Jeff Barcalow wrote:
Paul Webster wrote:
Jeff Barcalow wrote:
The stack trace is:
java.lang.VerifyError: (class: org/vssplugin/ui/view/CheckOutView,
method: <init> signature: ()V) Incompatible argument to function
at java.lang.Class.getDeclaredConstructors0(Native Method)
at java.lang.Class.privateGetDeclaredConstructors(Unknown Source)
at java.lang.Class.getConstructor0(Unknown Source)
at java.lang.Class.newInstance0(Unknown Source)
at java.lang.Class.newInstance(Unknown Source)
at
org.eclipse.core.internal.registry.osgi.RegistryStrategyOSGI.createExecutableExtension(RegistryStrategyOSGI.java:170)
That is an odd error from the vssplugin ...
RegistryStrategyOSGI.createExecutableExtension(*) always create the
no-arg constructor, and I'm sure CheckOutView has always provided one.
Is it complaining about the binary compatibility of the vss
CheckOutView? If CheckOutView didn't have a no-arg constructor I'd
expect that it was some kind of no method found error.
After M6 only PMC approved API changes were allowed. That's very
small number. Which API do you think has changed?
PW
I am not familiar with the vssplugin code, but I just grabbed
part of it from sourceforge. It compiles with 3.3M6, but with
3.3.M7 it has 32 compile errors. The VerifyError is because
CheckOutView has data members that inherit from eclipse objects,
but no longer have a consistent class hierarchy.
Most look errors like they are related to vssplugin classes that
inherited from
org.eclipse.team.internal.ui.actions.TeamAction
Maybe the vssplugin should not have inherited from a class in an
"internal" package.
It looks like in M6 TeamAction extends ActionDelegate while M7
extends AbstractHandler.
- 13 of the errors are because TeamAction added an abstract method
executed(IAction).
- 6 of the errors are because of a thrown exception change with
isEnabled()
- 4 of the errors are "cannot reduce the visibility of inherited
method"
AbstractHandler.isEnabled()
- 9 of the errors are because ActionDelegate objects are now
IActionDelegate.
It is relatively easy to fix the compile errors, but that does not mean
the
plug-in is working correctly.
-Jeff
I've posted a "potential" source code patch at:
http://sourceforge.net/tracker/?group_id=54205&atid=473013
I could not post the jar because of an attachment size limit.
Anyone interested in using this would have to get the rest
of the project from the CVS and then apply the changes.
I'm still testing the modified JAR myself.
-Jeff