Skip to content

Row opacity compounding, and the test count this repo stated twice - #30

Merged
Varnasr merged 1 commit into
mainfrom
claude/outdated-repos-maintenance-nbxnae
Sep 23, 2026
Merged

Varnasr merged 1 commit into
mainfrom
claude/outdated-repos-maintenance-nbxnae

Conversation

@Varnasr

@Varnasr Varnasr commented Sep 23, 2026

Copy link
Copy Markdown
Owner

What does this PR do?

.row span in the shared stack.css took an opacity on top of a token that had already been measured, which put the language tags under AA on the landing page.

The suite has 101 tests. CLAUDE.md's Testing section said 25, and so did the header comment in requirements.txt, while the Layout section of the same file said 101. Two numbers for one suite, in one repository, and nothing compared either against pytest. The count is read off a run now and stated once.

Type of change

  • Bug fix
  • Accessibility improvement
  • Documentation update

Checklist

  • PYTHONPATH=$(pwd) pytest tests/ — 101 passed on Python 3.11 against the pinned set
  • Landing page measured at 390x844; no horizontal overflow

Commits

  • Fix the row opacity compounding, and the test count this repo states twice

🤖 Generated with Claude Code

https://claude.ai/code/session_01MrvR2NXsJFVRCeJZFCPuNL


Generated by Claude Code

…twice

Two things, both a number or a rule that nothing checked.

The CSS. `.row span { opacity: .7 }` matched more than the tag it was
written for: `.t` is itself a span wrapping the title and the tag, so
the selector applied to the tag twice (0.7 x 0.7 = 0.49) and to the
title once. On this landing page that put 12 tags at #84827f on
#f6f2ea, 3.43:1 at 10.56px against AA's 4.5:1, and dimmed 12 row titles
that `.row b { font-weight: 600 }` exists to make prominent. Scoped to
`.row .t > span`: tag 6.98:1, title 17.39:1, hover 8.68:1.

assets/css/stack.css is byte-identical across the four stack
repositories by design, so the change lands in all four. Counted before:
InsightStack 18 elements, EquityStack 12, SignalStack 7, FieldStack 0
(different markup on its landing page). After, on all four: 0 serious
or critical axe violations at 1280x900 and 390x844, no overflow.

The test count. CLAUDE.md's Testing section said 25 tests and the
header comment in requirements.txt said 25, while CLAUDE.md's own
Layout section fifteen lines earlier said 101. The suite is 101, run
today against the pinned set on Python 3.11. Two numbers for one suite
in one repository, disagreeing with each other and with pytest, and
nothing compared any of them.

Both corrected to 101, with a line recording that the disagreement was
there and unnoticed. Note the suite still emits pydantic @validator
deprecation warnings, which CLAUDE.md already flags: the pin holds
pydantic at 2.13.5 and @validator is removed in 3, so whoever raises it
moves those to @field_validator in the same change.

Also checked while here: pip-audit over the pinned set reports no known
vulnerabilities. It flags setuptools 79.0.1, which is the venv's own
build tooling and is required by nothing in requirements.txt except
under dev, testing and build extras.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrvR2NXsJFVRCeJZFCPuNL
@Varnasr
Varnasr merged commit 507bee3 into main Sep 23, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant