fix(update): harden venv rebuild + verify core deps after install

Two complementary fixes for a silent partial-install failure that bit
``hermes update`` in the wild: a fresh checkout pulled 145 commits,
``rebuild_venv`` failed to recreate the venv on Windows because
``shutil.rmtree(ignore_errors=True)`` couldn't delete files held open by
the running ``hermes.exe`` shim. ``uv venv`` then refused with
"A directory already exists at: venv" and the update fell back to
installing on top of the stale venv. The resulting partial install
missed exactly one newly-added base dep — ``pathspec==1.1.1`` — which
``hermes desktop --build-only`` imports at the top of its content-hash
check. The desktop rebuild died with ModuleNotFoundError and the parent
update only logged "⚠ Desktop build failed (non-fatal)". Same root cause
made the "default: sync failed" line in the skill-sync stage, because
that sync subprocess hit the same missing import.

Fix 1: ``rebuild_venv`` retries with ``--clear``
------------------------------------------------
If ``uv venv`` fails with "already exists" in stderr (which is what uv
prints, and what uv's own hint tells you to fix with --clear), retry
once with ``--clear``. Only this specific failure pattern triggers the
retry — disk-full / interpreter-download failures still surface as
before so we don't mask real problems.

Fix 2: post-install dep verification
------------------------------------
Belt-and-suspenders so future uv resolver quirks (or any other cause of
partial installs) surface immediately instead of hours later in a
downstream subprocess. After ``_install_python_dependencies_with_optional_fallback``
runs, ``_verify_core_dependencies_installed``:

  1. Reads ``[project.dependencies]`` straight from pyproject.toml
     (so we don't trust the venv's stale metadata).
  2. Filters by environment markers via ``packaging.requirements.Requirement``
     so cross-platform exclusions (``ptyprocess ; sys_platform != 'win32'``)
     don't false-positive on Windows.
  3. Runs ``importlib.metadata.version()`` for each remaining dep inside
     the *target* venv interpreter (resolved from ``VIRTUAL_ENV``, not
     ``sys.executable``).
  4. If anything is missing, reinstalls the base group with
     ``--reinstall`` to force re-resolution. If a second probe still
     reports missing deps, force-installs each one with its pinned spec.
  5. Treats final failure as a warning rather than a hard error — a
     single broken-on-PyPI dep shouldn't block an otherwise-successful
     update — but the message points at ``hermes update --force`` and
     names the missing packages so the user knows what's wrong.

Tests
-----
- ``TestRebuildVenv::test_retries_with_clear_when_dir_already_exists`` —
  simulates the rmtree-couldn't-delete-it failure mode and asserts the
  ``--clear`` retry path is taken and succeeds.
- ``TestRebuildVenv::test_does_not_retry_when_first_failure_is_not_dir_exists``
  — guards against masking real failures (disk full, etc.).
- ``test_verify_core_dependencies.py`` — 7 tests covering the happy
  path, the regression (missing pathspec triggers --reinstall), the
  per-package fallback when --reinstall doesn't help, the platform-
  marker filter so Windows doesn't try to install ptyprocess, the
  missing-pyproject noop, and the VIRTUAL_ENV resolver.

Co-authored-by: Kyssta <218078013+kyssta-exe@users.noreply.github.com>
This commit is contained in:
teknium1
2026-06-04 06:05:41 -07:00
committed by Teknium
co-authored by Kyssta
parent 28ca4460a1
commit c136eb4de1
4 changed files with 547 additions and 6 deletions
+79
View File
@@ -145,6 +145,85 @@ class TestRebuildVenv:
result = rebuild_venv(uv_bin, venv_dir)
assert result is False
def test_retries_with_clear_when_dir_already_exists(self, tmp_path):
"""On Windows, rmtree can silently fail when an open handle holds a
file in the venv (running hermes.exe, gateway, AV scanner). uv then
refuses with ``Caused by: A directory already exists at: venv``.
Make sure we don't give up — retry with ``--clear`` to force uv past
the stale directory and rebuild successfully."""
venv_dir = tmp_path / "venv"
venv_dir.mkdir()
(venv_dir / "stale_open_handle").write_text("rmtree couldn't delete me")
uv_bin = str(tmp_path / "bin" / "uv")
call_log: list[list[str]] = []
def fake_run(cmd, **kwargs):
call_log.append(list(cmd))
m = MagicMock()
if cmd[1] == "venv" and "--clear" not in cmd:
# First attempt: uv refuses because dir still exists
m.returncode = 1
m.stderr = (
"error: Failed to create virtual environment\n"
" Caused by: A directory already exists at: venv\n"
"hint: Use the `--clear` flag or set `UV_VENV_CLEAR=1` to replace the existing directory\n"
)
m.stdout = ""
return m
if cmd[1] == "venv" and "--clear" in cmd:
# Retry: succeeds. Simulate uv writing the python shim.
m.returncode = 0
m.stderr = ""
m.stdout = ""
bin_dir = venv_dir / ("Scripts" if os.name == "nt" else "bin")
bin_dir.mkdir(parents=True, exist_ok=True)
python_name = "python.exe" if os.name == "nt" else "python"
(bin_dir / python_name).write_text("#!/bin/sh\necho Python 3.11.0")
return m
if "--version" in cmd:
m.returncode = 0
m.stdout = "Python 3.11.0"
m.stderr = ""
return m
m.returncode = 0
return m
with patch("hermes_cli.managed_uv.subprocess.run", side_effect=fake_run), \
patch("hermes_cli.managed_uv.shutil.rmtree"):
from hermes_cli.managed_uv import rebuild_venv
result = rebuild_venv(uv_bin, venv_dir)
assert result is True, "rebuild should succeed after --clear retry"
# We expect exactly two ``uv venv`` calls: one without --clear, one with.
venv_calls = [c for c in call_log if len(c) >= 2 and c[1] == "venv"]
assert len(venv_calls) == 2, f"expected 2 venv calls, got {venv_calls}"
assert "--clear" not in venv_calls[0], "first call should not pass --clear"
assert "--clear" in venv_calls[1], "retry must pass --clear"
def test_does_not_retry_when_first_failure_is_not_dir_exists(self, tmp_path):
"""If uv venv fails for some other reason (e.g. interpreter download
failed, disk full), we should NOT silently retry with --clear —
that would mask a real problem. Just surface the original failure."""
venv_dir = tmp_path / "venv"
uv_bin = str(tmp_path / "bin" / "uv")
call_log: list[list[str]] = []
def fake_run(cmd, **kwargs):
call_log.append(list(cmd))
m = MagicMock(returncode=1, stderr="error: No space left on device", stdout="")
return m
with patch("hermes_cli.managed_uv.subprocess.run", side_effect=fake_run), \
patch("hermes_cli.managed_uv.shutil.rmtree"):
from hermes_cli.managed_uv import rebuild_venv
result = rebuild_venv(uv_bin, venv_dir)
assert result is False
venv_calls = [c for c in call_log if len(c) >= 2 and c[1] == "venv"]
assert len(venv_calls) == 1, "should not retry on non-dir-exists failures"
assert "--clear" not in venv_calls[0]
# ---------------------------------------------------------------------------
# update_managed_uv