On Thu, 2005-07-14 at 18:07 +0200, Andreas Hartmann wrote:
> Thorsten Scherler wrote:
>
> [...]
>
> > should become
> > <map:read src="{lm:resources/shared/css/{3}.css}"
>
> How would that be used from Java classes?
>
> Like I would use sourceResolver.resolveURI("fallback://..."),
> is there a Java API for LocationMaps, or is it accessible as an
> InputModule only?
>
org.apache.forrest.locationmap.*
I am on my way out. I do not know whether it is only accessible as an
InputModule.
>
> > mime-type="text/css" />
> >
> > Then you can have in your locationmap:
> > <match pattern="resources/**.css" type="request">
> > <select type="exists">
> > <location src="pub/..." />
> > <location src="parent/" />
> > <location src="context://lenya/...">
> > </select>
> > <location src="fallback://resources..." />
> > </match>
> >
> > This way each pub can define their own fallback mechanism *without*
> > inventing new protocols all the time.
>
> IMO that's just another configuration of the publication template.
> A benefit is that we would use existing technology to implement
> fallback, but it adds complexity to the configuration.
>
hmm, adding complexity? I reckon it is doing the opposite. I would
recommend to have it in the core and each pub can have their own
overriding map.
...and using existing technology is IMO better to invent lenya specific
protocols.
The last point is that our 'configuration' is hell complicated and the
locationmap seems to help solving some aspects.
> But there is a major problem:
>
> Publication templating is an abstract concept which is an integral
> part of Lenya. The fallback mechanism for files is just one aspect.
> We need a central configuration point for a template (which cannot
> be a location map), otherwise we'll run into trouble.
>
I need to get into publication templating before I can say something to
it.
> Another point: How would loops be implemented?
> You have a chain of parents. Is it possible to use recursion?
>
The example has a loop.
> ----
>
> BTW, what I don't yet see is how this replaces the parent:// protocol?
>
Then please explain again what parent:// should do. Maybe I
misunderstood it.
> Thanks for the example,
>
de nada.
salu2
> -- Andreas
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
--
thorsten
"Together we stand, divided we fall!"
Hey you (Pink Floyd)
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
