Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Christophe David
> So, I again say that this would probably be better done by
> creating something like ?action=xml
> some sort.  (And in fact this is exactly what ?action=rss,
> ?action=rdf, ?action=atom do -- they produce xml-based views
> of pagelists.)


You are right.  The layout of the output when using these options is perfect.
What we need then is something that does the same as action=browse
(processing pagelists, conditonals, includes, etc.), but outputs the
result "as is".

I just had a look at feeds.php to see if I could try to do it myself,
but it does not appear to be a trivial task.

If you could give me a hint about the variables and fucntions that are
(not) needed to produce this result, I am prepared to spend some time
on this.

...unless of course you have a solution at hand.

Thanks a lot for your help.

Christophe

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 07:01:32PM +0200, Christophe David wrote:
> >Sure -- that's what the example I gave demonstrates -- see
> >   http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml
> >None of the css or formatting or HTML head/body/title/etc. elements
> 
> Not really :
> 
> See http://www.pmwiki.org/wiki/Test/XMLTest2?skin=xml
> 
> then, view source.
> ...
> 
> 
> 
> that are not acceptable in XML...

(Note that  are valid XML.)

I'm not sure that page is doing at all what you think it's doing.
In particular, the tags , , etc. are
not at all XML tags.  In the output source they're actually
coming across as "" and "" --
i.e., with the angle brackets escaped.

So, in order for this to produce actual XML output you'd have
to do the equivalent of EnableHTML on the page, so that HTML
sequences can be embedded directly.

But really I think this means that ?action=browse is not
the thing we want here, because ?action=browse and (:pagelist:)
are explicitly designed to produce XHTML, not XML, and they're
intended to escape any angle brackets and other character
entities coming from markup.

So, I again say that this would probably be better done by
creating something like ?action=xml, or perhaps ?action=pagelist
if the intent is to always produce an XML-based pagelist of 
some sort.  (And in fact this is exactly what ?action=rss,
?action=rdf, ?action=atom do -- they produce xml-based views
of pagelists.)

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 11:35:34AM -0500, Patrick R. Michaud wrote:
> On Fri, Jul 20, 2007 at 06:29:08PM +0200, Christophe David wrote:
> > But up to now, I could not manage to get rid of css and formatting
> > things that are not in my skin.
> > 
> > It this possible ?
> 
> Sure -- that's what the example I gave demonstrates -- see
>http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml
> ...

In fact, you can see how I created the skin by looking at
the files at

http://www.pmwiki.org/pmwiki/pub/skins/xml/

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Christophe David
> Sure -- that's what the example I gave demonstrates -- see
>http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml
> None of the css or formatting or HTML head/body/title/etc. elements

Not really :

See http://www.pmwiki.org/wiki/Test/XMLTest2?skin=xml

then, view source.

There are still things like




that are not acceptable in XML...

So we are close to it, but there are still a few things we would need
to get rid off when using skin=xml.

...hence my original question ;-)

If this could be done, it would open fantastic oportunities.

A simple example: Flash applications that need XML configuration files
could get them from PmWiki instead of hard coded XML text files,
making their maintenance or customizaton a breeze.

Christophe

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 06:29:08PM +0200, Christophe David wrote:
> >Okay, you're asking for something very different from what PmWiki
> >produces.  PmWiki doesn't have markup rules for producing 
> 
> Well, I would already be very happy to create the pagelist templates
> myself to generate
> 
> 
>   abc
>   cde
> 
> 
> as long as I could use a skin that would set the http header and not
> generate any other output than this (and possibly an  tag.
> 
> But up to now, I could not manage to get rid of css and formatting
> things that are not in my skin.
> 
> It this possible ?

Sure -- that's what the example I gave demonstrates -- see

   http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml

None of the css or formatting or HTML head/body/title/etc. elements
are in the output, only the output of the pagelist command.
(View the document source to see what it has -- don't try to
look at it with an XML parser, since it obviously won't parse
as XML yet.)

