fix: reintegrate PasswordStrength (#4111)

[singhh-piyush]

• This fix issue: Password strength code unused #4111

PR Description:

Removes PasswordStrength enum and associated logic from lib/models/users.py. This is dead code following the refactor in 88b91ae.

Changes:

• lib/models/users.py: Deleted PasswordStrength class and methods (value, color, strength, _check_password_strength).
• Imports: Removed Enum, auto, and tr from users.py.
• lib/models/__init__.py: Removed PasswordStrength from imports and __all__.

Tests and Checks

• I have tested the code!

Verification:

• Tests: pytest tests/ passed (13/13).
• Linting: Passed ruff, mypy, flake8, and pylint.
• Audit: grep -r "PasswordStrength" archinstall/ returned 0 results.
• Style: Tab indentation preserved.

[singhh-piyush]

• Fixes issue: Password strength code unused #4111

PR Description

Restores the PasswordStrength validation logic that was removed during the commit 88b91ae. Based on feedback, instead of removing the dead code I integrated it into the password prompt flow.

Changes:

• archinstall/lib/menu/util.py: Imported PasswordStrength and warn (leveraging the existing tr import). Integrated the strength() check immediately after the initial password input. It now emits a localized CLI warning for VERY_WEAK, WEAK, and MODERATE passwords before proceeding to the confirmation prompt.
• tests/test_password_strength.py: Added a new test suite with 7 cases covering all enum tiers and edge cases (empty strings, digit-only, etc.).

Tests and Checks

• I have tested the code locally.

Verification:

• Tests: pytest tests/ passed (20/20, 0 regressions).
• Linting: Passed ruff, mypy, flake8, and pylint via local pre-commit hooks.
• Style: Tab indentation enforced and verified.

[singhh-piyush]

Logic and tests for PasswordStrength integrated. I've reverted all locale changes to keep this PR focused on the requested fix .

Technical Note: The translation-check CI is currently failing on master independent of this PR. A duplicate msgid in locales/hi/LC_MESSAGES/base.po (line ~1491) causes locales_generator.sh to abort mid-run, leaving the committed locale files out of sync with the generator's output. I've verified this reproduces on a clean checkout of master with no other changes.

All logic has been verified locally with pytest - 20/20 passed, including 7 new cases covering every PasswordStrength tier and edge cases.

[singhh-piyush]

Thanks for the feedback! I integrated PasswordStrength into the password prompt as a non-blocking hint. The hint renders beneath the input field and updates on each keystroke with color based on strength tier.

• VERY_WEAK / WEAK - red

• MODERATE - yellow

• STRONG - green

Files changed

• util.py - password_hint closure, wired to first Input() call only
• components.py - #input-info label, on_input_changed handler, CSS classes
• helpers.py - type hint updated to tuple[str, str] | None

Tests

Tested locally in a QEMU VM running the Arch ISO. Screenshots attached.

Password	Result
abc	Too weak - too short (red)
abcdefg	Weak - add uppercase, lowercase, numbers and symbols (red)
Abcdefgh12345	Moderate - increase length (yellow)
Abcdefgh12345!@#	Strong (green)

pytest tests/ - 20/20 passed. mypy, ruff, flake8, pylint - clean.

[CelestifyX]

My personal opinion is that adding this functionality if it forces you to make a more complex password makes no sense at all, but if it’s just a warning and then it will skip this password, then that’s a different matter.

[singhh-piyush]

My personal opinion is that adding this functionality if it forces you to make a more complex password makes no sense at all, but if it’s just a warning and then it will skip this password, then that’s a different matter.

Hey, it's a non-blocking warning only. You can still proceed with any password regardless of strength, more of a hint/warning for the user.

i added this to address issue #4111 where PasswordStrength became dead code after get_password() was moved in commit 88b91ae. Rather than deleting the enum, this restores its original intent as a user-facing hint.

[singhh-piyush]

hey @svartkanin , you were right, warn() was bypassing the TUI entirely and only hitting the logs. Thanks for the feedback, I really appreciate it

I made changes as suggested. Extended the Input widget with an info_callback. Feedback now renders below the field as the user enters a password, with no interference with the validate attribute.

Could you take a look at the PR when you get a chance?

Solves issue #4111

[svartkanin]

I'll try getting around to test this this week

[singhh-piyush]

Thank you for the feedback. Sorry for the messy commit history. I was working on my local archinstall at the same time, testing some BTRFS formatting changes and some of those unrelated changes got added into this branch by accident. The PR is now clean.

All three points from the review addressed:

• Simplified to Password strength: Weak / Moderate / Strong as suggested.

• renamed to generic .input-hint-msg-error, .input-hint-msg-warning, .input-hint-msg-info.

• replaced tuple[str, str] with InputInfo dataclass and InputInfoType enum as suggested.

---

Regarding VERY_WEAK - it is already defined in the existing PasswordStrength enum in users.py and is returned by the strength() classifier. I used it in password_hint to handle all four possible return values from that function. On the UI side, VERY_WEAK and WEAK are collapsed into the same display string 'Password strength: Weak', so the user only sees three levels

Tested locally in a QEMU VM.

[svartkanin]

Could you rebase master, there was a bigger change merged which unfortunately conflicted with your branch