💾 Archived View for warmedal.se › ~bjorn › posts › wishing-for-stricter-gemtext.gmi captured on 2023-12-28 at 15:40:54. Gemini links have been rewritten to link to archived content

View Raw

More Information

⬅️ Previous capture (2021-11-30)

-=-=-=-=-=-=-

Wishing for Stricter Gemtext

This is going to be an unpopular opinion, but here goes. (I also want to apologize in advance to anyone using a screen reader; this text discusses gemtext syntax and I assume screen readers will be unable to discern the subtle differences in examples.)

I've already written a tool for gemtext to HTML conversion. But for reasons I now feel the need for a more general parser.

Now, the aforementioned tool uses a whole lot of regex magic to figure out which lines are what. And then there's of course some logic to handle whether we're inside a preformatted block or not (and because it translates to HTML there's also similar logic for whether we're inside a list or not, and I should probably add for whether or not we're in a blockquote block, but those issues are related to HTML rather than gemtext).

Anyway, that was then and now is now, and when I look at that code now and my current needs I feel a little unsatisfied with the needless parsing complexity of gemtext. Because even though it's very easy, it's still harder than it could be.

Consider these examples:

So, when parsing gemtext I need to consider the first one to three characters of the line, and whether or not it is within a preformatted block.

What if I always, without exception, just had to check the first three characters? The linetypes could be something like this:

#  First level heading.
## Second level heading.
###Third level heading.

=> Link line.
>  Block quote line.
>> Preformatted line.
And of course any other line, which is just plain text.

The URL and text on a link line should be separated by a single blankspace, not an arbitrary number of any type of whitespace.

I also think that plain text lines should be trimmed of leading and trailing whitespace when rendered, simply because I don't see any reason to keep them, and it would also allow plain text lines to start with a leading space followed by a character combination that would otherwise cause the line type to be different.

Maybe I should have sent this to the mailing list, but it's not a proposal for change. It's just unfinished thoughts that I'd like feedback on.

-- CC0 ew0k, 2021-07-23