opencode-config/tests/test_setup_memory.py

"""Tests for .opencode/scripts/setup-memory.sh — deterministic memory init.

Uses a local mock remote (``git init --bare``) instead of GitHub so the
tests are hermetic and offline-safe. ``tmp_path`` provides isolation.

The script is invoked via ``subprocess`` with env vars pointing at the
mock remote + a tmp dir. Each test asserts one of the deterministic
flow steps.

Script contract (``setup-memory.sh``):
  1. mkdir -p MEMORY_DIR
  2. clone REMOTE | git pull --ff-only
  3. remote set-url if origin != REMOTE
  4. install post-commit hook (auto-push) if missing/wrong
  5. memory index if .rag missing (best-effort, CLI optional)
  5b. generate JS wrapper for opencode-memory plugin (idempotent + backup)
  6. echo status

Exit 1 when ``OPENCODE_MEMORY_REMOTE`` is unset (no default — the
``.env.example`` provides the value; absence is a config error).
"""

import os
import subprocess
import sys
from pathlib import Path

import pytest

REPO_ROOT = Path(__file__).resolve().parent.parent
SCRIPT = REPO_ROOT / ".opencode" / "scripts" / "setup-memory.sh"
REPO_PYTHON = str(REPO_ROOT / ".venv" / "bin" / "python")

EXPECTED_HOOK = "#!/bin/bash\ngit push origin master 2>/dev/null || true\n"


def _seed_remote(remote_dir: Path) -> None:
    """Seed a bare remote with one commit on ``master``."""
    remote_dir.mkdir(parents=True, exist_ok=True)
    subprocess.run(["git", "init", "--bare", "-q", str(remote_dir)], check=True)
    seed = remote_dir.parent / "seed"
    seed.mkdir()
    subprocess.run(["git", "init", "-q", str(seed)], check=True)
    (seed / "README.md").write_text("# memory\n")
    subprocess.run(["git", "-C", str(seed), "add", "README.md"], check=True)
    subprocess.run(
        [
            "git",
            "-C",
            str(seed),
            "-c",
            "user.email=t@t",
            "-c",
            "user.name=t",
            "commit",
            "-qm",
            "init",
        ],
        check=True,
    )
    subprocess.run(["git", "-C", str(seed), "branch", "-M", "master"], check=True)
    subprocess.run(
        ["git", "-C", str(seed), "remote", "add", "origin", str(remote_dir)],
        check=True,
    )
    subprocess.run(["git", "-C", str(seed), "push", "-q", "origin", "master"], check=True)


def _run_script(
    memory_dir: Path,
    remote: str | None,
    wrapper: Path | None = None,
) -> subprocess.CompletedProcess[str]:
    """Run setup-memory.sh with env pointing at tmp paths.

    When ``remote`` is None, OPENCODE_MEMORY_REMOTE is removed from env
    (simulating the "no remote env" error path). ``wrapper`` sets
    ``MEMORY_WRAPPER_PATH`` so tests don't write to the real plugin dir.
    """
    env = {
        **os.environ,
        "OPENCODE_MEMORY_DIR": str(memory_dir),
        "GIT_TERMINAL_PROMPT": "0",
    }
    if remote is not None:
        env["OPENCODE_MEMORY_REMOTE"] = remote
    else:
        env.pop("OPENCODE_MEMORY_REMOTE", None)
    if wrapper is not None:
        env["MEMORY_WRAPPER_PATH"] = str(wrapper)
        wrapper.parent.mkdir(parents=True, exist_ok=True)
    env["MEMORY_WRAPPER_PYTHON"] = REPO_PYTHON
    env["OPENCODE_WORKSPACE"] = str(REPO_ROOT)
    return subprocess.run(
        ["bash", str(SCRIPT)],
        capture_output=True,
        text=True,
        check=False,
        env=env,
    )


