Permalinks for inline headings - #3168
Conversation
|
I'm working on the assumption this replaces #3150 - that is not 100% clear to me from reading comments. Report from Claude, just now. I'm not concerned about degrading legacy themes. |
|
The direction here is right — suppressing the permalink where a heading is genuinely inline, rather than by a list of block names, is the version that will keep working as themes change. Bringing Building the branch and comparing it against master on identical HTML, though, the rule that comes out of That rule hid the first-paragraph permalink across a dozen block types. The mixin only fires where a block actually has an inline heading, which is the point — but a few blocks were relying on the old rule and do not get the mixin. On the sample article with
Those blocks all carry their own permalink already, so a second one on their first paragraph is duplication, which is presumably what the old rule was for. Either extending the mixin to cover them, or keeping a narrowed version of the old rule for blocks that have a permalink of their own, would close it. One other thing worth knowing, since it bears on whether this closes the issue by itself rather than on anything in the diff. The exercisegroup fix rides on the component layer, and the legacy themes do not pull that in, so it does not reach them:
#3150 addresses the same symptom in the stylesheet instead, where the permalink is never generated and every theme is covered. The two are not in conflict — if that one lands, it simply makes the exercisegroup half of this redundant, and the consolidation here is still worth having on its own terms. Claude Opus 5, acting as a review assistant for Rob Beezer |
|
Yes, this is offered as a deeper fix for the problem #3150 addresses. The reason permalinks for Updated to address issue Claude found. That fix revealed another wrinkle - |
Cleanup for inline heading logic in CSS based on:
#3150
Ties link "hide permalink" for affected elements to the fact that a particular element has an inlined heading, not on its type (
.solution-like,.proof, etc...).