Sounds like what they want is what collaborationlistcontent/handler
does, specifically, at least for now.
On 09/04/17 06:30, James Hare wrote:
Why, exactly, do you want a wikitext intermediary between your JSON and
your HTML? The value of wikitext is that it’s a syntax that is easier to
edit than HTML. But if it’s not the native format of your data, nor can
browsers render it directly, what’s the point of having it?
The CollaborationKit extension [0] has two content models,
CollaborationHubContent and CollaborationListContent, and they have two
different strategies for parsing. CollaborationHubContent takes validated
JSON and puts out raw HTML; this is the most straightforward to work with.
CollaborationListContent instead puts out wikitext that is fed into the
parser, since it is expected that lists can be transcluded onto other
pages, and this means using wikitext (as transcluded HTML is not an
option). However, this creates a lot of limitations, including parser
restrictions that make sense in the context of arbitrary wikitext parsing
but not when the markup is provided directly by an extension. The long term
plan is for CollaborationListContent to put out HTML, since it’s more
straightforward than using a wikitext intermediary that the user does not
see anyway.
[0] https://www.mediawiki.org/wiki/Extension:CollaborationKit
On April 8, 2017 at 11:23:38 PM, Denny Vrandečić (vrande...@gmail.com)
wrote:
Here's my requirement:
- a wiki page is one JSON document
- when editing, the user edits the JSON directly
- when viewing, I have a viewer that turns the JSON into wikitext, and that
wikitext gets rendered as wikitext and turned into HTML by MediaWiki
I have several options, including:
1) hook for a tag like <json>, and write an extension that parses the
content between the tags and turns it into wikitext (not ideal, as I don't
use any of the existing support for JSON stuff, and also I could have
several such tags per page, which does not fit with my requirements)
2) I found the JsonConfig extension by yurik. This allows me to do almost
all of the things above - but it returns HTML directly, not wikitext. It
doesn't seem trivial to be able to return wikitext instead of HTML, but
hopefully I am wrong? Also, this ties in nicely with the Code Editor.
3) there is actually a JsonContentHandler in core. But looking through it
it seems that this suffers from the same limitations - I can return HTML,
but not wikitext.
3 seems to have the advantage to be more actively worked on that 2 (which
is not based on 3, probably because it is older than 3). So future goodies
like a Json Schema validator will probably go to 3, but not to 2, so I
should probably go to 3.
Writing this down, one solution could be to create the wikitext, and then
call the wikitext parser manually and have it create HTML?
I have already developed the extension in 1, and then fully rewritten it in
2. Before I go and rewrite it again in 3, I wanted to ask whether I am
doing it right, or if should do it completely differently, and also if
there are examples of stuff developed in 3, i.e. of extensions or features
using the JsonContent class.
Example:
I have a JSON document
{ "username": "Denny" }
which gets turned into wikitext
''Hello, [[User:Denny|Denny]]!''
which then gets turned into the right HTML and displayed to the user, e.g.
<i>Hello, <a href="...">Denny</a>!</i>
Cheers,
Denny
_______________________________________________
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
_______________________________________________
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l
_______________________________________________
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l