Block markup breaks <verbatim>

[link2xt]

Here is a testcase:

#title Test case

foo <verbatim>*not bold*
* Not a heading
</verbatim>baz

Emacs Muse and pandoc do not see a heading here.

I think it is a bug because it makes it very hard to generate Muse markup automatically: writer should take care not to place anything that may be recognized as a block start at the beginning of the line when wrapping the lines to a certain width.

I have not looked into the code, but looks like blocks are recognized at some kind of tokenization phase, before inline markup is parsed. If so, it may be hard to fix without reorganizing the parsing process to recognize <verbatim> before looking for blocks.

Feel free to keep fixing this bug low priority if it requires major changes to the parser architecture.

[melmothx]

@link2xt I'm a bit unsure about this. I've always considered <verbatim> as inline. And here the things are mixed up. What Emacs Muse sees if you remove the verbatim?

[link2xt]

Emacs Muse converts the test case above to:

<p>foo *not bold*
* Not a heading
baz</p>

If I remove verbatim, I get:

foo *not bold*
* Not a heading
baz

Then, Emacs Muse converts it to:

<p>foo <em>not bold</em>
<h2>Not a heading</h2></p>

<p>baz</p>

[melmothx]

Alexander <notifications@github.com> writes:

Then, Emacs Muse converts it to: <p>foo <em>not bold</em> <h2>Not a heading</h2></p> <p>baz</p>

Which is clearly broken as well. Ok, mumbling on this

…

-- Marco

[link2xt]

Seems like it does not expect headings to immediately follow paragraphs (see, </p> is misplaced).

But if I add a blank line, like this:

foo<verbatim>*not bold*

* Heading
</verbatim>
baz

i still get:

<p>foo*not bold*

* Heading

baz</p>

Emacs Muse has lots of bugs. It does not have anything resembling a parser, it is basically a number of regexp-triggered rewrites and even some search and replace operations on the whole text.

[link2xt]

I've always considered <verbatim> as inline.

I also treat <verbatim> as inline. The last example seems broken to me too, because there is a blank line inside inline markup. But in example without blankline it can be treated as inline <verbatim> with an escaped asterisk inside, that just happened to wrap on the next line.

Paragraphs are really special case, compared to other blocks. Unlike other blocks, like lists and tables, which can be stopped with any other blocks, paragraphs can't be stopped by other paragraphs. And I think they should not be stopped by blocks if they are in the middle of inline markup (inside link description, inside <verbatim> etc.), unless maybe blanklines, which do not make sense inside paragraphs, but can prevent the rest of the document from becoming a single paragraph due to unclosed tag.

[link2xt]

And I think they should not be stopped by blocks if they are in the middle of inline markup (inside link description, inside etc.), unless maybe blanklines

Also, blankline is not a block in some sense, and in pandoc I don't treat blanklines as special blocks that stop paragraphs. Instead, paragraph parser just doesn't accept blanklines.