Hard line breaks are not working in .md files

[WouterHouwen]

According to the specs of CommonMark, you can set a hard line break by adding 2 or more spaces at the end of a line that is not the end of a block.

Here is an example markdown file: line_break.md
The PDF output is this: line_break.pdf

The expected output is that "This should be on the next line" is on the next line.

The command I used to create the PDF is: rinoh line_break.md

commonmark.js shows it correctly https://spec.commonmark.org/dingus/?text=The%20following%20text%20should%20be%20on%20a%20new%20line%20%20%0A%22This%20should%20be%20on%20the%20next%20line%22%0A

CommonMark hard line break spec: https://spec.commonmark.org/0.29/#hard-line-break

Installed versions:

• rinohtype version: 0.5.3 (in development) (Installed with pip install git+https://github.com/brechtm/rinohtype.git@master on June 10th 2021)
• Sphinx version: 4.0.2
• Python version: 3.8.5
• operating system: Ubuntu 20.04.2 LTS

[brechtm]

recommonmark, the library rinohtype relies on for CommonMark support, translates this line break to a raw HTML node. rinohtype ignores raw nodes meant for other docutils writers (as it should).

Output from cm2pseudoxml line_break.md:

<document source="line_break.md">
    <paragraph>
        The following text should be on a new line
        <raw format="html" xml:space="preserve">
            <br />
        "
        This should be on the next line
        "

I think recommonmark outputs a raw node because docutils doesn't offer a hard line break node, as far as I know. Instead, recommonmark should transform a paragraph containing a hard line break to a line block so that it can be rendered properly by non-HTML writers.

You could open a ticket with recommonmark, but I see that package is now deprecated, so we might need to move away from it. The README recommends using MyST instead. Another option is pycmark.

[chrisjsewell]

You could open a ticket with recommonmark, but I see that package is now deprecated, so we might need to move away from it. The README recommends using MyST instead.

Heya, author of MyST here, certainly interested in rinohtype supporting MyST, since we were looking at in for https://github.com/executablebooks/jupyter-book 😄

cc @choldgraf

[brechtm]

Heya, author of MyST here, certainly interested in rinohtype supporting MyST, since we were looking at in for https://github.com/executablebooks/jupyter-book 😄

That would be a great use case for rinohtype!

If MyST can output to a plain docutils/Sphinx document tree, adding a MyST frontend to rinohtype should be trivial (see the CommonMark frontend code for reference). Is this something you want to try and implement? I'll be on vacation until late September, but I can help out when I'm back.

[chrisjsewell]

If MyST can output to a plain docutils/Sphinx document tree

It can indeed 👍
Yeh same, I probably won’t have much/any time in September anyway, but certainly something I or one of the team would like to try thanks, when I can find the time 😬

[brechtm]

@chrisjsewell Reminded of this by #370, I had a go at replacing recommonmark with MyST and it turned out to be dead-simple! I did see an issue with relative links, however: #370 (comment)

Also, the line break issue described in this very issue persists with MyST. (using trailing spaces to indicate a line break -- or anything else, really -- is a horrible choice though /rant)