On 4/11/07, Jukka Zitting <[EMAIL PROTECTED]> wrote:
Hi,
On 4/10/07, Alexandru Popescu ☀ <[EMAIL PROTECTED]> wrote:
> On a more theoretical level: isn't this a limitation of the spec? By
> the time you remove the event the path information attached to the
> event is quite useless, so the mechanism should try to publish more
> usefull information.
I think the remove events are most useful for ensuring coherency for
applications that cache parts of the repository content (a good
example would be an external search engine). Cache eviction only needs
the item path as the key of the cached entry.
... if you used that as a key and not the UUID or a "natural key"
based on the node properties.
What use cases do you think would benefit from more information in the
event records?
I don't have real scenarios right now, but I am seeing Sudhan example,
and I guess we can imagine quite a few (take for example workflow
notifications). Probably the best alternative would be to publish
through the event a shallow copy/dettached copy of the deleted node,
so that you still have most of the information available. I know that
this can be solved in your app with custom made code around special
removal ops, but if you look at this as a cross cutting concern, then
having more information published through event would lead to a lot of
possibilities.
bests,
./alex
--
.w( the_mindstorm )p.
_____________________________________
Alexandru Popescu, OSS Evangelist
TestNG/Groovy/AspectJ/WebWork/more...
Information Queue ~ www.InfoQ.com
BR,
Jukka Zitting
