Apex has always supported Kramdown-style IAL (Inline Attribute Lists), but I’ve been steadily adding more Pandoc compatibility. The latest release brings several new features that make Apex work better with Pandoc-style markdown while maintaining backward compatibility.
Thanks to somelinguist for the suggestions!
Table Captions Get More Flexible
Table captions now support three different formats. You can use the existing [Caption] and Table: Caption syntax, plus the new Pandoc-style : Caption format:
- “`markdown
- | Col1 | Col2 |
- |——|——|
- | A | B |
-
This is a Pandoc-style caption
“`
Even better, IAL attributes in table captions are now extracted and applied to the table element itself. So you can do things like:
- “`markdown
- | Col1 | Col2 |
- |——|——|
- | A | B |
-
My Table {#table-id .highlight}
“`
This applies both the id="table-id" and class="highlight" attributes directly to the <table> element, not just the caption.
Pandoc-Style IAL Everywhere
One of the biggest changes is full support for Pandoc-style IAL syntax without the colon prefix. Where you used to need {: #id .class}, you can now use {#id .class} in all contexts:
- Block-level IALs after headings, paragraphs, and other blocks
- Inline IALs on links, images, emphasis, and other inline elements
- Pure IAL paragraphs
- Table captions
Both formats work, so your existing Kramdown documents continue to work while new Pandoc-style documents are fully supported.
Fenced Divs
Pandoc’s fenced divs are now supported in unified mode (enabled by default). You can create custom block containers with attributes:
markdown
::::: {#warning .alert}
This is a warning box with custom styling.
:::::
Fenced divs support nesting, attributes, and all the usual IAL features. You can control them with the --divs and --no-divs command-line flags.
Bracketed Spans
Bracketed spans let you create inline HTML spans with attributes using a simple syntax:
markdown
This is [some text]{.highlight} with a span.
The [text]{IAL} syntax converts to <span class="highlight">some text</span>. It supports all IAL attribute types—IDs, classes, and key-value pairs—and processes markdown inside the spans.
Reference links take precedence, so if [text] matches a reference link definition, it stays a link. Otherwise, it becomes a span. Bracketed spans are enabled by default in unified mode, and you can control them with --spans and --no-spans flags.
Image Attributes Get Smarter
Image support has been significantly enhanced. You can now use IAL syntax on both inline and reference-style images:
“`markdown
{#my-image .photo width=50%}
[ref]: image.jpg “Title” {#ref-image .thumbnail width=300px}
![alt][ref]
“`
The width and height conversion follows Pandoc’s rules:
– Percentages like width=50% convert to style="width: 50%"
– Pixel values like width=300px convert to width="300" (the px suffix is stripped)
– Bare integers like width=300 stay as width="300"
– Other units like 5em or 10rem convert to style attributes
– Decimal pixel values like 100.5px also go to style attributes
This means you get clean, semantic HTML output that works well across different contexts. Reference image definitions also now preserve title attributes, so [ref]: url "title" {#id} includes the title in the final output.
Better IAL Detection
I’ve improved IAL detection throughout the codebase. IAL syntax with spaces (like { width=50% }) now works correctly, and IALs are properly stripped from output even when parsing fails, preventing raw IAL syntax from appearing in your HTML.
All of these features work together seamlessly. You can mix Kramdown and Pandoc syntax in the same document, use IALs everywhere, and get clean, semantic HTML output. The test suite has been expanded to cover all these new features, so everything should work reliably.
Check out the latest release for full release notes. And if you haven’t tried Apex yet, check out the Apex project and help me build the ultimate Markdown processor!
Leave a Reply