Feedback on PseudoXmlPlugin

This plugin definitely isn't production ready; it can't handle nested tags. Still, the simple functionality it provides met my modest needs. Feel free to improve the code

-- WillNorris - 12 Oct 2003

1. I wonder if it is possible to add the ending tag automatically when a newline is encountered, so that

<subtitle>My necessary subtitle

would be renamed: <span class="subtitle">My necessary subtitle</span>

• this would prevent multiline tags from working (not that they work at the moment) -- WillNorris - 14 Oct 2003
• would a preferences item listing a set of tags to autoclose at eol be sufficient? -- WillNorris - 20 Oct 2003

2. I would suggest to use "class" instead of "id" as id should be used only once per page.

• the plugin already outputs class unless you start the xml-like tag with # (e.g., <#id>id text</id>), or did i misunderstand? -- WillNorris - 14 Oct 2003

-- ArthurClemens - 12 Oct 2003

Version 1.001 handles nested tags now, but i may have significantly impacted the performance?

Also, I'm open to another name besides PseudoXmlPlugin (it started as SimpleXmlPlugin; SimplePseudoXmlPlugin also came to mind)

-- WillNorris - 14 Oct 2003

On nested tag, what about the following algorithm (I didn't look at your code yet):

1. when encountering <[a-zA-Z_0-9]+>, look for tag in a table of allowed tags (HTML spec + wiki pseudo tags + local additions)
   • if not found, replace by <span class="\1">
2. when encountering <[/[a-zA-Z_0-9]+>, look for tag in same table
   • if not found, replace by </span>

You may want to allow class and ID, as in:

<date#34>11 Oct 2003</date> would be renamed: <span class="date" id='date#34'>11 Oct 2003</span>

(not id=34 as it can conflict).

But I do not know if IDs are useful. they are potentially troublesome in case of conflicts. And users could just as well use <date_34>11 Oct 2003</date> to get this functionality without any need to code a special case.

-- ColasNahaboo - 14 Oct 2003

colas, for your example, did you mean to say <date_34>11 Oct 2003</date_34>? as written, those tags wouldn't be expanded. alternatively, it could be written <date#date_34>11 Oct 2003</date>. perhaps more verbose, but at the same time more flexible. the class and id combo is a good idea, but i'm not sure which format expansion is "right".

-- WillNorris - 20 Oct 2003

Well, since I propose to replace any XXX in </XXX> to </span>, anything will work, even... converting </> to </span>, allowing for the traditional sgml shorthand <date_34>11 Oct 2003</>

-- ColasNahaboo - 21 Oct 2003

I was going to suggest exactly the same thing until I read the above. I also thought of <> first as it doesn't require messing around with the shift key :). I wasn't aware </> was a sgml tag so maybe it is more logical but I'd still prefer <> for speed.

-- SamHasler - 20 Jan 2004

REX: XML Shallow Parsing with Regular Expressions

-- WillNorris - 21 Mar 2004

checked into CVS

-- WillNorris - 14 Feb 2005