Remove obsolete CSS features and enable obsolescence lint for CSS

[Elchi3]

Summary

Right now, our irrelevance guideline is not applied to the css/ folder because the lint has been disabled (from the start when we first implemented this rule, there were just too many data points to clean up at the time). We're now down to 25 entries there and I think it is time to finally enable this lint for CSS, too.

If you enable the lint on main you get the following messages:

  ✖ css.properties.-moz-user-focus - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.-webkit-mask-repeat-x - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.-webkit-mask-repeat-y - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.background-image.image-rect - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.background-repeat-x - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.background-repeat-y - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.break-after.paged_context.recto - Error → feature was never implemented.
  ✖ css.properties.break-before.paged_context.recto - Error → feature was never implemented.
  ✖ css.properties.caption-side.left - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.caption-side.right - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.properties.display.contents.focusable_elements - Error → feature was never implemented.
  ✖ css.properties.font-variant.greek_accented_characters - Error → feature was never implemented.
  ✖ css.properties.font-variant.uppercase_eszett - Error → feature was never implemented.
  ✖ css.properties.text-autospace.punctuation - Error → feature was never implemented.
  ✖ css.properties.text-autospace.replace - Error → feature was never implemented.
  ✖ css.selectors.-moz-broken - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.selectors.closed - Info → feature was implemented and has since been removed from all browsers dating back two or more years ago.
  ✖ css.selectors.empty.matches_whitespace - Error → feature was never implemented.
  ✖ css.selectors.scroll-button.next - Error → feature was never implemented.
  ✖ css.selectors.scroll-button.prev - Error → feature was never implemented.
  ✖ css.types.attr.type_function.url - Error → feature was never implemented.
  ✖ css.types.basic-shape.path.shape-outside - Error → feature was never implemented.
  ✖ css.types.image.paint.additional_parameters - Error → feature was never implemented.
  ✖ css.types.param.url - Error → feature was never implemented.
  ✖ css.types.param.url_fragments - Error → feature was never implemented.

I removed all of these except for:

• css.properties.break-after.paged_context.recto: The linked impl_url bug is fixed and 87 seems to be the version number
• css.properties.break-before.paged_context.recto: same

Test results and supporting details

Improved linting now passes for the CSS folder.
Future lints will prevent authors from adding all-false CSS features.

Related issues

None, I think.

[github-actions]

Tip: Review these changes grouped by change (recommended for most PRs), or grouped by feature (for large PRs).

[caugner]

LGTM, but I would reframe the PR title to removing these obsolete CSS features.

Not mandatory, but with PR stacks you could have the bottom PR removing the obsolete CSS features, another one on top that updates the non-obsolete CSS features, and a third one that removes the obsolete lint. This is slightly more work when doing these changes all manually, but simplifies the reviews, and allows to merge them all together (in three separate commits) when all are approved.