def test_fresh_init(tmp_path: Path) -> None:
    """Empty dir → run → clone, hook, remote, files present."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    r = _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js")
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    assert (mem / ".git").is_dir(), "repo not cloned"
    assert (mem / "README.md").exists(), "file not pulled"
    assert (mem / ".git" / "hooks" / "post-commit").exists(), "hook missing"
    got = subprocess.run(
        ["git", "-C", str(mem), "remote", "get-url", "origin"],
        capture_output=True,
        text=True,
        check=True,
    ).stdout.strip()
    assert got == str(remote), f"remote mismatch: {got}"


def test_existing_repo(tmp_path: Path) -> None:
    """.git exists → run → pull --ff-only, no destructive changes."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    # capture state before second run
    head_before = subprocess.run(
        ["git", "-C", str(mem), "rev-parse", "HEAD"],
        capture_output=True,
        text=True,
        check=True,
    ).stdout.strip()
    r = _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js")
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    head_after = subprocess.run(
        ["git", "-C", str(mem), "rev-parse", "HEAD"],
        capture_output=True,
        text=True,
        check=True,
    ).stdout.strip()
    assert head_before == head_after, "pull --ff-only changed HEAD unexpectedly"
    assert "pulling latest" in r.stdout


def test_wrong_remote(tmp_path: Path) -> None:
    """remote ≠ expected → run → set-url corrects the remote."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    # corrupt remote URL (tmp path — avoids S108 insecure-tmp warning)
    wrong = tmp_path / "wrong-remote"
    subprocess.run(
        ["git", "-C", str(mem), "remote", "set-url", "origin", str(wrong)],
        check=True,
    )
    r = _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js")
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    assert "fixing remote" in r.stdout
    got = subprocess.run(
        ["git", "-C", str(mem), "remote", "get-url", "origin"],
        capture_output=True,
        text=True,
        check=True,
    ).stdout.strip()
    assert got == str(remote), f"remote not fixed: {got}"


def test_missing_hook(tmp_path: Path) -> None:
    """hook missing → run → hook created with correct content."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    hook = mem / ".git" / "hooks" / "post-commit"
    hook.unlink()
    assert not hook.exists()
    r = _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js")
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    assert hook.exists(), "hook not recreated"
    assert hook.read_text() == EXPECTED_HOOK, f"hook content wrong: {hook.read_text()!r}"
    assert "installing post-commit hook" in r.stdout


def test_idempotent(tmp_path: Path) -> None:
    """run 3x → state identical after run 1 and run 3."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"

    def _snapshot() -> dict[str, str | None]:
        head = subprocess.run(
            ["git", "-C", str(mem), "rev-parse", "HEAD"],
            capture_output=True,
            text=True,
            check=True,
        ).stdout.strip()
        hook = mem / ".git" / "hooks" / "post-commit"
        return {
            "head": head,
            "hook": hook.read_text() if hook.exists() else None,
            "remote": subprocess.run(
                ["git", "-C", str(mem), "remote", "get-url", "origin"],
                capture_output=True,
                text=True,
                check=True,
            ).stdout.strip(),
        }

    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    snap1 = _snapshot()
    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    assert _run_script(mem, str(remote), wrapper=tmp_path / "wrapper" / "rag.js").returncode == 0
    snap3 = _snapshot()
    assert snap1 == snap3, f"non-idempotent:\n  run1={snap1}\n  run3={snap3}"


def test_no_remote_env(tmp_path: Path) -> None:
    """OPENCODE_MEMORY_REMOTE unset → exit 1, error message to stderr/stdout."""
    mem = tmp_path / "mem"
    r = _run_script(mem, remote=None)
    assert r.returncode == 1, f"expected exit 1, got {r.returncode}"
    combined = r.stderr + r.stdout
    assert "OPENCODE_MEMORY_REMOTE" in combined, f"missing error msg: {combined!r}"
    assert "ERROR" in combined or "not set" in combined


