@golba98/codexa 1.0.2 → 1.0.4

package/src/ui/BottomComposer.test.ts

@@ -1,674 +0,0 @@
-import assert from "node:assert/strict";
-import test from "node:test";
-import {
-  getCommandSuggestionState,
-  getComposerToFooterGapRows,
-  getComposerPersona,
-  getTokenBarDisplay,
-  getVisibleComposerStatusLine,
-  measureBottomComposerRows,
-} from "./BottomComposer.js";
-import { createLayoutSnapshot } from "./layout.js";
-import { getSlashCommandSuggestions } from "./slashCommands.js";
-import type { PendingModelSpec, VerifiedModelSpec } from "../core/models/modelSpecs.js";
-
-test("maps the idle state to the idle composer persona", () => {
-  assert.equal(getComposerPersona({ kind: "IDLE" }), "idle");
-});
-
-test("maps busy, answer, and error states to the right personas", () => {
-  assert.equal(getComposerPersona({ kind: "THINKING", turnId: 1 }), "busy");
-  assert.equal(getComposerPersona({ kind: "RESPONDING", turnId: 1 }), "busy");
-  assert.equal(getComposerPersona({ kind: "SHELL_RUNNING", shellId: 7 }), "busy");
-  assert.equal(getComposerPersona({ kind: "AWAITING_USER_ACTION", turnId: 2, question: "Need Redis?" }), "answer");
-  assert.equal(getComposerPersona({ kind: "ERROR", turnId: 3, message: "Boom" }), "error");
-});
-
-test("measures the standard composer rows from the rendered prompt state", () => {
-  const rows = measureBottomComposerRows({
-    layout: createLayoutSnapshot(100, 30),
-    uiState: { kind: "IDLE" },
-    mode: "auto-edit",
-    model: "gpt-5.4",
-    reasoningLevel: "medium",
-    tokensUsed: 1200,
-    value: "",
-    cursor: 0,
-  });
-
-  assert.equal(rows, 7);
-});
-
-test("adds a measured footer gap only when the viewport has room", () => {
-  assert.equal(getComposerToFooterGapRows(createLayoutSnapshot(100, 30)), 1);
-  assert.equal(getComposerToFooterGapRows(createLayoutSnapshot(100, 24)), 0);
-  assert.equal(getComposerToFooterGapRows(createLayoutSnapshot(39, 30)), 0);
-});
-
-test("uses the run footer row budget in cramped busy viewports", () => {
-  const rows = measureBottomComposerRows({
-    layout: createLayoutSnapshot(80, 24),
-    uiState: { kind: "THINKING", turnId: 1 },
-    value: "",
-    cursor: 0,
-  });
-
-  assert.equal(rows, 4);
-});
-
-test("does not render an exact slash command draft as a suggestion row", () => {
-  const exact = getCommandSuggestionState({
-    value: "/clear",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.equal(exact.showSuggestions, true);
-  assert.equal(exact.reserveSuggestionRow, true);
-  assert.deepEqual(exact.suggestions.map((suggestion) => suggestion.cmd), []);
-});
-
-test("keeps partial slash command suggestions visible", () => {
-  const partial = getCommandSuggestionState({
-    value: "/clea",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.equal(partial.showSuggestions, true);
-  assert.equal(partial.reserveSuggestionRow, true);
-  assert.deepEqual(partial.suggestions.map((suggestion) => suggestion.cmd), ["/clear"]);
-});
-
-test("surfaces the provider picker suggestion for root prefixes and alias input", () => {
-  const rootSuggestions = getCommandSuggestionState({
-    value: "/",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.equal(rootSuggestions.showSuggestions, true);
-  assert.ok(rootSuggestions.suggestions.map((suggestion) => suggestion.cmd).includes("/providers"));
-
-  const pSuggestions = getCommandSuggestionState({
-    value: "/p",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.ok(pSuggestions.suggestions.map((suggestion) => suggestion.cmd).includes("/providers"));
-
-  const shortProviderSuggestions = getCommandSuggestionState({
-    value: "/pro",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.ok(shortProviderSuggestions.suggestions.map((suggestion) => suggestion.cmd).includes("/providers"));
-
-  const prefixProviderSuggestions = getCommandSuggestionState({
-    value: "/pr",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.ok(prefixProviderSuggestions.suggestions.map((suggestion) => suggestion.cmd).includes("/providers"));
-
-  const providerSuggestions = getCommandSuggestionState({
-    value: "/provider",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.equal(providerSuggestions.showSuggestions, true);
-  assert.deepEqual(providerSuggestions.suggestions.map((suggestion) => suggestion.cmd), ["/providers"]);
-
-  const exactProviderSuggestions = getCommandSuggestionState({
-    value: "/providers",
