@golba98/codexa 1.0.1 → 1.0.3

package/src/core/providerLauncher/launcher.test.ts

@@ -1,238 +0,0 @@
-import assert from "node:assert/strict";
-import { EventEmitter } from "events";
-import test from "node:test";
-import { buildProviderLaunchSpec, launchProviderCli } from "./launcher.js";
-import type { ProviderConfig } from "./types.js";
-
-function makeProvider(overrides: Partial<ProviderConfig> = {}): ProviderConfig {
-  return {
-    id: "openai",
-    displayName: "OpenAI",
-    currentModel: "gpt-5.4",
-    backendType: "codex-cli-auth",
-    routeMode: "in-codexa",
-    enabled: true,
-    statusLabel: "Enabled",
-    launchCommand: { executable: "codex", args: [] },
-    isDefault: true,
-    isActiveRoute: true,
-    routeUnavailableReason: null,
-    ...overrides,
-  };
-}
-
-test("builds launch specs for enabled providers", () => {
-  const spec = buildProviderLaunchSpec(makeProvider({
-    launchCommand: { executable: "claude", args: ["--resume"] },
-  }), "C:\\Workspace");
-
-  assert.equal("status" in spec, false);
-  if ("status" in spec) return;
-  assert.equal(spec.executable, "claude");
-  assert.deepEqual(spec.args, ["--resume"]);
-  assert.equal(spec.cwd, "C:\\Workspace");
-});
-
-test("disabled providers fail before spawning", () => {
-  const result = buildProviderLaunchSpec(makeProvider({
-    id: "local",
-    displayName: "Local",
-    backendType: "local-openai-compatible",
-    enabled: false,
-    statusLabel: "Disabled",
-    launchCommand: null,
-    isDefault: false,
-  }), "C:\\Workspace");
-
-  assert.equal("status" in result, true);
-  assert.equal("status" in result ? result.status : "", "disabled");
-  assert.match("status" in result ? result.message : "", /Configure a command/i);
-});
-
-test("unsafe configured launch commands fail before spawning", () => {
-  const result = buildProviderLaunchSpec(makeProvider({
-    launchCommand: { executable: "codex & calc", args: [] },
-  }), "C:\\Workspace");
-
-  assert.equal("status" in result, true);
-  assert.equal("status" in result ? result.status : "", "spawn-error");
-  assert.match("status" in result ? result.message : "", /unsafe launch command/i);
-});
-
-test("missing command spawn errors become friendly launch results", async () => {
-  const child = new EventEmitter();
-  const spawnImpl = (() => {
-    queueMicrotask(() => {
-      const error = new Error("not found") as NodeJS.ErrnoException;
-      error.code = "ENOENT";
-      child.emit("error", error);
-    });
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace",
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "missing-command");
-  assert.match(result.message, /codex.*PATH/i);
-});
-
-test("missing command preflight fails before suspending raw mode", async () => {
-  const rawModes: boolean[] = [];
-  let didSpawn = false;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace",
-    stdin: {
-      isRaw: true,
-      setRawMode(enabled) {
-        rawModes.push(enabled);
-      },
-    },
-    commandExists: () => false,
-    spawnImpl: (() => {
-      didSpawn = true;
-      return new EventEmitter();
-    }) as unknown as typeof import("child_process").spawn,
-  });
-
-  assert.equal(result.status, "missing-command");
-  assert.equal(didSpawn, false);
-  assert.deepEqual(rawModes, []);
-});
-
-test("launch restores raw mode after child exits", async () => {
-  const child = new EventEmitter();
-  const rawModes: boolean[] = [];
-  const spawnImpl = (() => {
-    queueMicrotask(() => child.emit("close", 0, null));
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace",
-    stdin: {
-      isRaw: true,
-      setRawMode(enabled) {
-        rawModes.push(enabled);
-      },
-    },
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "completed");
-  assert.deepEqual(rawModes, [false, true]);
-});
-
-test("launch passes the workspace root as the child cwd", async () => {
-  const child = new EventEmitter();
-  let observedCwd = "";
-  let observedShell: boolean | undefined = undefined;
