Jörn Nettingsmeier schrieb:
> hi!
>
>
> i'm trying to implement proper checkout (aka locking) for tinymce.
> andreas told me to
>
>> - set the transaction policy of the usecase to "pessimistic"
>> - return the document's repository node in getNodesToLock()
>
> in order to find out how to do that, i looked for clues in other
> modules. in webdav, i found the Put.java usecase handler, and i figured
> the same locking that's good for webdav uploads should be ok for an
> editor usecase as well. but it seems there is a whole bunch of nodes
> being locked:
>
> /**
> * @see org.apache.lenya.cms.usecase.AbstractUsecase#getNodesToLock()
> */
> protected Node[] getNodesToLock() throws UsecaseException {
> try {
> Document doc = getSourceDocument();
> List nodes = new ArrayList();
> NodeSet set = SiteUtil.getSubSite(this.manager,
> doc.getLink().getNode());
> Document[] documents = set.getDocuments();
> for (int i = 0; i < documents.length; i++) {
> nodes.add(documents[i].getRepositoryNode());
> }
That's indeed strange, I don't know why the descendants of the
document are locked.
> SiteStructure structure = getSourceDocument().area().getSite();
> nodes.add(structure.getRepositoryNode());
Maybe locking the site structure is not even necessary, since the
WebDAV upload or any other editing step doesn't compromise the
site structure.
> return (Node[]) nodes.toArray(new Node[nodes.size()]);
> } catch (Exception e) {
> throw new UsecaseException(e);
> }
> }
>
> let me see. i edit a document
> "mysite.org/mypub/authoring/some/path/to/foo.html".
> does that mean i have to lock "/some.html", "/some/path.html",
> "/some/path/to.html" and finally "/some/path/to/foo.html"?
No, not at all.
> i can see how that made sense with the old repository implementation,
> but is it really necessary now that we have moved to UUIDs and each
> document is addressed directly?
> the answer probably depends on how we want to handle sitetree updates.
> there was some discussion about this a while ago, but i can't remember
> what was decided.
Me neither ...
> if i lock all those nodes, i can be sure that nobody makes changes to
> the relevant sub-sitetree while somebody is editing a file. but it also
> means that when someone is editing *any* document, you can't edit any of
> its ancestors, right?
I guess it would be sufficient to lock the to-be-edited document itself.
-- Andreas
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