-    allowCommands: true,
-    inputLocked: false,
-  });
-
-  assert.equal(exactProviderSuggestions.showSuggestions, true);
-  assert.deepEqual(exactProviderSuggestions.suggestions.map((suggestion) => suggestion.cmd), ["/providers"]);
-
-  const aliasMetadata = getSlashCommandSuggestions("/provider").find((suggestion) => suggestion.cmd === "/providers");
-  assert.deepEqual(aliasMetadata?.aliases, ["/provider"]);
-});
-
-test("keeps exact and partial slash command row budgets stable", () => {
-  const layout = createLayoutSnapshot(100, 30);
-  const base = {
-    layout,
-    uiState: { kind: "IDLE" } as const,
-    value: "",
-    cursor: 0,
-  };
-
-  const partialRows = measureBottomComposerRows({
-    ...base,
-    value: "/clea",
-    cursor: "/clea".length,
-  });
-  const exactRows = measureBottomComposerRows({
-    ...base,
-    value: "/clear",
-    cursor: "/clear".length,
-  });
-
-  assert.equal(exactRows, partialRows);
-});
-
-test("suppresses completed response status while a slash command draft is active", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "ANSWER_VISIBLE", turnId: 1 },
-      value: "/clear",
-      allowCommands: true,
-    }),
-    "",
-  );
-
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "ANSWER_VISIBLE", turnId: 1 },
-      value: "next prompt",
-      allowCommands: true,
-    }),
-    "✧ Codexa response complete",
-  );
-});
-
-test("shows Gemini-specific status when THINKING with google provider at 0 seconds", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 0,
-    }),
-    "Starting Gemini CLI",
-  );
-});
-
-test("includes elapsed timer in Gemini status after first second", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 3,
-    }),
-    "Starting Gemini CLI  00:03",
-  );
-});
-
-test("shows reassurance message in Gemini status at 5 seconds", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 5,
-    }),
-    "Gemini CLI is still starting. The upstream CLI can take a moment  00:05",
-  );
-});
-
-test("shows still waiting message in Gemini status at 15 seconds", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 15,
-    }),
-    "Still waiting for Gemini CLI  00:15",
-  );
-});
-
-test("shows generic thinking status for unknown/local provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "local",
-      runElapsedSeconds: 10,
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("shows Starting Claude Code at 0 seconds for anthropic provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 0,
-    }),
-    "Starting Claude Code",
-  );
-});
-
-test("includes elapsed timer in Claude Code status after first second", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 3,
-    }),
-    "Starting Claude Code  00:03",
-  );
-});
-
-test("shows reassurance message at 5 seconds for Claude Code", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 5,
-    }),
-    "Claude Code is still starting. The upstream CLI can take a moment  00:05",
-  );
-});
-
-test("shows still waiting message at 15 seconds for Claude Code", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 15,
-    }),
-    "Still waiting for Claude Code  00:15",
-  );
-});
-
-test("shows Starting Codex CLI at 0 seconds for openai provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "openai",
-      runElapsedSeconds: 0,
-    }),
-    "Starting Codex CLI",
-  );
-});
-
-test("includes elapsed timer in Codex CLI status after first second", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "openai",
-      runElapsedSeconds: 3,
-    }),
-    "Starting Codex CLI  00:03",
-  );
-});
-
-test("shows Gemini ready status when RESPONDING with google provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "RESPONDING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-    }),
-    "✧ Gemini ready",
-  );
-});
-
-test("shows Claude ready status when RESPONDING with anthropic provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "RESPONDING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-    }),
-    "✧ Claude ready",
-  );
-});
-
-test("shows Codex ready status when RESPONDING with openai provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "RESPONDING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "openai",
-    }),
-    "✧ Codex ready",
-  );
-});
-
-test("shows generic thinking status when RESPONDING with unknown provider", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "RESPONDING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "local",
