Semantic syntax highlighter (without inline CSS)

Question

I’m looking for a tool that takes code as input and outputs this code as syntax highlighted HTML+CSS.

It can either be a local tool (running natively on GNU/Linux) or a self-hosted web app (Linux server). In either case, it must be FLOSS.

Input

I only need support for "Web languages" (i.e., XML, HTML, CSS and JavaScript). But support for more languages is welcome, of course.

The tool doesn’t have to guess the language; I’m fine with selecting it.

Output

It should escape characters that might conflict with (X)HTML, e.g. < as <.

It should use only div and span elements to markup the code.

It should not use the style attribute! Either use the style element (so I can copy the CSS and paste it into my stylesheet) or offer a separate CSS file.

For that, it should use class attributes (data-* is okay, too). At best it uses semantic class values, and some kind of prefix to avoid name conflicts.

Just to give a poor example, here is an HTML input followed by an acceptable output:

<p>Lheben teil brann i annon ar <b id="foo">neledh</b> neledhi gar godrebh.</p>

<span class="sh-html-tag">&lt;p&gt;</span>Lheben teil brann i annon ar <span class="sh-html-tag">&lt;b <span class="sh-html-attribute">id="foo"</span>&gt;</span>neledh<span class="sh-html-tag">&lt;/b&gt;</span> neledhi gar godrebh.<span class="sh-html-tag">&lt;/p&gt;</span>

Answer 1

I once wrote that for iFAQMaker, which is no longer maintained. An example of code output you can find in the Demo page:

Input are basically text files with macros supported. This example uses the sqlfile macro, resulting in the text syntax highlighted as SQL, plus a link to download the original file. There are *file macros available for SQL, PHP, and Shell, and you can extend the API with additional ones if needed.

How does it fit your requirements:

• self-hosted web app (Linux server): iFAQMaker is a PHP application/script.
• FLOSS: Yes (GPLv2)
• XML, HTML, CSS and JavaScript: As it ships, only SQL, PHP, and Shell. Given a keyword list, you can easily add all languages you want.
• escape characters: Yes, it does.
• should use only div and span elements to markup the code: Yupp.
• not use the style attribute: Nope. Uses CSS classes with a specific (external) stylesheet

Not a perfect match (as you'd have to create the keyword lists for your major languages first), but I hope it's useful nevertheless. I'm using it a lot on my websites for exactly this: displaying scripts inline syntax-highlighted, while at the same time offering them for download.