On Apr 19, 2007, at 10:25 AM, Martin Marinschek wrote:
oh, yes. I had overlooked the return null at the end of the method -
that was different before.
Brief update here - I ended up changing the behavior of
ManagedBeanResolver to go ahead and return the managed bean right
away due to :
Sure!
regards,
Martin
On 4/19/07, Paul McMahan [EMAIL PROTECTED] wrote:
Cycle reference check should be fixed now in r530517. thanks again for the
peer review.
Best wishes,
Paul
On Apr 19, 2007, at 10:25 AM, Martin Marinschek wrote:
But still - you are short-circuiting the cyclic
Thanks for the review Martin. I agree that the persistence mechanism
for managed beans was already working OK. i.e. when a managed bean
has scope none it was not persisted in a scope. But the problem I
encountered was that the CompositeELResolver was not able to resolve
managed beans
Hi Paul,
oh, yes. I had overlooked the return null at the end of the method -
that was different before.
But still - you are short-circuiting the cyclic reference check now -
if I have a bean now which has scope none, and has a managed-property
referring to the bean again, then we'll run into
On Apr 19, 2007, at 10:25 AM, Martin Marinschek wrote:
But still - you are short-circuiting the cyclic reference check now -
if I have a bean now which has scope none, and has a managed-property
referring to the bean again, then we'll run into an infinite loop,
right?
Yes that's true, and
Cycle reference check should be fixed now in r530517. thanks again
for the peer review.
Best wishes,
Paul
On Apr 19, 2007, at 10:25 AM, Martin Marinschek wrote:
But still - you are short-circuiting the cyclic reference check now -
if I have a bean now which has scope none, and has a
Just wanted to invite some peer review for this change I just
committed for MYFACES-1588. The problem was that managed beans in
scope none weren't accessible via the resolver. The change I made
passes the test cases but there might be a more elegant way to
implement it.
Also, I have an
I don't think anyone has run the cactus tests in about six months. They
aren't a part of the CI loop either.
Dennis Byrne
On 4/18/07, Paul McMahan [EMAIL PROTECTED] wrote:
Just wanted to invite some peer review for this change I just
committed for MYFACES-1588. The problem was that managed
right,
I think they used to be Bill's sandbox ;)
-M
On 4/18/07, Dennis Byrne [EMAIL PROTECTED] wrote:
I don't think anyone has run the cactus tests in about six months. They
aren't a part of the CI loop either.
Dennis Byrne
On 4/18/07, Paul McMahan [EMAIL PROTECTED] wrote:
Just wanted to
Hi Paul,
if you do the first change (introduce a scope where put does nothing),
I don't see why the second one needs to be done - putting will do
nothing, so you don't need the extra-check for none, right?
regards,
Martin
On 4/18/07, Paul McMahan [EMAIL PROTECTED] wrote:
Just wanted to
10 matches
Mail list logo
