Headers that are all numbers do not create ID attribute

[bulrush15]

When converting Markdown to XHTML, this Markdown header:

# 2012

Converts to this: <h1>2012</h1>

when it should convert to: <h1 id="2012">2012</h1>

Any other header that contains at least one letter work properly. I started using 4 digit years for some of my book headers and this bug popped up.

This MD header: # This is a header

Converts properly to: <h1 id="Thisisaheader">This is a header</h1>

The proper (and unique) IDs are required for XHTML when making EPUB books. I take care of the uniqueness.

[pdl]

Since XHTML does not permit IDs which are wholly numeric, I don't think the behaviour you suggest will help you in the end. You might need some other workaround.

[bulrush15]

That makes sense. I do have a workaround for this case. But the programmers could prefix or suffix the ID with a letter to make it compatible with XHML, or have an option to do this. As it stands right now, Text::MultiMarkdown is not totally compatible with XHTML.

[pdl]

It is compatible with XHTML, at least in this respect.

The Text::Multimarkdown module simply recognises that trying to consistently implement the heading IDs feature of MMD would create invalid HTML under certain conditions, and chooses standards compliance over strict consistency, which I think is a reasonable decision given that markdowns are for writing HTML documents.

I don't think there's an obviously better solution here - whatever arbitrary prefix you choose (and it needs to be a prefix, as the restriction is that the ID must begin with a letter) could cause conflicting ids anyway. The MMD test suite doesn't indicate either way what the expected behaviour is. There is an option heading_ids which if false will simply not generate the heading ids, which leaves you free to make your own, and that's what I'd do in this case if I were you.

[bulrush15]

What I meant was, this is not totally compatible with XHTML for EPUB books. In EPUB, every link destination from the table of contents, like headers, must have a unique ID. (I convert all headers in my MMD to be entries in the EPUB TOC.) To make sure of uniqueness in a whole EPUB book, I append a serial number to each header. So, I do have to do some processing of these all numeric headers after it's converted from MMD to XHTML, and then to EPUB. It's not a big deal, I fixed it on my own. Thanks for the info.