-  const spawnImpl = ((_executable: string, _args: string[], options: { cwd?: string; shell?: boolean }) => {
-    observedCwd = options.cwd ?? "";
-    observedShell = options.shell;
-    queueMicrotask(() => child.emit("close", 0, null));
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace\\Project",
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "completed");
-  assert.equal(observedCwd, "C:\\Workspace\\Project");
-  assert.equal(observedShell, false);
-});
-
-test("launch wraps Windows batch commands without enabling shell mode", async () => {
-  if (process.platform !== "win32") return;
-
-  const child = new EventEmitter();
-  let observedExecutable = "";
-  let observedArgs: string[] = [];
-  let observedShell: boolean | undefined = undefined;
-  const spawnImpl = ((executable: string, args: string[], options: { shell?: boolean }) => {
-    observedExecutable = executable;
-    observedArgs = args;
-    observedShell = options.shell;
-    queueMicrotask(() => child.emit("close", 0, null));
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider({
-    launchCommand: { executable: "codex.cmd", args: ["--resume"] },
-  }), {
-    cwd: "C:\\Workspace",
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "completed");
-  assert.equal(observedExecutable, "cmd.exe");
-  assert.deepEqual(observedArgs, ["/d", "/s", "/c", "call", "codex.cmd", "--resume"]);
-  assert.equal(observedShell, false);
-});
-
-test("launch runs suspend and resume hooks around raw mode changes", async () => {
-  const child = new EventEmitter();
-  const events: string[] = [];
-  const spawnImpl = (() => {
-    events.push("spawn");
-    queueMicrotask(() => child.emit("close", 0, null));
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace",
-    stdin: {
-      isRaw: true,
-      setRawMode(enabled) {
-        events.push(enabled ? "raw-on" : "raw-off");
-      },
-    },
-    beforeLaunch: () => events.push("before"),
-    afterLaunch: () => events.push("after"),
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "completed");
-  assert.deepEqual(events, ["before", "raw-off", "spawn", "raw-on", "after"]);
-});
-
-test("launch restores terminal state after child spawn error", async () => {
-  const child = new EventEmitter();
-  const events: string[] = [];
-  const spawnImpl = (() => {
-    events.push("spawn");
-    queueMicrotask(() => {
-      const error = new Error("blocked") as NodeJS.ErrnoException;
-      error.code = "EPERM";
-      child.emit("error", error);
-    });
-    return child;
-  }) as unknown as typeof import("child_process").spawn;
-
-  const result = await launchProviderCli(makeProvider(), {
-    cwd: "C:\\Workspace",
-    stdin: {
-      isRaw: true,
-      setRawMode(enabled) {
-        events.push(enabled ? "raw-on" : "raw-off");
-      },
-    },
-    beforeLaunch: () => events.push("before"),
-    afterLaunch: () => events.push("after"),
-    commandExists: () => true,
-    spawnImpl,
-  });
-
-  assert.equal(result.status, "spawn-error");
-  assert.deepEqual(events, ["before", "raw-off", "spawn", "raw-on", "after"]);
-});

package/src/core/providerLauncher/registry.test.ts

@@ -1,324 +0,0 @@
-import assert from "node:assert/strict";
-import test from "node:test";
-import { buildProviderRegistry, getDefaultProviderId } from "./registry.js";
-import { checkLocalProvider, resetLocalProviderStateForTests } from "../providerRuntime/local.js";
-import { resetGeminiRouteValidationCacheForTests } from "../providerRuntime/gemini.js";
-import { resetGeminiExecutableCacheForTests } from "../executables/geminiExecutable.js";
-import { setProviderActiveRoute } from "./workspaceConfig.js";
-import { resolveActiveProviderRoute } from "../providerRuntime/registry.js";
-
-test("provider registry exposes the default launcher providers", () => {
-  const providers = buildProviderRegistry({ activeModel: "gpt-5.4" });
-
-  assert.deepEqual(providers.map((provider) => provider.id), ["openai", "anthropic", "google", "local"]);