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("shows error message in status for google provider in ERROR state regardless of elapsed time", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "ERROR", turnId: 1, message: "Gemini CLI failed to start" },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 30,
-    }),
-    "Gemini CLI failed to start",
-  );
-});
-
-test("shows error message in status for anthropic provider in ERROR state regardless of elapsed time", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "ERROR", turnId: 1, message: "Claude Code failed to start" },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 30,
-    }),
-    "Claude Code failed to start",
-  );
-});
-
-test("getTokenBarDisplay returns null percentage (not 0) for an unknown spec", () => {
-  const spec: PendingModelSpec = {
-    status: "unknown",
-    contextWindow: null,
-    maxOutputTokens: null,
-    sourceUrl: "",
-    verifiedAt: null,
-    error: null,
-  };
-  const display = getTokenBarDisplay(5_000, spec);
-  assert.equal(display.percentage, null, "percentage must be null, not 0, for unknown specs");
-  assert.equal(display.usedText, "Context");
-  assert.equal(display.limitText, "Unknown");
-  assert.equal(display.isEstimatedLimit, false);
-  assert.equal(display.hasKnownLimit, false);
-});
-
-test("getTokenBarDisplay returns correct non-null percentage for a verified spec", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 200_000,
-    maxOutputTokens: 64_000,
-    sourceUrl: "",
-    verifiedAt: 0,
-  };
-  const display = getTokenBarDisplay(10_000, spec);
-  assert.equal(display.percentage, 5);
-  assert.notEqual(display.percentage, null);
-  assert.equal(display.isEstimatedLimit, false);
-  assert.equal(display.hasKnownLimit, true);
-  assert.equal(display.usedText, "10,000", "usedText should be exact number with thousands separator");
-});
-
-test("getTokenBarDisplay formats refreshed LM Studio context meter", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 32_000,
-    maxOutputTokens: 32_000,
-    sourceUrl: "lmstudio-api",
-    verifiedAt: 0,
-  };
-  const display = getTokenBarDisplay(0, spec);
-  assert.equal(display.usedText, "0");
-  assert.equal(display.limitText, "32,000");
-  assert.equal(display.percentage, 0);
-  assert.equal(display.hasKnownLimit, true);
-});
-
-test("getTokenBarDisplay does not add ~ prefix for a documented verified context window", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 200_000,
-    maxOutputTokens: 64_000,
-    sourceUrl: "",
-    verifiedAt: 0,
-  };
-  const display = getTokenBarDisplay(10_000, spec);
-  assert.equal(display.isEstimatedLimit, false);
-  assert.ok(!display.limitText.startsWith("~"), "verified limitText must not start with ~");
-});
-
-// ─── externalCliStatus: provider readiness gate ───────────────────────────────
-
-test("shows 'Codexa is thinking' (not startup message) when provider is ready and THINKING — google", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 2 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 0,
-      externalCliStatus: "ready",
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("shows 'Codexa is thinking' even at 20 seconds elapsed when provider is ready — google", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 2 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 20,
-      externalCliStatus: "ready",
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("shows 'Codexa is thinking' (not startup message) when provider is ready and THINKING — anthropic", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 2 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "anthropic",
-      runElapsedSeconds: 0,
-      externalCliStatus: "ready",
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("shows 'Codexa is thinking' (not startup message) when provider is ready and THINKING — openai", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 2 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "openai",
-      runElapsedSeconds: 0,
-      externalCliStatus: "ready",
-    }),
-    "✧ Codexa is thinking",
-  );
-});
-
-test("still shows startup messages when externalCliStatus is 'starting' — google at 0 seconds", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 0,
-      externalCliStatus: "starting",
-    }),
-    "Starting Gemini CLI",