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Christophe David
> Okay, you're asking for something very different from what PmWiki
> produces.  PmWiki doesn't have markup rules for producing 

Well, I would already be very happy to create the pagelist templates
myself to generate


   abc
   cde


as long as I could use a skin that would set the http header and not
generate any other output than this (and possibly an  tag.

But up to now, I could not manage to get rid of css and formatting
things that are not in my skin.

It this possible ?

Christophe

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 06:07:07PM +0200, Christophe David wrote:
> >Wouldn't we also need at least a  tag at
> >the beginning?
> 
> The specs are :
> "The root element can be preceded by an optional XML declaration. This
> element states what version of XML is in use (normally 1.0); it may
> also contain information about character encoding and external
> dependencies.
> 
> "
> 
> So yes, this may be added.
> 
> 
> >As a start, see
> >  http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml
> 
> I cannot check easly the http header right now, but it needs to be text/XML

I think the canonical form is "text/xml", and yes, the skin is
already setting the http header.

> To please the XML parser, the data may not contain '<' or '>', unless
> enclosed between
> 
> 
> 
> So we want in this example
> 
> "
> 
> 
> 
> 
> 
> 
> 
> 

Okay, you're asking for something very different from what PmWiki
produces.  PmWiki doesn't have markup rules for producing 
elements, or things like , or placing things inside of
CDATA contexts.  We could possibly create such markups, but that's
quite a bit more complex than a simple skin would typically
provide.

It seems that what you're really looking for is something more
along the lines of an RSS feed, which _does_ generate XML output
directly from pagelists.  And feeds.php is fully configurable, 
meaning it's possible to define custom tags and components in 
the XML output.

See http://www.pmwiki.org/wiki/PmWiki/WebFeeds .

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Christophe David
> Wouldn't we also need at least a  tag at
> the beginning?

The specs are :
"The root element can be preceded by an optional XML declaration. This
element states what version of XML is in use (normally 1.0); it may
also contain information about character encoding and external
dependencies.

"

So yes, this may be added.


> As a start, see
>   http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml

I cannot check easly the http header right now, but it needs to be text/XML

To please the XML parser, the data may not contain '<' or '>', unless
enclosed between



So we want in this example

"









Christophe


---

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 04:22:51PM +0200, Christophe David wrote:
> >Sure... just create a skin for it, and omit the 
> >and  directives.
> >I'll be glad to create one for this, if someone can let me
> >know what belongs and doesn't belong in the resulting output.
> 
> Well, basically, a page containg
> 
> 
>   123
> 
> 
> shown with the "XML" skin should return
> 
> an HTTP header "Content-Type: text/XML"
> 
> then nothing else than the text above.

Wouldn't we also need at least a  tag at 
the beginning?

As a start, see

  http://www.pmwiki.org/wiki/Test/XMLTest?skin=xml

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Christophe David
> Sure... just create a skin for it, and omit the 
> and  directives.
> I'll be glad to create one for this, if someone can let me
> know what belongs and doesn't belong in the resulting output.

Well, basically, a page containg


   123


shown with the "XML" skin should return

an HTTP header "Content-Type: text/XML"

then nothing else than the text above.

this would open in a browser as XML with the default XSL of the browser.

On my current installation, I get other formatting code, possibly
coming from recipes, even with a stripped down skin.

Christophe

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users


Re: [pmwiki-users] PmWiki as XML source ?

2007-07-20 Thread Patrick R. Michaud
On Fri, Jul 20, 2007 at 03:06:03PM +0200, Christophe David wrote:
> Many applications use XML formatted data obtained from a web server.
> 
> PmWiki pagelists formats are so powerful that the information returned
> could be used by other applications expecting XML, if the there was a
> way let PmWiki return  "Content-Type: text/XML" in the HTTP header,
> and to remove the stylesheet stuff that is returned "around" what is
> defined in the skin template.
> 
> Is there a way to do that ?

Sure... just create a skin for it, and omit the 
and  directives.

I'll be glad to create one for this, if someone can let me
know what belongs and doesn't belong in the resulting output.

Pm

___
pmwiki-users mailing list
pmwiki-users@pmichaud.com
http://www.pmichaud.com/mailman/listinfo/pmwiki-users