Virtual environment: handle pip updates (#129)

Instead of putting our payload into pip and making pip3 and pip3.12
symlinks to pip, put our payload into pip_patched and symlink all pip
variants to that. Then we can check for a problematic update by testing
whether readlink() of the current executing file still points to
pip_patched. If it doesn't, put it back.

Resolves #128

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

Author: hoodmane

pyodide_build/out_of_tree/venv.py

@@ -174,6 +174,45 @@ def platform_tags():
         sysconfig._init_config_vars()
         del os.environ["_PYTHON_SYSCONFIGDATA_NAME"]
         """
+        # Handle pip updates.
+        #
+        # The pip executable should be a symlink to pip_patched. If it is not a
+        # link, or it is a symlink to something else, pip has been updated. We
+        # have to restore the correct value of pip. Iterate through all of the
+        # pip variants in the folder and remove them and replace with a symlink
+        # to pip_patched.
+        """
+        from pathlib import Path
+
+        file_path = Path(__file__)
+
+
+        def pip_is_okay():
+            try:
+                return file_path.readlink() == file_path.with_name("pip_patched")
+            except OSError as e:
+                if e.strerror != "Invalid argument":
+                    raise
+            return False
+
+
+        def maybe_repair_after_pip_update():
+            if pip_is_okay():
+                return
+
+            venv_bin = file_path.parent
+            pip_patched = venv_bin / "pip_patched"
+            for pip in venv_bin.glob("pip*"):
+                if pip == pip_patched:
+                    continue
+                pip.unlink(missing_ok=True)
+                pip.symlink_to(venv_bin / "pip_patched")
+
+
+        import atexit
+
+        atexit.register(maybe_repair_after_pip_update)
+        """
     )
 
 
@@ -183,26 +222,17 @@ def create_pip_script(venv_bin):
     # Python in the shebang. Use whichever Python was used to invoke
     # pyodide venv.
     host_python_path = venv_bin / f"python{get_pyversion()}-host"
-    pip_path = venv_bin / "pip"
-    pyversion = get_pyversion()
-    other_pips = [
-        venv_bin / "pip3",
-        venv_bin / f"pip{pyversion}",
-        venv_bin / f"pip-{pyversion}",
-    ]
+    pip_path = venv_bin / "pip_patched"
 
     # To support the "--clear" and "--no-clear" args, we need to remove
     # the existing symlinks before creating new ones.
-    if host_python_path.exists():
-        host_python_path.unlink()
-    if (venv_bin / "python-host").exists():
-        (venv_bin / "python-host").unlink()
-    if pip_path.exists():
-        pip_path.unlink()
-    for pip in other_pips:
-        if pip.exists():
-            pip.unlink()
-            pip.symlink_to(pip_path)
+    host_python_path.unlink(missing_ok=True)
+    (venv_bin / "python-host").unlink(missing_ok=True)
+    for pip in venv_bin.glob("pip*"):
+        if pip == pip_path:
+            continue
+        pip.unlink(missing_ok=True)
+        pip.symlink_to(pip_path)
 
     host_python_path.symlink_to(sys.executable)
     # in case someone needs a Python-version-agnostic way to refer to python-host

pyodide_build/tests/test_venv.py

@@ -221,3 +221,38 @@ def test_pip_install(base_test_dir, packages):
             venv_path.glob(f"**/{package.replace('-', '_')}-*.dist-info")
         )
         assert len(dist_info_dirs) > 0, f"{package} not found in the venv"
+
+
+@pytest.mark.integration
+def test_pip_downgrade(base_test_dir):
+    """Test that our monkeypatched pip can upgrade/downgrade itself"""
+    venv_path = base_test_dir / "test_venv"
+
+    venv.create_pyodide_venv(venv_path, [])
+    venv_pip_path = venv_path / "bin" / "pip"
+    assert venv_pip_path.exists(), "pip wasn't found in the virtual environment"
+
+    result = subprocess.run(
+        [
+            str(venv_pip_path),
+            "install",
+            "--upgrade",
+            "pip==24.0",
+            "-v",
+            "--disable-pip-version-check",
+        ],
+        capture_output=True,
+        text=True,
+        check=False,
+    )
+    assert result.returncode == 0, f"Failed to downgrade pip: {result.stderr}"
+
+    result = subprocess.run(
+        [str(venv_pip_path), "--version"],
+        capture_output=True,
+        text=True,
+        check=False,
+    )
+    assert result.returncode == 0, result.stderr
+    assert result.stdout.startswith("pip 24.0")
+    assert venv_pip_path.readlink() == venv_pip_path.with_name("pip_patched")