-  );
-});
-
-test("still shows 'Still waiting' when externalCliStatus is 'starting' at 15 seconds — google", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 15,
-      externalCliStatus: "starting",
-    }),
-    "Still waiting for Gemini CLI  00:15",
-  );
-});
-
-test("still shows startup messages when externalCliStatus is 'idle' (first prompt, not yet starting)", () => {
-  assert.equal(
-    getVisibleComposerStatusLine({
-      uiState: { kind: "THINKING", turnId: 1 },
-      value: "",
-      allowCommands: true,
-      activeProviderId: "google",
-      runElapsedSeconds: 0,
-      externalCliStatus: "idle",
-    }),
-    "Starting Gemini CLI",
-  );
-});
-
-test("regression: second prompt with ready provider never shows 'Still waiting for Gemini CLI'", () => {
-  const statusLine = getVisibleComposerStatusLine({
-    uiState: { kind: "THINKING", turnId: 2 },
-    value: "",
-    allowCommands: true,
-    activeProviderId: "google",
-    runElapsedSeconds: 20,
-    externalCliStatus: "ready",
-  });
-  assert.ok(
-    !/Still waiting for Gemini CLI|Starting Gemini CLI|Checking Gemini/i.test(statusLine),
-    `Expected no startup text but got: "${statusLine}"`,
-  );
-});
-
-// ─── getTokenBarDisplay — new format ─────────────────────────────────────────
-
-test("getTokenBarDisplay(0, 64k spec) shows '0' usedText and '64,000' limitText, not Unknown", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 64_000,
-    maxOutputTokens: 64_000,
-    sourceUrl: "",
-    verifiedAt: 0,
-  };
-  const display = getTokenBarDisplay(0, spec);
-  assert.equal(display.hasKnownLimit, true, "64k verified spec must have known limit");
-  assert.equal(display.usedText, "0", "0 tokens used renders as '0'");
-  assert.equal(display.limitText, "64,000");
-  assert.equal(display.percentage, 0);
-});
-
-test("getTokenBarDisplay uses Math.floor — 1999/200000 rounds down to 0%", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 200_000,
-    maxOutputTokens: 200_000,
-    sourceUrl: "",
-    verifiedAt: 0,
-  };
-  const display = getTokenBarDisplay(1_999, spec);
-  // floor(1999/200000*100) = floor(0.9995) = 0; Math.round would give 1
-  assert.equal(display.percentage, 0, "percentage must use Math.floor, not Math.round");
-});
-
-test("getTokenBarDisplay unknown spec regression — hasKnownLimit is false", () => {
-  const spec: PendingModelSpec = {
-    status: "unknown",
-    contextWindow: null,
-    maxOutputTokens: null,
-    sourceUrl: "",
-    verifiedAt: null,
-    error: null,
-  };
-  assert.equal(getTokenBarDisplay(99_999, spec).hasKnownLimit, false);
-});
-
-// ─── getTokenBarDisplay — estimated specs ─────────────────────────────────────
-
-test("getTokenBarDisplay with isEstimated spec returns limitText with ~ prefix and compact format", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 400_000,
-    maxOutputTokens: 400_000,
-    sourceUrl: "known-registry",
-    verifiedAt: 0,
-    isEstimated: true,
-  };
-  const display = getTokenBarDisplay(0, spec);
-  assert.equal(display.limitText, "~400K");
-  assert.equal(display.isEstimatedLimit, true);
-  assert.equal(display.hasKnownLimit, true);
-});
-
-test("getTokenBarDisplay with isEstimated spec still has correct percentage", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 400_000,
-    maxOutputTokens: 400_000,
-    sourceUrl: "known-registry",
-    verifiedAt: 0,
-    isEstimated: true,
-  };
-  const display = getTokenBarDisplay(40_000, spec);
-  assert.equal(display.percentage, 10);
-  assert.equal(display.hasKnownLimit, true);
-});
-
-test("getTokenBarDisplay with isEstimated uses compact M suffix for 1M context", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 1_048_576,
-    maxOutputTokens: 1_048_576,
-    sourceUrl: "known-registry",
-    verifiedAt: 0,
-    isEstimated: true,
-  };
-  const display = getTokenBarDisplay(0, spec);
-  assert.equal(display.limitText, "~1.0M");
-  assert.equal(display.isEstimatedLimit, true);
-});
-
-test("getTokenBarDisplay with isEstimated: false uses comma format (regression guard)", () => {
-  const spec: VerifiedModelSpec = {
-    status: "verified",
-    contextWindow: 400_000,
-    maxOutputTokens: 400_000,
-    sourceUrl: "",
-    verifiedAt: 0,
-    isEstimated: false,
-  };
-  const display = getTokenBarDisplay(0, spec);
-  assert.equal(display.limitText, "400,000");
-  assert.equal(display.isEstimatedLimit, false);
-});