# ── git insteadOf helper (GITHUB_TOKEN) ──────────────────────────────────────


def test_script_contains_git_insteadof() -> None:
    """setup-memory.sh contains `git config --global url.insteadOf` before clone."""
    src = SCRIPT.read_text()
    assert "git config --global url." in src, "missing git config --global url. call"
    assert ".insteadOf https://github.com/" in src, "missing insteadOf directive"
    # The insteadOf block must appear BEFORE the git clone line.
    insteadof_pos = src.index(".insteadOf")
    clone_pos = src.index("git clone --origin")
    assert insteadof_pos < clone_pos, "insteadOf helper must precede git clone"


def test_script_contains_github_token_check() -> None:
    """setup-memory.sh checks GITHUB_TOKEN (guarded, not hard-fail if empty)."""
    src = SCRIPT.read_text()
    assert "GITHUB_TOKEN" in src, "missing GITHUB_TOKEN reference"
    # Must be a conditional check (if [ -n ... ]), not an unconditional hard fail.
    assert '-n "${GITHUB_TOKEN:-}"' in src or '-n "$GITHUB_TOKEN"' in src, (
        "GITHUB_TOKEN must be a guarded check (if [ -n ... ]), not a hard requirement"
    )


def test_insteadof_idempotent(tmp_path: Path) -> None:
    """Running the script N times does not duplicate the insteadOf entry.

    `git config --global url.<prefix>.insteadOf` overwrites the value on each
    call (idempotent by git semantics). We verify by invoking setup-memory.sh
    twice with a fake GITHUB_TOKEN against a mock remote and checking that
    `git config --global --get-all` returns exactly one match after each run.
    """
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"

    # Use an isolated HOME so --global writes to a tmp gitconfig (no pollution).
    home = tmp_path / "home"
    home.mkdir()

    env = {
        **os.environ,
        "OPENCODE_MEMORY_DIR": str(mem),
        "OPENCODE_MEMORY_REMOTE": str(remote),
        "GITHUB_TOKEN": "fake-token-for-test",
        "HOME": str(home),
        "GIT_TERMINAL_PROMPT": "0",
        "MEMORY_WRAPPER_PATH": str(tmp_path / "wrapper" / "rag.js"),
        "MEMORY_WRAPPER_PYTHON": REPO_PYTHON,
        "OPENCODE_WORKSPACE": str(REPO_ROOT),
    }
    (tmp_path / "wrapper").mkdir(exist_ok=True)

    # First run: clone path triggers the insteadOf helper.
    r1 = subprocess.run(["bash", str(SCRIPT)], capture_output=True, text=True, check=False, env=env)
    assert r1.returncode == 0, f"run1 failed: stdout={r1.stdout}\nstderr={r1.stderr}"

    # Read the global gitconfig written by the script.
    gitconfig = home / ".gitconfig"
    assert gitconfig.exists(), f"gitconfig not created at {gitconfig}"
    content1 = gitconfig.read_text()
    assert "x-access-token:fake-token-for-test@github.com" in content1
    assert "insteadOf" in content1

    # Count insteadOf occurrences — should be exactly 1 (no duplication).
    insteadof_count_1 = content1.count("insteadOf")
    assert insteadof_count_1 == 1, f"expected 1 insteadOf after run1, got {insteadof_count_1}"

    # Second run: existing repo → pull path (clone branch skipped). The helper
    # is inside the clone branch, so it won't re-fire — but we assert the
    # config is NOT duplicated regardless (single insteadOf entry preserved).
    r2 = subprocess.run(["bash", str(SCRIPT)], capture_output=True, text=True, check=False, env=env)
    assert r2.returncode == 0, f"run2 failed: stdout={r2.stdout}\nstderr={r2.stderr}"
    content2 = gitconfig.read_text()
    insteadof_count_2 = content2.count("insteadOf")
    assert insteadof_count_2 == 1, (
        f"expected 1 insteadOf after run2, got {insteadof_count_2} (duplicate?)"
    )


def test_no_github_token_skips_insteadof(tmp_path: Path) -> None:
    """When GITHUB_TOKEN is unset, the insteadOf helper is skipped (no error)."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"

    home = tmp_path / "home"
    home.mkdir()