-  assert.equal(providers[0]?.displayName, "OpenAI");
-  assert.equal(providers[0]?.currentModel, "gpt-5.4");
-  assert.deepEqual(providers[0]?.launchCommand, { executable: "codex", args: [] });
-  assert.deepEqual(providers[1]?.launchCommand, { executable: "claude", args: [] });
-  assert.deepEqual(providers[2]?.launchCommand, { executable: "gemini", args: [] });
-  assert.equal(providers[3]?.enabled, false);
-  assert.equal(providers[3]?.launchCommand, null);
-});
-
-test("workspace config can set the default provider", () => {
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: { workspaceDefaultProviderId: "anthropic" },
-  });
-
-  assert.equal(getDefaultProviderId({ workspaceDefaultProviderId: "anthropic" }), "anthropic");
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.isDefault, true);
-  assert.equal(providers.find((provider) => provider.id === "openai")?.isDefault, false);
-});
-
-test("provider registry keeps workspace default separate from active route", () => {
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: {
-      workspaceDefaultProviderId: "anthropic",
-      activeRoute: {
-        providerId: "openai",
-        modelId: "gpt-5.4",
-      },
-    },
-  });
-
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.isDefault, true);
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.isActiveRoute, false);
-  assert.equal(providers.find((provider) => provider.id === "openai")?.isDefault, false);
-  assert.equal(providers.find((provider) => provider.id === "openai")?.isActiveRoute, true);
-});
-
-test("unvalidated local provider remains disabled until endpoint discovery succeeds", () => {
-  resetLocalProviderStateForTests();
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: {
-      workspaceDefaultProviderId: "anthropic",
-      activeRoute: {
-        providerId: "local",
-        modelId: "llama-local",
-      },
-    },
-  });
-
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.isDefault, true);
-  assert.equal(providers.find((provider) => provider.id === "local")?.isActiveRoute, true);
-  assert.equal(providers.find((provider) => provider.id === "local")?.routeMode, "in-codexa");
-  assert.equal(providers.find((provider) => provider.id === "local")?.enabled, false);
-});
-
-test("google can be selected as an active in-Codexa route", () => {
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: {
-      activeRoute: {
-        providerId: "google",
-        modelId: "gemini-3-flash-preview",
-        backendKind: "gemini-cli-auth",
-      },
-    },
-  });
-
-  assert.equal(providers.find((provider) => provider.id === "google")?.isActiveRoute, true);
-  assert.equal(providers.find((provider) => provider.id === "google")?.routeMode, "in-codexa");
-  assert.equal(providers.find((provider) => provider.id === "google")?.currentModel, "gemini-3-flash-preview");
-  assert.equal(providers.find((provider) => provider.id === "openai")?.isActiveRoute, false);
-});
-
-test("anthropic can be selected as an active in-Codexa route", () => {
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: {
-      activeRoute: {
-        providerId: "anthropic",
-        modelId: "claude-sonnet-4-20250514",
-        backendKind: "anthropic-api-key",
-      },
-    },
-  });
-
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.isActiveRoute, true);
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.routeMode, "in-codexa");
-  assert.equal(providers.find((provider) => provider.id === "anthropic")?.currentModel, "claude-sonnet-4-20250514");
-  assert.equal(providers.find((provider) => provider.id === "openai")?.isActiveRoute, false);
-});
-
-
-test("discovered local models enable local provider and display selected model", async () => {
-  resetLocalProviderStateForTests();
-  const localOverride = {
-    currentModel: "google/gemma-4-26b-a4b",
-  };
-  await checkLocalProvider({
-    override: localOverride,
-    fetchImpl: (async (input) => {
-      if (String(input).includes("/api/v0/")) {
