Obtaining Syntax Mark Information from MDAST

[eliyap]

Hi All!

Is it possible to get syntax marks positions / ranges from MDAST?

For **strong** or _emph_, it's straightforwards.

Fenced code is trickier. They can open with 3+ backticks or tildes, and the ending fence may be longer than the opening fence.
AFAIK, MDAST exposes Code's overall range, lang and meta attributes, but not opening / closing range.

It's possible to search for value within the Literal range, but official support help in corner cases:

• `````` (an empty fence with 3 open + 3 close backticks)
• ```````` (3 open + 5 close? 4 open + 4 close?)

Lastly, I'd love this for Blockquotes, which are most difficult. > are interleaved within contents, may / may not be present (laziness clause), may / may not have a space after them, and may / may not be nested.

Appreciate any thoughts!

[wooorm]

What you’re looking for is called a CST. See also: #59.

micromark allows creating one, and was made for that, but one hasn’t been created yet.

Why do you want this?

[eliyap]

Sorry for the late response

Thank you for the clarification, a CST is exactly what I’m looking for!
My motivation is to improve “smart syntax”, see https://pencil.md (previously discussed here)

Take the following

I met a traveller from an antique land, who said:
> I met a traveller from an antique land, who said:
> > I met a traveller from an antique land, who said: ...
> > et cetera
> > et cetera

Say the user (in my app) strikes out / erases the entire last line – including the >'s!
In this case, I consider them to be targeting the text, not the syntax marks. (though we might clean up the syntax marks if the block is empty)

So we would wish to obtain:

I met a traveller from an antique land, who said:
> I met a traveller from an antique land, who said:
> > I met a traveller from an antique land, who said: ...
> > et cetera
> > ~~et cetera~~

instead of

I met a traveller from an antique land, who said:
> I met a traveller from an antique land, who said:
> > I met a traveller from an antique land, who said: ...
> > et cetera
~~> > et cetera~~

AFAIK, the MDAST can't tell me which > belongs to which blockquote, and due to the laziness clause, an educated guess is tricky.

I understand that CSTs aren't in high demand and might take a lot of work, so it it's not on the table that's ok too!

[wooorm]

laziness: yeah, tricky indeed!

CST: It’s not really on the todolist for me to work on unfortunately, I have a ton of other things to do with a higher priority.

Best!