Thanks for the suggestions! I've checked the javassist versions - I
initially had two different versions of 3.6. I think you are right that this
is more likely a javassist issue than a tapestry issue. I have a workaround
on our development machines (both linux and windows), but the resulting war
still broke on our main server, which made me nervous! If I get a better
understanding of the cause I'll post here.
Thanks again,
Adriaan
On Feb 11, 2008 8:41 PM, Howard Lewis Ship <[EMAIL PROTECTED]> wrote:
> I don't know about the cause, though you if you are deploying inside
> JBoss, you may be seeing a conflict on the version of Javassist.
>
> The workarounds for these cases are to encapsulate the affected
> instance variable inside its own (private) accessor methods. That
> simplifies the method that Tapestry instruments. I had previously
> done this a few times before a later release of Javassist fixed the
> problem ... or so I thought.
>
> You may also do ok by getting the value from the parameter into a
> local variable outside the try ... finally block. try ... finally
> support is a late addition to Javassist and is perhaps buggy.
>
> On Feb 11, 2008 9:58 AM, Adriaan Joubert <[EMAIL PROTECTED]> wrote:
> > Hi Howard,
> >
> > More information. The actual error is
> >
> > "bad frame_type in StackMapTable"
> >
> > and is thrown from javassist.bytecode.StackMapTable.stackMapFrames
> (int,int).
> > Not sure whether this helps, but it looks as if the stack is messed up.
> >
> > There is a try {...} finally{..} in the code, and if I remove this it
> runs
> > through. I've spent several hours trying to produce a small test case
> (the
> > actual site includes a 500k line internal library), and, even though it
> > looks very similar it runs through fine. I'll continue trying to produce
> a
> > test case, but in the meantime is there anything I can do to identify
> the
> > cause?
> >
> > Thanks,
> >
> > Adriaan
> >
>
>
>
> --
> Howard M. Lewis Ship
>
> Creator Apache Tapestry and Apache HiveMind
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