-        return new Response(null, { status: 404 });
-      }
-      return new Response(JSON.stringify({
-        data: [{ id: "google/gemma-4-26b-a4b" }],
-      }), { status: 200 });
-    }) as typeof fetch,
-  });
-
-  const providers = buildProviderRegistry({
-    activeModel: "gpt-5.4",
-    workspaceConfig: {
-      activeRoute: {
-        providerId: "local",
-        modelId: "google/gemma-4-26b-a4b",
-        backendKind: "local-openai-compatible",
-      },
-      providers: {
-        local: localOverride,
-      },
-    },
-  });
-
-  const local = providers.find((provider) => provider.id === "local");
-  assert.equal(local?.enabled, true);
-  assert.equal(local?.currentModel, "google/gemma-4-26b-a4b");
-  assert.equal(local?.backendType, "local-openai-compatible");
-  assert.equal(local?.statusLabel, "Enabled");
-  assert.deepEqual(local?.launchCommand, null);
-  resetLocalProviderStateForTests();
-});
-
-test("LM Studio loaded Local model replaces stale active route in provider registry", async () => {
-  resetLocalProviderStateForTests();
-  try {
-    await checkLocalProvider({
-      override: {
-        currentModel: "google/gemma-4-26b-a4b",
-        defaultModel: "google/gemma-4-26b-a4b",
-        baseUrl: "http://localhost:1234/v1",
-      },
-      fetchImpl: (async (input) => {
-        if (String(input).includes("/api/v0/")) {
-          return new Response(JSON.stringify({
-            data: [{
-              id: "qwen/qwen3.6-27b",
-              state: "loaded",
-              loaded_context_length: 32000,
-              max_context_length: 262144,
-              capabilities: ["tool_use"],
-            }],
-            object: "list",
-          }), { status: 200 });
-        }
-        return new Response(JSON.stringify({
-          data: [{ id: "google/gemma-4-26b-a4b" }, { id: "qwen/qwen3.6-27b" }],
-        }), { status: 200 });
-      }) as typeof fetch,
-    });
-
-    const providers = buildProviderRegistry({
-      activeModel: "gpt-5.4",
-      workspaceConfig: {
-        activeRoute: {
-          providerId: "local",
-          modelId: "google/gemma-4-26b-a4b",
-          backendKind: "local-openai-compatible",
-        },
-        providers: {
-          local: {
-            currentModel: "google/gemma-4-26b-a4b",
-            defaultModel: "google/gemma-4-26b-a4b",
-            baseUrl: "http://localhost:1234/v1",
-          },
-        },
-      },
-    });
-
-    const local = providers.find((provider) => provider.id === "local");
-    assert.equal(local?.currentModel, "qwen/qwen3.6-27b");
-    assert.equal(local?.contextLengthLabel, "32,000");
-    assert.equal(local?.contextLengthSource, "lmstudio-api");
-    assert.equal(local?.capabilityProfile?.supportsToolCalls, true);
-    assert.equal(local?.capabilityProfile?.supportsVision, null);
-  } finally {
-    resetLocalProviderStateForTests();
-  }
-});
-
-// ─── Provider status label / color sync tests ────────────────────────────────
-
-async function withNoGeminiConfig<T>(callback: () => T | Promise<T>): Promise<T> {
-  const savedGemini = process.env.GEMINI_API_KEY;
-  const savedGoogle = process.env.GOOGLE_API_KEY;
-  const savedExe = process.env.GEMINI_EXECUTABLE;
-  const savedCliPath = process.env.GEMINI_CLI_PATH;
-  try {
-    delete process.env.GEMINI_API_KEY;
-    delete process.env.GOOGLE_API_KEY;
-    delete process.env.GEMINI_EXECUTABLE;
-    delete process.env.GEMINI_CLI_PATH;
-    resetGeminiRouteValidationCacheForTests();
-    resetGeminiExecutableCacheForTests();
-    return await callback();
-  } finally {
-    if (savedGemini === undefined) delete process.env.GEMINI_API_KEY; else process.env.GEMINI_API_KEY = savedGemini;
-    if (savedGoogle === undefined) delete process.env.GOOGLE_API_KEY; else process.env.GOOGLE_API_KEY = savedGoogle;
-    if (savedExe === undefined) delete process.env.GEMINI_EXECUTABLE; else process.env.GEMINI_EXECUTABLE = savedExe;
-    if (savedCliPath === undefined) delete process.env.GEMINI_CLI_PATH; else process.env.GEMINI_CLI_PATH = savedCliPath;
