On Feb 9, 2009, at 6:56 PM, Thomas Mortagne wrote:

Hi devs,

Currently when a document is saved it is rendered with a custom URLFactory to catch links.

IMO this is wrong because: - it's impossible to really support generated links since the document is renderer with a specific context in a specific time etc... - it could makes saving document real slow if the document contains big scripts which should not have to worry about that kind of very crappy rendering

I understand it was easier to do this that way for 1.0 since it was difficult to parse the document content to find links but the new rendering does not have this issue anymore. It's now very easy to get the XDOM and all the links it contains.

So I propose to only take the links found in XDOM before the transformation.

Here is my +1

+1 Thanks -Vincent http://xwiki.com http://xwiki.org http://massol.net

Thomas Mortagne wrote:

Hi devs,

Currently when a document is saved it is rendered with a custom URLFactory to catch links.

IMO this is wrong because: - it's impossible to really support generated links since the document is renderer with a specific context in a specific time etc... - it could makes saving document real slow if the document contains big scripts which should not have to worry about that kind of very crappy rendering

I understand it was easier to do this that way for 1.0 since it was difficult to parse the document content to find links but the new rendering does not have this issue anymore. It's now very easy to get the XDOM and all the links it contains.

So I propose to only take the links found in XDOM before the transformation.

+1 Also, generated links are not real connections/dependencies. A search result does not really establish a link to a document, nor does Most Recent Changes, so we should not persist this temporary/contextual link between two documents. -- Sergiu Dumitriu http://purl.org/net/sergiu/