cursedops 0.16.0 → 0.16.1

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "cursedops",
-	"version": "0.16.0",
+	"version": "0.16.1",
 	"description": "The build-and-ops answers this generation's apps wrote independently and identically: finding a generation's roots — and printing a command that runs when pasted — without knowing a path, the generation's whole-tree laws run over one repo from a checkout or a worktree, macOS launchd agent install/replace/remove and the live port a job serves, the scaffolding and verdicts of a deployed smoke (origin probe, the smoke's own environment, a network that lies about DNS, a settled version), the static-serving helpers eight apps copied — the path-traversal guard among them — the API floor that keeps an unmatched /api/... from ever being answered with the app shell, the commit and dirty flag a checkout-served process reports, the Cloudflare Worker deploy toolkit four apps copied (the deploy sequence, exact-set secrets over a pipe, origin-first rollback, the curl edge fetch, the row-for-row D1 import proof, the billed-CPU tail check around a deploy's walk and smoke, and each app's worker:secrets and worker:smoke main as one function of its data), the relay a Worker fronts a Mac-bound app with (the Durable Object, the frames, the Mac's dialer and key rotation — lifted from station for roms — and the signed-in stage walk's skeleton and the relay app's whole worker:deploy), the Worker import-graph and await-port checks every Worker app's suite runs over its own source, and the public-surface ratchet three published libraries each carried a forked copy of. Mechanism only — no app knows its name from here. Bun, zero runtime dependencies (typescript is an optional peer, for public-surface only), ships source.",
 	"type": "module",
 	"scripts": {

package/src/workerDeploy.ts

@@ -54,7 +54,8 @@
  * refusal without wrangler, git or a network.
  */
 import { spawn, spawnSync } from "node:child_process";
-import { existsSync, readdirSync, readFileSync } from "node:fs";
+import { existsSync, mkdtempSync, readdirSync, readFileSync, rmSync } from "node:fs";
+import { tmpdir } from "node:os";
 import { join } from "node:path";
 import { liveColumns, liveColumnsQueries, missingColumnStatements, schemaTables } from "cursedops/d1-columns";
 import { wranglerRows } from "cursedops/d1-import";
@@ -345,6 +346,19 @@ export interface DeployDeps {
 	 * generation's lease; absent (or `null` with no state root), nothing is leased.
 	 */
 	lease?: (app: string, what: string) => StageLeaseResult | null;
+	/**
+	 * A fresh file for wrangler's ND-JSON output (`WRANGLER_OUTPUT_FILE_PATH`), which is where a failed
+	 * `wrangler deploy` says WHY in a form a program can read ({@link transientDeployFailure}).
+	 * {@link workerDeployDeps} makes one under the OS temp dir; absent, a failed deploy is not retried.
+	 */
+	wranglerOutput?: () => WranglerOutputFile;
+}
+
+/** One `wrangler` run's output file: the env that points wrangler at it, its text, and its removal. */
+export interface WranglerOutputFile {
+	env: Record<string, string>;
+	read: () => string;
+	dispose: () => void;
 }
 
 /**
@@ -420,6 +434,15 @@ export function workerDeployDeps(cwd: string): DeployDeps {
 		},
 		git: (args) => (spawnSync("git", [...args], { cwd, encoding: "utf8" }).stdout ?? "").trim(),
 		read: (path) => readFileSync(join(cwd, path), "utf8"),
+		wranglerOutput: () => {
+			const dir = mkdtempSync(join(tmpdir(), "cursedops-wrangler-"));
+			const path = join(dir, "output.ndjson");
+			return {
+				env: { WRANGLER_OUTPUT_FILE_PATH: path },
+				read: () => (existsSync(path) ? readFileSync(path, "utf8") : ""),
+				dispose: () => rmSync(dir, { recursive: true, force: true }),
+			};
+		},
 		record: (line) => {
 			recordDeployEvent(forgeState(cwd), line);
 		},
@@ -572,6 +595,38 @@ export function wranglerError(stdout: string): { reason: string; api: boolean }
 	return { reason, api: name === "APIError" || (kind === "error" && /cloudflare api/i.test(reason)) };
 }
 
+/**
+ * Why a failed `wrangler deploy` is worth ONE more try — or `null` when it would fail the same way.
+ *
+ * Measured 2026-10-11 (forge task 3053): roms' production deploy uploaded the Worker and its routes,
+ * then Cloudflare answered the cron-schedules PUT with `Internal authentication error: internal
+ * server error [code: 10500]`. wrangler exited 1, the ship's roms lane went red, and the ship batch
+ * re-ran a 50-minute `ship run` to redo a step that a second request would have done. A deploy is
+ * idempotent — the same script, the same triggers — so a Cloudflare-side failure is retried once.
+ * A bundling error, a refused binding or a script that throws at startup (10021) is not: wrangler's
+ * own `command-failed` line ({@link WranglerOutputFile}) must name the server's fault or a retry-after.
+ */
+export function transientDeployFailure(ndjson: string): string | null {
+	type Failed = { code?: unknown; message?: unknown; retry_after_ms?: unknown };
+	let failed: Failed | null = null;
+	for (const line of ndjson.split("\n")) {
+		if (!line.trim()) continue;
+		try {
+			const entry = JSON.parse(line) as (Failed & { type?: unknown }) | null;
+			if (entry?.type === "command-failed") failed = entry;
+		} catch {
+			// a torn line from a killed run — the ones around it still count
+		}
+	}
+	if (!failed) return null;
+	const message = typeof failed.message === "string" ? failed.message : "";
+	const first = message.split("\n").find((l) => l.trim()) ?? "";
+	const reason = (detail: string) => `${first.trim() || "wrangler deploy failed"} — ${detail}`;
+	if (typeof failed.retry_after_ms === "number") return reason(`Cloudflare asked for a retry after ${failed.retry_after_ms} ms`);
+	const server = message.match(/internal server error|\[code: (?:10500|10013|7403)\]|\b50[234]\b|ETIMEDOUT|ECONNRESET|socket hang up|fetch failed/i);
+	return server ? reason(`"${server[0]}" is Cloudflare's side`) : null;
+}
+
 /**
  * Did one schema chunk APPLY? Its exit code alone does not say.
  *
@@ -895,7 +950,22 @@ export function runWorkerDeploy(spec: WorkerDeploySpec, deps: DeployDeps): Deplo
 		? ["--var", `${spec.stampVars.commit}:${commit}`, "--var", `${spec.stampVars.dirty}:${shipsDirty ? "true" : "false"}`]
 		: [];
 	enter("upload");
-	if (deps.run(["bunx", "wrangler", "deploy", ...envArgs, ...stamp], spec.credential) !== 0) return stop("deploy", "wrangler deploy failed.");
+	const deployArgv = ["bunx", "wrangler", "deploy", ...envArgs, ...stamp];
+	const deployOnce = (): { code: number; transient: string | null } => {
+		const output = deps.wranglerOutput?.();
+		try {
+			const code = deps.run(deployArgv, { ...spec.credential, ...output?.env });
+			return { code, transient: code !== 0 && output ? transientDeployFailure(output.read()) : null };
+		} finally {
+			output?.dispose();
+		}
+	};
+	let deployed = deployOnce();
+	if (deployed.code !== 0 && deployed.transient) {
+		error(`  🟡 ${deployed.transient} — retrying the deploy once`);
+		deployed = deployOnce();
+	}
+	if (deployed.code !== 0) return stop("deploy", "wrangler deploy failed.");
 	if (spec.secrets) {
 		step("attach the secrets — exactly this deployment's set, over a pipe");
 		enter("secrets");