What happened: HTML sections like abstract, conformance etc. are no longer recognized. The spec is using format: "markdown", where the main file is HTML, but it includes some markdown files (seen here).
But what I was expecting?: The documentation says, "_When format is set to "markdown", you can use a mix of HTML and markdown_" (see here). This worked before, but not anymore.
This worked before, but not anymore.
~Can you tell when it worked? It might be a regression.~
nvm. Regression from #2715
This seems to be a problem with indentation logic. The quick "fix" here is to unindent everything to zero-indentation.
cc: @saschanaz
The real problem is that data-include-format is missing. The indentation logic is fired only for markdown formats.
Edit: OMG we have no documentation for this.
Seems we need to modify the check as it really should fire for all data-include whose ancestor is markdown, not just for markdown-formatted inclusion.
So, I should remove format: "markdown" from configuration, since my main file is HTML, but include data-include-format="markdown" on every data-include, right?
Yup, that's my current understanding.
I tested it locally, but can't confirm. Removing format: "markdown" and adding data-include-format="markdown" on every include (except for the JSON examples) just leads to everything being interpreted as markdown. So it's even worse than before. 馃槙
Test branch: https://github.com/dini-ag-kim/oer-service-card/tree/test-data-include-format
Ah, I guess we also need data-format=markdown for each section.
@sidvishnoi could you explore a fix mentioned here https://github.com/w3c/respec/issues/2850#issuecomment-617152537 ?
@saschanaz I tried various combinations and fiddled around with indentation and markdown logic, but it didn't work for nested sections. Should we try to unindent sections in a depth-first order?