Topic on Extension talk:CodeEditor

Is there a technical reason or did no one get around to writing one?

CodeEditor uses the Ace library and it dors not support wikitext highlighting.

Yes, I noticed :) But why not? Is there a technical reason why there's no syntax highlighter for wikitext or is the problem simply that nobody has tried to write one yet?

The reason is that Wikitext is not really a formal language, but a mess.

For 15 years it has been tried to create a Backus-Naur description, without full success. There is also no syntaxhighlight presentation.

However, you may search for a thing called CodeMirror.

So you think it's impossible to write a syntax highlighter without a full-fledged parser?

“Impossible” is a big word.

Theoretically, yes. Practically, low outcome.

• You might do some markup for bracket pairs.
• Interpretation is dependant on localization and state of the farm.
• Interpretation depends on localization.

Consider:

• [[CSI: Miami]]
  • Currently in article space.
  • As soon as csi will be etablished as wiki language, it is an interwiki link.
• [[Image:map.png|Legend]]
  • Media transclusion, Image: rather than File:
  • Could be localised file namespace name: [[Fichier:map.png|Legend]] is image transclusion in French.
  • If Fichier: is not file namespace name, that is linking to article or other namespace.
• {{Something}}
  • Could be anything: template transclusion, parser function, localised variable name.
• so <what> now
  • If what is known HTML or tag extension (may be localised name), this is a tag.
  • Otherwise, it is mathematical: A < B < C
• A template transclusion may be embedded within syntax and evaluates to some string under certain conditions. No way.

Huge efforts, limited resulta.

• You need to put constant set of keywords at syntax repository, but interpretation depends on local wiki environment.

That's all true, yes. A parser would need detailed knowledge about the wiki itself. That knowledge is not supremely hard come by:

https://www.mediawiki.org/w/api.php?action=query&meta=siteinfo&siprop=namespaces%7Cgeneral%7Cinterwikimap%7Cextensiontags%7Cfunctionhooks%7Cvariables%7Cfileextensions%7Cmagicwords%7Cspecialpagealiases

but still, that knowledge is required. And the parser would also need a database of extension tags, parser functions, magic words and such. That's where it gets really tedious...

I'm asking all of these questions because I'm determined to write a wikitext parser in JavaScript. Not a complete parser. It won't output HTML or resolve templates, but it will parse everything into an abstract syntax tree. It's ambitious, I know :|

You should read the (by now) heavily outdated documentation page Markup_spec.

Oh and Parsoid of course is a JS parser for wikitext.

Thank you!