On 5/31/07, Andreas Hartmann <[EMAIL PROTECTED]> wrote:
Richard Frovarp schrieb:
> Is there a way to call a page via it's UUID from a browser? My issue is
> that we translate from UUIDs to URLs when a page is being edited.

That's a very good point.

> If an
> object being targeted (image or link) changes path while the page is
> being edited, it then becomes lost. Links really don't matter if they
> are left as lenya-document as they aren't rendered. However, images need
> to be rendered. If there is some way of accessing a document via HTTP
> using it's UUID, a pre-processing instruction can be put in to handle
> creating those URLs, and things will be good.

We should certainly introduce such a URL space. It will also be helpful
for debugging. Should we use a particular prefix, or does the UUID space
itself provide enough safety from name clashes?
  /{pubId}/{area}/{uuid}.png

In theory, this would be sufficient:
- It is very unlikely to have a URL with this syntax which is
  not a UUID.
- If the "natural" URL space uses UUIDs, there is no risk for clashes
  because UUIDs are unique.

The problem is that we'd need to check if a document with this
UUID exists for each call which matches
  /{pubId}/{area}/????????-????-????-????-????????????
to avoid masking the "natural" URL space of the publication.
Considering this, it might make sense to use a particular prefix, e.g.:
  /{pubId}/{area}/lenya-uuid/{uuid}.png

WDYT?
-- Andreas


This feature was implemented in 1.3.  The design used one protocol
(cleverly named "content:") for access to all documents and resources.
The protocol finds the resource by testing a variety of formats
including the UUID. (UUIDs are strictly formatted UNIDs.)  See the
second half of 13HELP for the specs:
http://svn.apache.org/repos/asf/lenya/branches/revolution/1.3.x/13HELP.txt

XMAPs can just pass the resource identifier portion of the URL to the
protocol and retrieve the correct resource.  That avoids issues about
needing a special prefix.  The protocol works fine internally so no
other code cares whether the identifier is a path or UNID.

Links should always be stored as UNIDs.  Editors must translate
path-based identifiers to UNIDs before storing documents.  The "Add
link" and "Add image" functions should display resource titles to hide
UNIDs from users.

solprovider

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to