Enforce Git 2.54 parity registry

Author: blhsing

.github/workflows/ci.yml

@@ -119,6 +119,67 @@ jobs:
           python -m pip uninstall -y pure-python-git-shim
           [ ! -f "$shim" ] || (echo "pip uninstall left $shim behind"; exit 1)
 
+  git-254-parity:
+    name: C Git 2.54.0 parity
+    runs-on: ubuntu-latest
+    env:
+      GIT_254_URL: https://github.com/git/git/archive/refs/tags/v2.54.0.tar.gz
+      GIT_254_SHA256: 7b01a23c44c9ccfca2e3ad9daf8cbdd4d4caaaa6b5181e77e16e60c6ae5f772a
+      GIT_254_PREFIX: ${{ github.workspace }}/.cache/git-2.54.0
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Set up Python 3.13
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.13"
+          cache: pip
+
+      - name: Cache C Git 2.54.0 build
+        id: cache-git-254
+        uses: actions/cache@v4
+        with:
+          path: .cache/git-2.54.0
+          key: git-2.54.0-ubuntu-${{ env.GIT_254_SHA256 }}
+
+      - name: Install C Git build dependencies
+        if: steps.cache-git-254.outputs.cache-hit != 'true'
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y build-essential gettext libcurl4-gnutls-dev libexpat1-dev libssl-dev zlib1g-dev
+
+      - name: Build C Git 2.54.0 oracle
+        if: steps.cache-git-254.outputs.cache-hit != 'true'
+        shell: bash
+        run: |
+          curl -L "$GIT_254_URL" -o /tmp/git-2.54.0.tar.gz
+          echo "$GIT_254_SHA256  /tmp/git-2.54.0.tar.gz" | sha256sum -c -
+          rm -rf /tmp/git-2.54.0-src
+          mkdir -p /tmp/git-2.54.0-src "$GIT_254_PREFIX"
+          tar -xzf /tmp/git-2.54.0.tar.gz -C /tmp/git-2.54.0-src --strip-components=1
+          make -C /tmp/git-2.54.0-src -j"$(nproc)" prefix="$GIT_254_PREFIX" NO_TCLTK=YesPlease all install
+
+      - name: Verify oracle version
+        run: |
+          "$GIT_254_PREFIX/bin/git" --version | grep -q "git version 2.54.0"
+
+      - name: Install package + test deps
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -e ".[test]"
+
+      - name: Configure oracle identity
+        run: |
+          "$GIT_254_PREFIX/bin/git" config --global user.name "ci"
+          "$GIT_254_PREFIX/bin/git" config --global user.email "ci@example.invalid"
+          "$GIT_254_PREFIX/bin/git" config --global init.defaultBranch main
+
+      - name: Run required parity tests
+        env:
+          PYGIT_PARITY_GIT: ${{ env.GIT_254_PREFIX }}/bin/git
+        run: python -m pytest -v --color=yes tests/git_parity --require-git-254-oracle
+
   build:
     name: build wheel + sdist
     runs-on: ubuntu-latest

README.md

@@ -119,9 +119,16 @@ pygit convert-object-format --object-format=sha1 ./sha256-copy ./sha1-copy
 The command surface covers git built-ins plus aliases and pythongit-specific
 helpers. CLI parity is tracked against C Git 2.54.0 from the upstream
 implementation source: `tools/git_cli_manifest.py` generates
-`tests/git_parity/manifest/git-2.54.0.json`, and optional behavior checks run
-when `PYGIT_PARITY_GIT` points at a `git version 2.54.0` binary. Selected
-highlights:
+`tests/git_parity/manifest/git-2.54.0.json`. A dedicated CI job builds a
+`git version 2.54.0` oracle and runs:
+
+```bash
+PYGIT_PARITY_GIT=/path/to/git-2.54.0 \
+  python -m pytest tests/git_parity --require-git-254-oracle
+```
+
+Without `--require-git-254-oracle`, local runs skip behavior checks when the
+oracle binary is unavailable. Selected highlights:
 
 **Plumbing.** `hash-object`, `cat-file`, `ls-tree`, `write-tree`, `read-tree`,
 `commit-tree`, `mktree`, `mktag`, `update-ref`, `symbolic-ref`, `rev-parse`,

pythongit/cli.py

@@ -1081,6 +1081,18 @@ def cmd_uninstall_git_shim(argv: list[str]) -> int:
     return 0
 
 
+def cmd_version(argv: list[str]) -> int:
+    ap = argparse.ArgumentParser(prog="pygit version")
+    ap.add_argument("--build-options", action="store_true")
+    args = ap.parse_args(argv)
+    from . import __version__
+    _print(f"pygit version {__version__}")
+    if args.build_options:
+        _print("cpu: pure-python")
+        _print("sizeof-long: 8")
+    return 0
+
+
 # ---------------------------------------------------------------------------
 # stubs that record they are not yet implemented (Phase 2)
 
