OK, I checked this, but SecurityImpl hasPermission is never used as
they are overridden by ACLSecurityImpl which is the implementation
used. So, no need to change anything here. getActionNode in
NamespaceConfig *is* used, but only in initialization I guess, which
should not cause any problem as well.

I have deprecated methods without tokens that access the stores in
Security and getUri in Namespace.

Oliver 


On Fri, 29 Oct 2004 10:38:47 -0700, Warwick Burrows
<[EMAIL PROTECTED]> wrote:
> 
> Oliver,
> 
> The solution is to replace the calls to
> org.apache.slide.common.Namespace.getUri(String) in both hasPermission()
> methods in SecurityImpl.java with the companion method from the same class
> getUri(SlideToken, String). The safest option is to replace all getUri()
> calls  in this way so that we don't run into the problem again. Ideally it
> would be best to replace _all_ calls to getUri(String) in the entire Slide
> tree and then remove the getUri(String) method so that noone can
> accidentally use it in the wrong circumstances again. The only other place I
> see that getUri(String) is used other than hasPermission() is in
> org.apache.slide.common.NamespaceConfig in getActionNode() so if it can also
> be replaced then getUri(String) can be removed.
> 
> The trick to replacing the getUri(String) call is that you then need to have
> the SlideToken for the current request handy to pass into getUri(SlideToken,
> String). If it is already available in one of the parameters to
> hasPermission() then great. I wasn't sure. Otherwise it may need to be
> passed down the call sequence somehow or fetched from request-specific
> memory. Does Slide maintain request specific context of some kind that the
> SlideToken could be attached to? Then the token wouldn't have to be passed
> along it could be fetched from the request specific store when needed.
> 
> 
> 
> Warwick
> 
> > -----Original Message-----
> > From: Oliver Zeigermann [mailto:[EMAIL PROTECTED]
> > Sent: Friday, October 29, 2004 11:46 AM
> > To: Slide Users Mailing List
> > Subject: Re: Server seems to hang on create new
> > WebdavResource... (yikes, still hung)
> >
> >
> > > Oliver, you implemented the transactional file system
> > store, correct?
> > > Do you know whether the other stores acquire locks at the
> > start of a
> > > transaction as seems to be the case with the DB store? If
> > so then this
> > > could be the same problem (31907) or the access control variant
> > > (31908)?
> >
> > Yes, the tx store does locking very similar to the RDBMS
> > stores. Concerning 31908, how can we fix this? Have you
> > already digged into it?
> >
> > Oliver
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> >
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> 
>

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

Reply via email to