ccorn

It should be noted that this change breaks previous CSS-based workarounds. If users never put empty lines before definitions and put the following kludge in their (S)CSS: ~~~ css /*...

I find that this behavior is [actually specified](https://kramdown.gettalong.org/syntax.html#definition-lists) ([source](https://github.com/gettalong/kramdown/blob/master/doc/syntax.page#L729)). However, my impression is that this behavior might be the reason for style rules like the [`margin-top` of blocks and countermeasures...

I have introduced a boolean option `:auto_dd_para` for the proposed parser change so that you can experiment a bit. Since it is off by default and not activated for the...

By the way, I get 2838 test runs without MathjaxNode tests whereas the CI only does 2818 test runs with MathjaxNode tests included. The numbers are for Ruby 2.4. No...

Merged changes from master, then rebased and squashed.

Changed `:auto_dd_para` criterion to a semantic version (several block children) instead of the lexical one (contains blank line followed by something else). This also handles examples like ~~~ kramdown item...

Added more tests for the :auto_dd_para option. Take a look at [`test/testcases/block/13_definition_list/auto_dd_para.text`](https://github.com/gettalong/kramdown/compare/master...ccorn:dd-para#diff-0d8be09aab060bcf4d55e72e8f066251) to get a glimpse of the current semantics: * Plain `para` is wrapped in `` regardless of the...

Yes, it is a parser option affecting the AST. It deliberately changes the grammar in order to get rid of (presumably unintented) mixed block/inline sequences. It does not affect how...

I'll have to update the wording in the specification to more closely match the implementation. The current (semantic) implementation is said to take effect when there are several blocks. In...

Just noticed the following: ~~~ bash $ kramdown