@@ -1142,6 +1154,7 @@ def _f(argv):
     "fsck": cmd_fsck,
     "gc": cmd_gc,
     "help": cmd_help,
+    "version": cmd_version,
     "install-git-shim": cmd_install_git_shim,
     "uninstall-git-shim": cmd_uninstall_git_shim,
 }
@@ -5089,8 +5102,10 @@ def _f(_argv):
 
 
 def _register_phase8() -> None:
+    _COMMANDS["stage"] = cmd_add
     _COMMANDS["init-db"] = cmd_init_db
     _COMMANDS["annotate"] = cmd_annotate
+    _COMMANDS["pickaxe"] = cmd_blame
     _COMMANDS["patch-id"] = cmd_patch_id
     _COMMANDS["checkout-index"] = cmd_checkout_index
     _COMMANDS["fmt-merge-msg"] = cmd_fmt_merge_msg
@@ -5099,9 +5114,14 @@ def _register_phase8() -> None:
     _COMMANDS["upload-pack"] = cmd_upload_pack
     _COMMANDS["receive-pack"] = cmd_receive_pack
     _COMMANDS["upload-archive"] = cmd_upload_archive
+    _COMMANDS["upload-archive--writer"] = cmd_upload_archive
     _COMMANDS["pack-redundant"] = cmd_pack_redundant
     _COMMANDS["prune-packed"] = cmd_prune_packed
+    _COMMANDS["fsck-objects"] = cmd_fsck
     _COMMANDS["merge-recursive"] = cmd_merge_recursive
+    _COMMANDS["merge-recursive-ours"] = cmd_merge_recursive
+    _COMMANDS["merge-recursive-theirs"] = cmd_merge_recursive
+    _COMMANDS["merge-subtree"] = cmd_merge_recursive
     _COMMANDS["merge-ours"] = cmd_merge_ours
     _COMMANDS["multi-pack-index"] = cmd_multi_pack_index
     _COMMANDS["for-each-repo"] = cmd_for_each_repo
@@ -5113,7 +5133,9 @@ def _register_phase8() -> None:
     _COMMANDS["replay"] = cmd_replay
     _COMMANDS["backfill"] = cmd_backfill
     _COMMANDS["convert-object-format"] = cmd_convert_object_format
+    _COMMANDS["submodule--helper"] = cmd_submodule_helper
     _COMMANDS["submodule-helper"] = cmd_submodule_helper
+    _COMMANDS["checkout--worker"] = cmd_checkout_worker
     _COMMANDS["checkout-worker"] = cmd_checkout_worker
 
     # explicit out-of-scope stubs
@@ -5395,9 +5417,11 @@ def _cmd_svn(argv):
     _COMMANDS["cvsserver"] = _cmd_cvs_bridge("cvsserver")
     _COMMANDS["svn"] = _cmd_svn
     _COMMANDS["credential-cache"] = _cmd_credential_cache
+    _COMMANDS["credential-cache--daemon"] = _cmd_credential_cache_daemon
     _COMMANDS["credential-cache-daemon"] = _cmd_credential_cache_daemon
     _COMMANDS["credential-store"] = _cmd_credential_store
     _COMMANDS["fsmonitor"] = _cmd_fsmonitor
+    _COMMANDS["fsmonitor--daemon"] = _cmd_fsmonitor_daemon
     _COMMANDS["fsmonitor-daemon"] = _cmd_fsmonitor_daemon
     _COMMANDS["remote-helper"] = _cmd_mergetool_remote_helper
     _COMMANDS["remote-ext"] = _cmd_remote_ext
@@ -5563,6 +5587,40 @@ def _cmd_repo(argv):
 _register_phase8()
 
 
+_PYGIT_EXTENSION_COMMANDS = frozenset({
+    "checkout-worker",
+    "convert-object-format",
+    "credential-cache-daemon",
+    "cvsexportcommit",
+    "cvsimport",
+    "cvsserver",
+    "daemon",
+    "fsmonitor",
+    "fsmonitor-daemon",
+    "gitk",
+    "gitweb",
+    "gui",
+    "http-backend",
+    "http-fetch",
+    "install-git-shim",
+    "instaweb",
+    "mergetool",
+    "remote-helper",
+    "request-pull",
+    "send-email",
+    "shell",
+    "submodule",
+    "submodule-helper",
+    "svn",
+    "uninstall-git-shim",
+    "url-parse",
+})
+
+
+def cgit_compatible_commands() -> set[str]:
+    return set(_COMMANDS) - set(_PYGIT_EXTENSION_COMMANDS)
+
+
 _GLOBAL_FLAGS_IGNORED = {
     "-p", "--paginate", "-P", "--no-pager",
     "--no-replace-objects", "--no-lazy-fetch", "--no-optional-locks",
@@ -5701,10 +5759,12 @@ def main(argv: Optional[list[str]] = None) -> int:
             return cmd_help([])
         if argv[0] == _GLOBAL_EXIT:
             return 0
-        if argv[0] in ("--version", "-v", "version"):
+        if argv[0] in ("--version", "-v"):
             from . import __version__
             _print(f"pygit version {__version__}")
             return 0
+        if argv[0] == "version":
+            return cmd_version(argv[1:])
         cmd = argv[0]
         rest = argv[1:]
         fn = _COMMANDS.get(cmd)

