Syntax highlighting

Question

For syntax highlighting, StackExchange uses a general-purpose syntax highlighter (probably something like prettify?), which guesses what to highlight based on the most common languages: C-like languages, Python, etc (and often gets highlighting of other languages wrong, e.g. when you have a Haskell function with a "prime" in it, everything after the ' is highlighted as if quoted.)

For this TeX.SE, it seems worthwhile to use a syntax highlighter that is more tuned to (La)TeX. How could we go about this?

---

A toy example of something that ought to be highlighted correctly:

\begin{enumerate}
  \item This is # not a comment
  \item This is % a comment
\end{enumerate}

A good syntax highlighter would recognize which lines are comments, and also probably do something with 'begin', 'end' and 'item' (as the LaTeX Wikibook does).

(It seems a weaker version of this has been asked before as Syntax Highlighting Hints, and there's even a somewhat simple hack that is not implemented. Irrespective of all that history, consider this another request for TeX-specific syntax highlighting on this website, not necessarily using any of the more general solutions.)

Answer 1

I had a look at the prettify, and despite my rexp-fu is seriously lacking, I pulled together something:

// PR_(STRING|KEYWORD|COMMENT|TYPE|LITERAL|PUNCTUATION|PLAIN|TAG|DECLARATION|SOURCE|ATTRIB_(NAME|VALUE)
PR.registerLangHandler(
  PR.createSimpleLexer(
    [
      ['opn', /^(?=[^\\])(?:\{|\\\[)/, null, '{['],
      ['clo', /^(?=[^\\])(?:\}|\\\])/, null, '}]']
    ],
    [
      [PR.PR_TAG, /^\\(?:text|math)?(?:rm|it|sl|tt|sc)\b/],
      [PR.PR_KEYWORD, /^\\(?:[^\\])\w*\b/],
      [PR.PR_COMMENT, /^(?=[^\\])%[^\r\n]*/],
      [PR.PR_PUNCTUATION, /^(?:\\\\|\\;|\\,|\\!|~|\ |\^)/],
      [PR.PR_PUNCTUATION, /^(?=[^\\])(?:\$){1,2}/],
      [PR.PR_PUNCTUATION, /^(?:\\\[|\\\])/],
      [PR.PR_LITERAL,
/^(?:by|at|to|spread)? ?(?:-)?(?:\d+)?\.?\d+ ?(?:true)? ?(?:pt|pc|in|bp|cm|mm|dd|cc|em|ex|mu|\\fil)\b/],
      [PR.PR_LITERAL, /^(?=[^\\])#\d?/],
      [PR.PR_LITERAL, /^(?=[^\\])[&_]/]
    ]),
  ['tex']);

And the accompanying colours from the site-design:

/* CSS */
.prettyprint { display: block; color: #393318; padding: 2px 4px;
  border: 1px solid #efefef; background-color: #f6f0df; }
.kwd { color: #8a4a0b; }
.com { color: #868686; }
.lit { color: #953838; }
.opn, .clo { color: #145680; }
.pun { color: #145680; }
.tag { color: #4c9067; font-weight: bold }

You know, something to get this train running!
I made liberal use of the "PUNCTUATION", "LITERAL", etc. because I'm not sure what they are corresponding in TeX.

Here is a pretty picture:

Answer 2

We have implemented this prettify plugin and will enable it on the next deploy (tonight). If there are changes required please update this post so we will be sure to see it.

Answer 3

Not that I don't think that in-site support for this is a bad idea, but pastie.org has support for TeX syntax highlighting in the meantime. Of course, the original question should include the code so that we don't loose everything if pastie.org goes down, and then just add a link to the same code on pastie.org that says [highlighted] or something like that at the bottom.

Answer 4

GeSHi is a generic syntax highlighter for PHP. It can be customized for using it with LaTeX code. There exist LaTeX language files.

Here's the example above with the default LaTeX syntax highlighting:

Of course it could be adjusted to the requirements and the design of our site.

Answer 5

SE is using Google Code Prettify (as mentioned on meta.SO), which doesn't yet support TeX. So the easiest way to get highlighting for this site is to add support in code prettify.