stopped/route state.
40
+ Identify incomplete or synthetic-only evidence without expanding release claims.