-    resetGeminiRouteValidationCacheForTests();
-    resetGeminiExecutableCacheForTests();
-  }
-}
-
-test("google shows 'Needs config' statusLabel when no API key or CLI validation is present", async () => {
-  await withNoGeminiConfig(() => {
-    const providers = buildProviderRegistry({ activeModel: "gpt-5.4" });
-    const google = providers.find((p) => p.id === "google");
-    assert.equal(google?.statusLabel, "Needs config");
-    assert.notEqual(google?.routeUnavailableReason, null);
-  });
-});
-
-test("google shows 'Enabled' statusLabel when GEMINI_API_KEY is set", async () => {
-  const saved = process.env.GEMINI_API_KEY;
-  try {
-    process.env.GEMINI_API_KEY = "test-key";
-    resetGeminiRouteValidationCacheForTests();
-    const providers = buildProviderRegistry({ activeModel: "gpt-5.4" });
-    const google = providers.find((p) => p.id === "google");
-    assert.equal(google?.statusLabel, "Enabled");
-    assert.equal(google?.routeUnavailableReason, null);
-  } finally {
-    if (saved === undefined) delete process.env.GEMINI_API_KEY; else process.env.GEMINI_API_KEY = saved;
-    resetGeminiRouteValidationCacheForTests();
-  }
-});
-
-test("google 'Needs config' survives workspace config override with no enabled field", async () => {
-  await withNoGeminiConfig(() => {
-    const providers = buildProviderRegistry({
-      activeModel: "gpt-5.4",
-      workspaceConfig: {
-        providers: {
-          google: { currentModel: "gemini-3-flash-preview", currentReasoning: "medium" },
-        },
-      },
-    });
-    const google = providers.find((p) => p.id === "google");
-    assert.equal(google?.statusLabel, "Needs config");
-    assert.notEqual(google?.routeUnavailableReason, null);
-  });
-});
-
-test("'Active' text is display-layer only — registry statusLabel stays 'Enabled' for active route when configured", async () => {
-  const saved = process.env.GEMINI_API_KEY;
-  try {
-    process.env.GEMINI_API_KEY = "test-key";
-    resetGeminiRouteValidationCacheForTests();
-    const providers = buildProviderRegistry({
-      activeModel: "gpt-5.4",
-      workspaceConfig: {
-        activeRoute: { providerId: "google", modelId: "gemini-3-flash-preview", backendKind: "gemini-cli-auth" },
-      },
-    });
-    const google = providers.find((p) => p.id === "google");
-    assert.equal(google?.isActiveRoute, true);
-    assert.equal(google?.statusLabel, "Enabled");
-  } finally {
-    if (saved === undefined) delete process.env.GEMINI_API_KEY; else process.env.GEMINI_API_KEY = saved;
-    resetGeminiRouteValidationCacheForTests();
-  }
-});
-
-test("resolveActiveProviderRoute returns google route when configured via API key", async () => {
-  const saved = process.env.GEMINI_API_KEY;
-  try {
-    process.env.GEMINI_API_KEY = "test-key";
-    resetGeminiRouteValidationCacheForTests();
-    const route = resolveActiveProviderRoute({
-      workspaceConfigActiveRoute: {
-        providerId: "google",
-        modelId: "gemini-3-flash-preview",
-        backendKind: "gemini-cli-auth",
-      },
-      currentModel: "gpt-5.4",
-      currentReasoning: "medium",
-    });
-    assert.equal(route.providerId, "google");
-    assert.equal(route.modelId, "gemini-3-flash-preview");
-  } finally {
-    if (saved === undefined) delete process.env.GEMINI_API_KEY; else process.env.GEMINI_API_KEY = saved;
-    resetGeminiRouteValidationCacheForTests();
-  }
-});
-
-test("setProviderActiveRoute does not update activeRoute when google is not configured", async () => {
-  await withNoGeminiConfig(() => {
-    const original: import("./types.js").ProviderWorkspaceConfig = {
-      activeRoute: { providerId: "openai", modelId: "gpt-5.4", backendKind: "codex-cli-auth" },
-    };
-    const result = setProviderActiveRoute(original, {
-      providerId: "google",
-      modelId: "gemini-3-flash-preview",
-      backendKind: "gemini-cli-auth",
-    });
-    assert.equal(result.activeRoute?.providerId, "openai");
-  });
-});