DRAFT: HTML support

[ctrlcctrlv]

I don't know what to do about the options thing given I don't think it's possible to give arguments to raw handlers, but wanted to send you the patch so the work isn't duplicated.

[Omikhleia]

Relates to #13

[Omikhleia]

In packages/markdown/init.lua you are loading it via the class, so there wouldn't be any reason to load it too via require(), is there? (and the commands would be registered via the class if you rename your registerHTMLcommands to registerCommands, simply.

[Omikhleia]

Why is this removed? We'd still want to support the standard Markdown linebreaks.

[ctrlcctrlv]

Mistake during testing thanks

[Omikhleia]

SILE does not have these, right? We'd need to map them

• to raise/lower (in the general case)
• or perhaps even, initially to the package-provided textsuperscript/textsubscript (if containing only text).

[ctrlcctrlv]

It has them.

[Omikhleia]

It doesn't seem so (at least, there's no "sup"/"sub" in the base and plain class), they only exist it seems in math mode.

[ctrlcctrlv]

Heh, with all due respect, you do so much work it seems you forget some of it!

markdown.sile/packages/markdown/commands.lua

Line 97 in cd90084

self.class:loadPackage("textsubsuper")

[Omikhleia]

Heh, with all due respect, you do so much work it seems you forget some of it!

Heh :) But in that case the commands are textsuperscript and textsubscript then (not sup and sub, hence my initial question).

In the current implementation, they'd work if the content is text-only. (If it's not, they ought to fail, although we may reconsider our options)

[ctrlcctrlv]

Huh, strange, my document still compiles though. But you're right, the text is not superscripted.

[Omikhleia]

I don't know what to do about the options thing given I don't think it's possible to give arguments to raw handlers

It actually is --

% Include example
\include[src=somefile.md, smart=false, bracketed_spans=false, subscript=false]
% Raw handler example
\begin[type=markdown, smart=false, bracketed_spans=false, subscript=false]{raw}
H~2~O and 2^10^ [Doe]{.smallcaps}
\end{raw}

=> I have just added the missing support for it (#19) 😸

[ctrlcctrlv]

I don't know what to do about the options thing given I don't think it's possible to give arguments to raw handlers

It actually is --

Via command line as well?

[Omikhleia]

Via command line as well?

Ah, no, I don't think it's actually possible by command line. Well that would be a question for SILE (passing main inputter's options via the CLI), but we'll be ready for that case too, then.

[ctrlcctrlv]

Thanks, I was only testing in command line thus my confusion. 🙇

[Omikhleia]

A div cannot be made an hbox (or it won't be properly line broken) -- same for a span actually.

But more generally, Pandoc has lots of extensions regarding HTML in Markdown, e.g. markdown_in_html_blocks which allows mixing HTML and Markdown, if I read that correctly.

I have one more concern which comes to mind: I'd ideally like both the pandocast and markdown packages to have the same minimal set of features. This imply, for any test we write, looking at what the Pandoc AST would contain, to at least ensure reasonable parity, whichever route is taken.

[Omikhleia]

I am closing the PR:

• The larger picture is documented in Closer look at inline HTML in Markdown #29 (also referring to this PR so it may be consulted in due time)
• The minimal support for <br> and <wbr> is done ine Support raw inline HTML br and wbr elements #30

But thanks again for the incentive and the ideas.