Accessing parent sitemaps with cocoon:../

[Jeff Turner <je...@apache.org>]

Hi,

The current sitemap is available through cocoon:/foo
Root sitemap is available through cocoon://foo
How about making parent sitemaps available through cocoon:../foo,
cocoon:../../foo, etc?

Context: I'm currently trying to get Forrest working in samples/forrest,
and find myself having to hardcode 'cocoon://samples/forrest/...' refs
everywhere, because subsitemaps need access to pipelines in the main
forrest sitemap.  Having a cocoon:../ syntax would make things much more
portable.


--Jeff

Re: Accessing parent sitemaps with cocoon:../

[Geoff Howard <co...@leverageweb.com>]

At 09:25 AM 3/26/2003, you wrote:
>Jeff Turner wrote:
>>Hi,
>>The current sitemap is available through cocoon:/foo
>>Root sitemap is available through cocoon://foo
>>How about making parent sitemaps available through cocoon:../foo,
>>cocoon:../../foo, etc?
>>Context: I'm currently trying to get Forrest working in samples/forrest,
>>and find myself having to hardcode 'cocoon://samples/forrest/...' refs
>>everywhere, because subsitemaps need access to pipelines in the main
>>forrest sitemap.  Having a cocoon:../ syntax would make things much more
>>portable.
>
>I've been historically against this in order to limit about and force 
>better SoC at the URL-space level, but if you really found a need for it, 
>I'd be happy to change my vote from -1 to +0.
>
>Stefano.

I've needed this in the past and wasn't aware there was an opinion against 
it.  I think I see the concern with SoC but surely hardcoding the path is 
as bad or worse.  Is there another solution?  Syntax aside, would it be 
better to limit access to root, current and immediate parent (which maybe 
could default to root?).

Geoff 

Re: Accessing parent sitemaps with cocoon:../

[Stefano Mazzocchi <st...@apache.org>]

Jeff Turner wrote:
> Hi,
> 
> The current sitemap is available through cocoon:/foo
> Root sitemap is available through cocoon://foo
> How about making parent sitemaps available through cocoon:../foo,
> cocoon:../../foo, etc?
> 
> Context: I'm currently trying to get Forrest working in samples/forrest,
> and find myself having to hardcode 'cocoon://samples/forrest/...' refs
> everywhere, because subsitemaps need access to pipelines in the main
> forrest sitemap.  Having a cocoon:../ syntax would make things much more
> portable.

I've been historically against this in order to limit about and force 
better SoC at the URL-space level, but if you really found a need for 
it, I'd be happy to change my vote from -1 to +0.

Stefano.