Dan Allen

At the moment, the Asciidoctor processor isn't really a source to source processor. I've done that on various projects, but it's always a bit hacky as that's not been a...

Basically, what I do is parse the document with the sourcemap option enabled and parse option disabled. Then I grab the source lines, then parse the document. I then use...

> Is this something which must wait for #1793? I'd say not really because generating / manipulating a source document is very different from generating output documents. There is information...

> Each of these only produces the source document without modifications, even though doc.convert produces html5 with the correct changes. What I had originally suggested was on-the-fly AST transformations. Those...

In other words, the title of this issue is actually two different things. > Creating document programatically There's (a) creating the document programmatically in order to produce and output document,...

Nope. The best way to describe is that the processor was built with a single goal in mind, to make an output document. Creating an input document requires additional information...

FYI, the asciidoc-coalescer.rb script has been redesigned and published as the [asciidoctor-reducer](https://github.com/asciidoctor/asciidoctor-reducer). It has also been released to RubyGems.org as a gem named **asciidoctor-reducer**. You can find installation instructions in...

AsciiDoc.py supports this notation, though it outputs the CSS into the style attribute on the HTML table tag.

One open question is how to handle left and right alignment. Currently, the left and right roles float the element to the left and right, respectively. That gives different behavior...

This customization was first proposed in Asciidoctor PDF. https://github.com/asciidoctor/asciidoctor-pdf/issues/858#issuecomment-529310500 In that issue, I suggested that the change should be made to core (here). So let's consider the design proposed in...