The “NoChange” event means
that we don’t know the type of the object we’re trying to bind from
(or it’s a getter/setter with no change event). Usually when we insert
that event it’s accompanied by a binding warning to indicate that the
type of the property is unknown (I forgot to ask if you have binding warnings
enabled and whether they come out in the failure case, I’m guessing no
which is a real shame). Since we can’t reliably detect the type of the property
I think we stop creating watchers for the deeper properties since there’d
be no point. When we do know the type, the extra watchers would then be
created.
Matt
From: Joe Berkovitz
[mailto:[EMAIL PROTECTED]
Sent: Saturday, March 26, 2005
7:37 AM
To: [email protected]
Subject: RE: [flexcoders] Re:
Cyclical References
I am sure that Matt is right about not being able to reliably detect
the circularity (as he says, he'd have fixed it!),
but there _does_
seem to be a consistent feature of broken property
bindings. The
watcher setup code in generated AS in the broken
case seems to contain
references to a special "NoChange"
event, and the number of watchers
approximately doubles when compared to the
generated code for the
non-broken case. I have been able to use
this as a sign that the
compiler bug is probably biting.
I wouldn't take this as a reliable indicator,
since no doubt there is a
bona fide reason that the compiler would emit this
NoChange code in
certain situations. It's smoke, but not
fire.
.
. . . ...j
--- Matt Chotin <[EMAIL PROTECTED]>
wrote:
> I wish I could offer more, obviously it's a
bug we'd rather not have.
> If we
> had a tool that showed the circularity we'd
probably also have the
> fix for
> it :-) As for doing the wrong thing silently,
we did not ship the
> product
> knowing that this case could happen. We
thought the rare case was
> the
> compiler telling you there was a circularity
at all, not binding
> failing to
> generate some of its code.
>
>
>
> I'm hoping that a lot of the work we're doing
on the next version
> will
> address these kinds of issues.
>
>
>
> Matt
>
>
>
> _____
>
> From: Eric Raymond
[mailto:[EMAIL PROTECTED]
> Sent: Friday, March 25, 2005 6:14 PM
> To: [email protected]
> Subject: [flexcoders] Re: Cyclical References
>
>
>
>
> Apparently it's not so rare...!
>
> If you can't fix the problwm, perhaps there's
a tool which will make
> it easier to detect and/or track down these
issues?
>
> Silently doing the wrong thing in some cases
is not a great position
> statement for a tools vendor. Of course I
would care less about this
> if the bug wasn't biting me now, but ....
>
> --- In [email protected],
Matt Chotin <[EMAIL PROTECTED]> wrote:
> >>
> > Thanks to Joe for posting his
workaround. As he says, this is
> supposed to
> > be a rare case which is why we did not
re-write the compiler
>
>
>
>
>
>
>
>
> Yahoo! Groups Sponsor
>
>
>
> ADVERTISEMENT
>
>
<http://us.ard.yahoo.com/SIG=1292r5uv8/M=298184.6191685.7192823.3001176/D=gr
>
oups/S=1705007207:HM/EXP=1111889654/A=2593423/R=0/SIG=11el9gslf/*http:/www.n
> etflix.com/Default?mqso=60190075> click
here
>
>
>
>
<http://us.adserver.yahoo.com/l?M=298184.6191685.7192823.3001176/D=groups/S=
> :HM/A=2593423/rand=896517425>
>
>
>
> _____
>
> Yahoo! Groups Links
>
> * To visit your
group on the web, go to:
> http://groups.yahoo.com/group/flexcoders/
> <http://groups.yahoo.com/group/flexcoders/>
>
> * To
unsubscribe from this group, send an email to:
> [EMAIL PROTECTED]
>
<mailto:[EMAIL PROTECTED]>
>
> * Your use of
Yahoo! Groups is subject to the Yahoo!
> <http://docs.yahoo.com/info/terms/>
Terms of Service.
>
>
| Yahoo! Groups Sponsor |
ADVERTISEMENT
![click here]() | |
![]() |
Yahoo! Groups Links
