[pavel.tsekov]

No problem - will do. Once I am ready I'll let you know.

[pavel.tsekov]

I've changed the code to utilize java.beans.XMLEncoder since it seemed appropriate. I produces something like that:



org.jboss.aop.pointcut.ast.ParseException


6640507307072022813



Now, do we really want a single XML file for all previous releases or you can live with one file for a release ? I think that creating a single file for all releases will complicate the tool without any real advantages. Well, there are some advantages - the fact that we need to validate against j2se 1.4 and 1.5 will double the number of produced files so if we have a single file repository we'll only have two files. Please, advise.

[pavel.tsekov]

Sorry, an entry in the database looks like this:



org.jboss.aop.pointcut.ast.ParseException


6640507307072022813

[pavel.tsekov]

Ouch. The board strips the XML tags :(

[brian.stansberry]

Use the 'code' button. :)

[pavel.tsekov]

 <object class="org.jboss.tools.ClassVersionInfo">
 <void property="name">
 <string>org.jboss.aop.pointcut.ast.ParseException</string>
 </void>
 <void property="serialVersion">
 <long>6640507307072022813</long>
 </void>
 </object>

Thanks, for the hint!

[manik]

If you need to create multiple XML files, then you should do it in a subdirectory of /etc - perhaps /etc/compatibility-db or something.

[pavel.tsekov]

I've just committed the changes to the SerialVersionUID tool to generate XML files.

[pavel.tsekov]

The following SVUID database files were added to the CVS repository:

U etc/svuid-db/svuid-14.1.3.0.GA.xml
U etc/svuid-db/svuid-14.1.4.0.GA.xml
U etc/svuid-db/svuid-50.1.3.0.GA.xml
U etc/svuid-db/svuid-50.1.4.0.GA.xml

[pavel.tsekov]

Forgot to mention it in the previous post - the database was based on the jars found in jboss-cache-dist-1.4.0.GA.zip and jboss-cache-dist-1.3.0.GA.zip located on sf.net.

[pavel.tsekov]

I've just commited the test cases for JBossCache 1.3.0 GA and 1.4.0 GA.
Those tests can by executed by using the "compat-tests" target. Both tests fail due to issues with jgroups and some issues with jbosscache classes - see below. Please, advise whether I should mark those failures as expected.

Failed classes when testing against 1.4.0 GA:

org.jboss.cache.marshall.JBCMethodCall
org.jboss.serial.objectmetamodel.DataContainer
org.jgroups.Message
org.jgroups.blocks.DistributedHashtable
org.jgroups.blocks.DistributedLockManager$LockDecree
org.jgroups.blocks.NotificationBus$Info
org.jgroups.blocks.ReplicatedHashtable
org.jgroups.blocks.ReplicatedTree$Node
org.jgroups.blocks.ReplicatedTree$Request
org.jgroups.blocks.VotingAdapter$VoteResult
org.jgroups.demos.Draw$DrawPanel
org.jgroups.demos.MyCanvas
org.jgroups.demos.QuoteClient
org.jgroups.demos.TotalOrder
org.jgroups.demos.wb.SendDialog
org.jgroups.protocols.ENCRYPT$EncryptHeader
org.jgroups.protocols.FD_PROB$FdHeader
org.jgroups.protocols.FD_SIMPLE$FdHeader
org.jgroups.protocols.JMS$JMSAddress
org.jgroups.protocols.PerfEntry
org.jgroups.protocols.SMACK$SmackHeader
org.jgroups.protocols.TOTAL$Header
org.jgroups.protocols.TunnelHeader
org.jgroups.protocols.VIEW_SYNC$ViewSyncHeader
org.jgroups.protocols.pbcast.JoinRsp
org.jgroups.protocols.pbcast.STATE_TRANSFER$StateHeader
org.jgroups.util.List
org.jgroups.util.TimedWriter$Timeout

Failed classes when testing against 1.3.0 GA:

org.jboss.cache.OptimisticTreeNode
org.jboss.cache.optimistic.DefaultDataVersion
org.jboss.cache.optimistic.FqnComparator
org.jgroups.ChannelException
org.jgroups.MergeView
org.jgroups.Message
org.jgroups.View
org.jgroups.ViewId
org.jgroups.blocks.DistributedHashtable
org.jgroups.blocks.DistributedLockManager$LockDecree
org.jgroups.blocks.NotificationBus$Info
org.jgroups.blocks.ReplicatedHashtable
org.jgroups.blocks.ReplicatedTree$Node
org.jgroups.blocks.ReplicatedTree$Request
org.jgroups.blocks.RequestCorrelator$Header
org.jgroups.blocks.VotingAdapter$VoteResult
org.jgroups.debug.Debugger
org.jgroups.demos.Draw$DrawPanel
org.jgroups.demos.MyCanvas
org.jgroups.demos.QuoteClient
org.jgroups.demos.Topology
org.jgroups.demos.TotalOrder
org.jgroups.demos.applets.DrawApplet
org.jgroups.demos.wb.GraphPanel
org.jgroups.demos.wb.SendDialog
org.jgroups.demos.wb.Whiteboard
org.jgroups.protocols.ENCRYPT$EncryptHeader
org.jgroups.protocols.ExampleHeader
org.jgroups.protocols.FC$FcHeader
org.jgroups.protocols.FD$FdHeader
org.jgroups.protocols.FD_PROB$FdHeader
org.jgroups.protocols.FD_SIMPLE$FdHeader
org.jgroups.protocols.FD_SOCK$FdHeader
org.jgroups.protocols.FLOW_CONTROL$FCInfo
org.jgroups.protocols.HTOTAL$HTotalHeader
org.jgroups.protocols.JMS$JMSAddress
org.jgroups.protocols.PerfHeader
org.jgroups.protocols.SMACK$SmackHeader
org.jgroups.protocols.StateTransferRequest
org.jgroups.protocols.TOTAL$Header
org.jgroups.protocols.TOTAL_OLD$TotalHeader
org.jgroups.protocols.TransportedVectorTime
org.jgroups.protocols.TunnelHeader
org.jgroups.protocols.UNICAST$UnicastHeader
org.jgroups.protocols.WanPipeAddress
org.jgroups.protocols.pbcast.Digest
org.jgroups.protocols.pbcast.GMS$GmsHeader
org.jgroups.protocols.pbcast.JoinRsp
org.jgroups.protocols.pbcast.NakAckHeader
org.jgroups.protocols.pbcast.STABLE$StableHeader
org.jgroups.protocols.pbcast.STATE_TRANSFER$StateHeader
org.jgroups.stack.IpAddress
org.jgroups.util.List
org.jgroups.util.TimedWriter$Timeout

[pavel.tsekov]

I've just merged org.jboss.tools.SerialVersionUID and org.jboss.cache.compat.SerialVersionUIDUnitTestCase into HEAD.
From an earlier comment on this discussion I understand that testing 2.x.x
against 1.x.x doesn't make sense except in some rare cases. So I haven't merged the xml database file - please, let me know if you want me to add them to HEAD.

[manik]

Are the failures above on HEAD? There is no point comparing HEAD (2.x.x) against 1.x.x classes as you mentioned.

[pavel.tsekov]

No - the failures above are produced by comparing Branch_JBossCache_1_4_0 to 1.4.0 GA and 1.3.0 GA. I, do understand that 2.x.x is incompatible with 1.x.x as already discussed earlier on this thread.

[belaban]

Most of the JGroups exceptions are meaningless, because JGroups doesn't use Java serialization anyways. I know that some JGroups exceptions *might* get serialized when used with method calls, but I don't see any of them in the output.
Manik ?