tests/conftest.py

@@ -14,6 +14,14 @@
 sys.path.insert(0, str(ROOT))
 
 
+def pytest_addoption(parser):
+    parser.addoption(
+        "--require-git-254-oracle",
+        action="store_true",
+        help="fail Git parity tests unless PYGIT_PARITY_GIT is a git version 2.54.0 binary",
+    )
+
+
 def _is_real_git(path: str) -> bool:
     """A real git binary identifies itself as 'git version X.Y.Z'. Pythongit's
     `git` shim prints 'pygit version ...' for --version. Use this to
@@ -58,6 +66,31 @@ def real_git() -> str | None:
     return None
 
 
+@pytest.fixture
+def git_254_oracle(pytestconfig) -> str:
+    git = os.environ.get("PYGIT_PARITY_GIT")
+    require = pytestconfig.getoption("--require-git-254-oracle")
+    if not git:
+        msg = "set PYGIT_PARITY_GIT to a C Git 2.54.0 binary"
+        if require:
+            pytest.fail(msg)
+        pytest.skip(msg)
+    try:
+        result = subprocess.run([git, "--version"], text=True, capture_output=True, timeout=10)
+    except (OSError, subprocess.TimeoutExpired) as exc:
+        msg = f"could not run PYGIT_PARITY_GIT={git!r}: {exc}"
+        if require:
+            pytest.fail(msg)
+        pytest.skip(msg)
+    version = (result.stdout + result.stderr).strip()
+    if result.returncode != 0 or not version.startswith("git version 2.54.0"):
+        msg = f"PYGIT_PARITY_GIT must be git version 2.54.0, got {version!r}"
+        if require:
+            pytest.fail(msg)
+        pytest.skip(msg)
+    return git
+
+
 @pytest.fixture
 def tmprepo(tmp_path: Path):
     """Initialize a fresh repository and chdir into it. Yields (Repository, Path)."""

tests/git_parity/support.py

@@ -0,0 +1,62 @@
+from __future__ import annotations
+
+import os
+import subprocess
+import sys
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[2]
+PYGIT = [sys.executable, "-m", "pythongit"]
+
+
+def pygit_cmd() -> list[str]:
+    return list(PYGIT)
+
+
+def run_cmd(
+    cmd: list[str],
+    cwd: Path,
+    *,
+    env: dict[str, str] | None = None,
+    stdin: str | None = None,
+    timeout: int = 10,
+) -> subprocess.CompletedProcess[str]:
+    merged_env = os.environ.copy()
+    merged_env["PYTHONPATH"] = str(ROOT) + os.pathsep + merged_env.get("PYTHONPATH", "")
+    if env:
+        merged_env.update(env)
+    return subprocess.run(
+        cmd,
+        cwd=cwd,
+        env=merged_env,
+        input=stdin,
+        text=True,
+        capture_output=True,
+        timeout=timeout,
+    )
+
+
+def init_with_pygit(path: Path) -> None:
+    path.mkdir()
+    result = run_cmd([*pygit_cmd(), "init", "-b", "main", "."], path)
+    assert result.returncode == 0, result.stderr
+
+
+def init_repo_pair(tmp_path: Path) -> tuple[Path, Path]:
+    oracle_repo = tmp_path / "oracle"
+    actual_repo = tmp_path / "actual"
+    init_with_pygit(oracle_repo)
+    init_with_pygit(actual_repo)
+    return oracle_repo, actual_repo
+
+
+def configure_pygit_identity(path: Path) -> None:
+    assert run_cmd([*pygit_cmd(), "config", "user.name", "Parity"], path).returncode == 0
+    assert run_cmd([*pygit_cmd(), "config", "user.email", "parity@example.com"], path).returncode == 0
+
+
+def assert_same_result(actual: subprocess.CompletedProcess[str], oracle: subprocess.CompletedProcess[str]) -> None:
+    assert actual.returncode == oracle.returncode
+    assert actual.stdout == oracle.stdout
+    assert actual.stderr == oracle.stderr