    env = {
        **os.environ,
        "OPENCODE_MEMORY_DIR": str(mem),
        "OPENCODE_MEMORY_REMOTE": str(remote),
        "HOME": str(home),
        "GIT_TERMINAL_PROMPT": "0",
        "MEMORY_WRAPPER_PATH": str(tmp_path / "wrapper" / "rag.js"),
        "MEMORY_WRAPPER_PYTHON": REPO_PYTHON,
        "OPENCODE_WORKSPACE": str(REPO_ROOT),
    }
    env.pop("GITHUB_TOKEN", None)
    (tmp_path / "wrapper").mkdir(exist_ok=True)

    r = subprocess.run(["bash", str(SCRIPT)], capture_output=True, text=True, check=False, env=env)
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    gitconfig = home / ".gitconfig"
    if gitconfig.exists():
        content = gitconfig.read_text()
        assert "insteadOf" not in content, "insteadOf set without GITHUB_TOKEN"


# ── JS wrapper for opencode-memory plugin (step 5b) ──────────────────────────


def test_wrapper_generated(tmp_path: Path) -> None:
    """Run → wrapper file exists at MEMORY_WRAPPER_PATH with delegate content."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    wrapper = tmp_path / "wrapper" / "rag.js"
    r = _run_script(mem, str(remote), wrapper=wrapper)
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    assert wrapper.exists(), f"wrapper not generated at {wrapper}"
    assert wrapper.stat().st_mode & 0o100, "wrapper not executable"
    assert "generated wrapper" in r.stdout, f"missing generate msg: {r.stdout!r}"


def test_wrapper_idempotent(tmp_path: Path) -> None:
    """Run 2x → wrapper content identical, 2nd run reports 'wrapper correct'."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    wrapper = tmp_path / "wrapper" / "rag.js"
    r1 = _run_script(mem, str(remote), wrapper=wrapper)
    assert r1.returncode == 0, f"run1: stdout={r1.stdout}\nstderr={r1.stderr}"
    content1 = wrapper.read_text()
    r2 = _run_script(mem, str(remote), wrapper=wrapper)
    assert r2.returncode == 0, f"run2: stdout={r2.stdout}\nstderr={r2.stderr}"
    content2 = wrapper.read_text()
    assert content1 == content2, "wrapper content changed between runs"
    assert "wrapper correct" in r2.stdout, f"2nd run should report correct: {r2.stdout!r}"


def test_wrapper_content(tmp_path: Path) -> None:
    """Wrapper contains python memory CLI delegation + spawnSync."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    wrapper = tmp_path / "wrapper" / "rag.js"
    r = _run_script(mem, str(remote), wrapper=wrapper)
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    body = wrapper.read_text()
    assert "python" in body, f"wrapper missing python ref: {body!r}"
    assert "-m" in body and "src.memory" in body, "wrapper missing src.memory module"
    assert "spawnSync" in body, "wrapper missing spawnSync"
    assert "process.argv.slice" in body, "wrapper missing argv passthrough"


def test_wrapper_backup_original(tmp_path: Path) -> None:
    """Pre-existing wrapper with different content → .orig backup saved once."""
    remote = tmp_path / "remote.git"
    _seed_remote(remote)
    mem = tmp_path / "mem"
    wrapper = tmp_path / "wrapper" / "rag.js"
    wrapper.parent.mkdir(parents=True, exist_ok=True)
    original = "# old rag-cli shim\nconsole.log('old');\n"
    wrapper.write_text(original)
    r = _run_script(mem, str(remote), wrapper=wrapper)
    assert r.returncode == 0, f"stdout={r.stdout}\nstderr={r.stderr}"
    backup = wrapper.with_suffix(".js.orig")
    assert backup.exists(), ".orig backup not created"
    assert backup.read_text() == original, ".orig does not preserve original content"
    assert wrapper.read_text() != original, "wrapper not replaced with delegate"
    # Second run with different wrong content: .orig must NOT be overwritten.
    wrapper.write_text("# another wrong\n")
    r2 = _run_script(mem, str(remote), wrapper=wrapper)
    assert r2.returncode == 0, f"run2: stdout={r2.stdout}\nstderr={r2.stderr}"
    assert backup.read_text() == original, ".orig overwritten on 2nd backup"


if __name__ == "__main__":
    sys.exit(pytest.main([__file__, "-v"]))