Skip to content

Fix theme toggle inline - #57

Merged
Fivell merged 2 commits into
activeadmin-plugins:masterfrom
yeti-switch:fix-theme-toggle-inline
Oct 3, 2026
Merged

Fivell merged 2 commits into
activeadmin-plugins:masterfrom
yeti-switch:fix-theme-toggle-inline

Conversation

@dmitry-sinina

Copy link
Copy Markdown
Contributor

No description provided.

  The switch's li was display: flex, which is block-level. ActiveAdmin lays
  the utility nav out with inline items, so the switch broke the row and
  stacked the username, theme switch and logout. Use inline-flex, and make
  css_check fail if any utility-nav item becomes block-level.
@Fivell
Fivell merged commit 2555059 into activeadmin-plugins:master Oct 3, 2026
2 checks passed
@Fivell
Fivell deleted the fix-theme-toggle-inline branch October 3, 2026 14:36
Fivell added a commit to yeti-switch/active_admin_theme that referenced this pull request Oct 3, 2026
Two findings from the Copilot review of this branch, both real.

$skinStatusTagTextColor ended up declared twice when this branch rebased onto
activeadmin-plugins#57, which had added its own declaration: once beside the status tag fills and
once beside the table header colours. Sass keeps the first !default and drops
the rest, so the second was dead code that could drift from the live one in
silence. Both said #ffffff, so nothing rendered differently — it was waiting to.

The README check only walked the documented names, so a new variable could
ship with no row in the table and still pass, while the success line claimed
the table matched every declaration. It now compares both sets: a declaration
with no row fails, and so does a second declaration of the same name. The
count comes from the declarations rather than the rows.

Checked by adding $skinTotallyUndocumented and a second $skinStatusTagOkColor
and watching each fail with its own message.
Fivell added a commit to yeti-switch/active_admin_theme that referenced this pull request Oct 3, 2026
Two findings from the Copilot review of this branch, both real.

$skinStatusTagTextColor ended up declared twice when this branch rebased onto
activeadmin-plugins#57, which had added its own declaration: once beside the status tag fills and
once beside the table header colours. Sass keeps the first !default and drops
the rest, so the second was dead code that could drift from the live one in
silence. Both said #ffffff, so nothing rendered differently — it was waiting to.

The README check only walked the documented names, so a new variable could
ship with no row in the table and still pass, while the success line claimed
the table matched every declaration. It now compares both sets: a declaration
with no row fails, and so does a second declaration of the same name. The
count comes from the declarations rather than the rows.
Fivell added a commit to yeti-switch/active_admin_theme that referenced this pull request Oct 5, 2026
… marker

Five findings from the Copilot review of this branch, all real.

$skinStatusTagTextColor ended up declared twice when this branch rebased onto
activeadmin-plugins#57, which had added its own declaration. Sass keeps the first !default and
drops the rest, so the second was dead code waiting to drift. One now.

The README check only walked the documented names, so a variable could ship
with no row while the success line claimed the table matched every
declaration. It compares both sets now, and the count comes from the
declarations.

A blank cell was being recorded as documentation: the name went into the
documented set, which exempted it from the undocumented check, while the
mismatch check skipped it for having no value — so it passed on both sides.
Blank cells are no longer recorded.

A name listed twice in the table silently kept the last row, so the stylesheet
could agree with a row no reader reaches first. Duplicate rows now fail, the
same way duplicate declarations do.

The unsorted sort marker sat at 0.4 opacity, which against the header fill is
2.13:1 in light and 2.73:1 in dark. It is the only thing separating a sortable
heading from a plain one, so WCAG 1.4.11 asks 3:1 of it. At 0.6 it is 3.40 and
4.22.

The comment above the tag fills also read as though the new palette were the
one scoring 2.35 to 3.78; those were the fills it replaced.

Each hole was reproduced before and after: a blank dark default, a row listed
twice, a second declaration, an undocumented variable.
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.

2 participants