* Rename markdown to markdown-legacy
* Change how default renderWikiTextPragma value is displayed
To prevent out-of-sync, dynamically display the default value of
renderWikiTextPragma from the shadow tiddler instead of hard coding
the text in the "usage.tid".
* Repackage remarkable-based markdown plugin as markdown-legacy
- Rename plugin title to $:/plugins/tiddlywiki/markdown-legacy
- Add support for "text/markdown" MIME type and set that as the default
when creating new markdown tiddlers
* Create new markdown plugin
* add support to text/markdown MIME type
* remove linkify and linkNewWindow config options
- linkify feature should be controlled by "extlink" TW parser rule;
enabling markdown's linkify option will interfere with parsing
- remove the possibility to open external links in the same tab/window
to match TW's behavior
* Ignore latex-parser wikirule in rednerWikiTextPragma
* Prevent camel-case link text from generating a link
* Update editions/markdowndemo
* Produce better parse tree
* Improve markdown/tiddlywiki integration
- widget block should not interrupt paragraph
- ignore tw-syntax links inside markdown-syntax links
- remove repeated renderWikiTextPragma parsing
- more efficient findNextMatch when examining tw rules
* Update user docs
* Replace includes() with indexOf() for legacy browsers
* Markdown: Let WikiText parsing handle the creation of LaTeX widgets.
When embedding LaTeX snippets in inline HTML nodes, such as TiddlyRemember
macros or HTML tables, the parsing of latex nodes breaks the WikiText by
splitting it into pieces around the latex node.
This commit fixes the issue by converting the Remarkable katex nodes back to
text, using a newline to indicate a block katex snippet. This is then re-parsed
by the WikiText KaTeX plugin.
TESTED:
Created a test wiki with:
```
$ node tiddlywiki.js test --init markdowndemo
$ node tiddlywiki.js test --listen
```
* Verified markdown + KaTeX support still works as expected.
* Verified that embedding LaTeX snippets in inline HTML works (e.g. `<a
href="https://example.com/">$x^2$</a>`).
* Verified the markdown + KaTeX support works as expected with renderWikiText
set to `false`.
* Style: Remove spaces between if and opening parentheses.
* markdown: Don't emit paragraph tags when a paragraph is "tight".
Motivation: Since the upgrade to remarkable.js (#3876), lists are rendered as
HTML like this:
```
<ul>
<li><p>One</p></li>
<li><p>Two</p></li>
<li><p>Three</p></li>
</ul>
```
The paragraph nodes insert blocks that break the visual flow of the list and are
unexpected e.g. compared to WikiText markup's rendering of a bulleted list.
Solution: remarkable.js annotates certain paragraph nodes as "tight", and in the
bulleted list case, the paragraph nodes wrapping the text of each list item are
marked tight.
remarkable uses the tight property to [elide paragraph tags in its
renderer](58b6945f20/lib/rules.js (L136-L142)).
This change implements the equivalent logic in TiddlyWiki's markdown rendering:
If a paragraph is marked tight, then we elide the `<p>` tag wrapping its children.
* Use ES5 Array.concat instead of ES6 spread operator.
* Add optional KaTeX support to the tiddlywiki/markdown plugin.
Uses the remarkable-katex plugin 1.1.8 by Brad Howes to enable KaTeX support if
the tiddlywiki/katex plugin is installed. Fixes#2984.
TESTED:
Created a test wiki with:
```
$ node tiddlywiki.js test --init markdowndemo
$ node tiddlywiki.js test --listen
```
* Verified markdown support works without the tiddlywiki/katex plugin enabled.
* Verified markdown support works with the tiddlywiki/katex plugin enabled.
* Verified KaTeX (both inline and blocks) work as expected when the
tiddlywiki/katex plugin is enabled.
* Mention remarkable-katex plugin usage in the readme Tiddler.
* Include the remarkable-katex license as a tiddler.
* Include the Remarkable license.
* Include unminified original source of remarkable-katex 1.1.8.
* Corrected issue with markdowns softbreaks
It wasn't respecting the $:/config/markdown/breaks setting.
It was effectively always "true".
Also, no more errors when encountering hardbreaks.
* Corrected introduced md bug
Made it so after each md type, accumulated text may be flushed.
* Changed convertNodes to use switch statement
It's faster, less text, easier to read, and now debuggers don't play
duck-duck-goose with every if-else block.
* whitespace text is no longer forgotten
* Fixed issue with text after md pattern and \n
Fixes#855 - albeit it’s a bit of a kluge. The internal format returned
by `markdown.toHTMLTree()` isn’t sufficiently well documented for me to
be confident that the new code is resilient to all the things that
might be returned.
