/**
 * @file repo-links.test.js
 * @description Guards the two staleness classes that shipped broken commands to
 * users for weeks: (1) GitHub-shaped references in a repo that is self-hosted on
 * Gitea — the `owner/repo` marketplace shorthand resolves to github.com ONLY
 * (verified against the CLI: it rejects `host/owner/repo` outright and ignores
 * GH_HOST / GITHUB_HOST / CLAUDE_CODE_MARKETPLACE_URL), plus raw.githubusercontent
 * / *.github.io / `/blob/<branch>` URLs that 404 against Gitea; and (2) a stale
 * committed `openapi.yaml`, which had silently drifted a whole minor version
 * behind package.json. Pure file reads plus one in-process spec render.
 * @author Nguyễn Ngọc Trí Vĩ <vinnt@smartgift.vn>
 */

const { describe, it } = require("node:test");
const assert = require("node:assert/strict");
const fs = require("node:fs");
const path = require("node:path");
const { execFileSync } = require("node:child_process");

const REPO_ROOT = path.join(__dirname, "..", "..");

// Only git-tracked text files: keeps node_modules, client/dist and every
// build artifact out without hand-maintaining an ignore list.
const TRACKED_EXT = /\.(md|json|js|cjs|mjs|ts|tsx|sh|ya?ml|html|css)$/;

// This file states every banned pattern verbatim — in the rules below and in
// the comments explaining them — so it must exclude itself or it fails on its
// own source. Derived from __filename rather than hardcoded so a rename can't
// silently re-arm the self-report.
const SELF = path.relative(REPO_ROOT, __filename);

function trackedFiles() {
  // No GIT_* scrub here, deliberately: under this repo's pre-commit hook
  // GIT_INDEX_FILE points at the index being committed, so `ls-files` lints
  // exactly what is about to land. Scrubbing it would lint the previous commit
  // instead and let a newly staged violation through.
  const out = execFileSync("git", ["ls-files", "-z"], {
    cwd: REPO_ROOT,
    encoding: "utf8",
    maxBuffer: 20_000_000,
  });
  return (
    out
      .split("\0")
      .filter((f) => f && TRACKED_EXT.test(f))
      .filter((f) => f !== SELF)
      // Dated design docs and plans record what was true when written; rewriting
      // history to satisfy a lint would be a lie about the decision record.
      .filter((f) => !f.startsWith("docs/superpowers/"))
      // Generated from server/openapi-extra/* (which IS linted): the drift test
      // below covers it, so linting both would only double-report one fix.
      .filter((f) => f !== "openapi.yaml" && !f.endsWith("package-lock.json"))
      // Third-party files copied unchanged (server/vendor/redoc): their links are upstream's.
      .filter((f) => !f.includes("/vendor/"))
  );
}

// Each rule: what to look for, and what to do instead. A `skip` pattern is
// tested against the MATCHED TEXT ONLY, never the surrounding line — a skip
// written against nearby prose silently never fires. Add one only when the
// exception is visible inside the match itself.
const RULES = [
  {
    name: "GitHub owner/repo marketplace shorthand",
    // `claude plugin marketplace add Smartgift-AI/AgentHub` — resolves to
    // github.com and fails, because this repo lives on Gitea.
    re: /plugin marketplace add\s+(?!https?:\/\/)(?!\.{0,2}\/)(?!["'$])[A-Za-z0-9][\w.-]*\/[A-Za-z0-9][\w.-]*/g,
    fix: "use the full https:// clone URL — the shorthand is GitHub-only",
    // Anthropic's own marketplace really does live on GitHub.
    skip: [/anthropics\/claude-plugins-official/],
  },
  {
    name: "GitHub URL for a Smartgift repo",
    re: /(?:https?:\/\/)?(?:www\.)?github\.com\/[Ss]martgift[\w-]*\/[\w.-]+/g,
    fix: "point at git.smartgift.io.vn",
    skip: [],
  },
  {
    name: "raw.githubusercontent.com",
    // Our repos are not on GitHub, so every raw URL 404s. Same-origin paths or
    // a Gitea /raw/branch/<ref>/ URL instead.
    re: /raw\.githubusercontent\.com\/[\w-]+\/[\w.-]+/g,
    fix: "use a same-origin path, or a Gitea /raw/branch/<ref>/ URL",
    // A third-party project that really is on GitHub: the public model price
    // list server/lib/pricing-sync.js reads.
    skip: [/raw\.githubusercontent\.com\/BerriAI\/litellm$/],
  },
  {
    name: "GitHub Pages host for a Smartgift site",
    re: /[\w-]*smartgift[\w-]*\.github\.io/gi,
    fix: "there is no GitHub Pages site — link the Gitea docs directory",
    skip: [],
  },
  {
    name: "GitHub /blob/<branch>/ path syntax on a Gitea URL",
    // Gitea serves files at /src/branch/<ref>/, never /blob/<ref>/.
    re: /git\.smartgift\.io\.vn\/[^\s"')\]]*\/blob\/[\w.-]+\//g,
    fix: "Gitea serves files at /src/branch/<ref>/",
    skip: [],
  },
  {
    name: "retired repository name",
    // Gitea 301-redirects the old name, so these never fail loudly — they just
    // rot until someone turns the redirect off.
    re: /Claude-Code-Monitor|Claude-Code-Agent-Monitor/g,
    fix: "the repository is Smartgift-AI/AgentHub",
    skip: [],
  },
];

describe("repo links and generated files", () => {
  const files = trackedFiles();

  it("finds tracked text files to lint", () => {
    assert.ok(files.length > 50, `expected a populated file list, got ${files.length}`);
  });

  for (const rule of RULES) {
    it(`no ${rule.name}`, () => {
      const hits = [];
      for (const rel of files) {
        const text = fs.readFileSync(path.join(REPO_ROOT, rel), "utf8");
        // A fresh regex per file: the /g rules carry lastIndex between calls.
        for (const m of text.matchAll(new RegExp(rule.re.source, rule.re.flags))) {
          if (rule.skip.some((s) => s.test(m[0]))) continue;
          const line = text.slice(0, m.index).split("\n").length;
          hits.push(`${rel}:${line}: ${m[0].trim()}`);
        }
      }
      assert.deepEqual(hits, [], `${rule.fix}\n  ${hits.join("\n  ")}`);
    });
  }

  it("committed openapi.yaml matches the generator output", () => {
    const { render, OUT } = require("../../scripts/generate-openapi-yaml");
    assert.equal(
      fs.readFileSync(OUT, "utf8"),
      render().text,
      "openapi.yaml is stale — run `npm run openapi:yaml` and commit the result"
    );
  });
});