RE: Wanted: SerialVersionUID experts ;)

[Jeroen Frijters]

Stuart Ballard wrote:
> Tom pointed out a little problem with Japitools' serialVersionUID
> calculation.

It's Japi bug. Classpath gets it right.

I don't have time to look into it right now (alright, I admit, I'm too
lazy), but I'm pretty sure that we're (read: I am) using the wrong class
modifiers. Can you try this patch:

Index: src/net/wuffies/japi/ClassFile.java
===================================================================
RCS file:
/cvsroot/japitools/japitools/src/net/wuffies/japi/ClassFile.java,v
retrieving revision 1.12
diff -u -r1.12 ClassFile.java
--- src/net/wuffies/japi/ClassFile.java 21 Sep 2005 01:55:11 -0000
1.12
+++ src/net/wuffies/japi/ClassFile.java 23 Sep 2005 22:03:33 -0000
@@ -1194,7 +1194,7 @@
 
            data_out.writeUTF(getName());
 
-           int modifiers = raw_access_flags;
+           int modifiers = access_flags;
            // just look at interesting bits
            modifiers = modifiers & (Modifier.ABSTRACT  | Modifier.FINAL
|
                Modifier.INTERFACE | Modifier.PUBLIC);

Regards,
Jeroen