Hi, On Fri, 2008-08-01 at 11:44 -0700, Kelly O'Hair wrote: > Best guess is that some kind of change to javac is triggering > this failure, but I'm not convinced anything is 'wrong' with javac, > just a change in the class files created implicitly. Stay tuned... > > In the corba repository I did find this problem: > > diff -r 80a0f46a6203 make/org/omg/CORBA/Makefile > --- a/make/org/omg/CORBA/Makefile Thu Jul 17 11:28:32 2008 -0700 > +++ b/make/org/omg/CORBA/Makefile Fri Aug 01 11:30:41 2008 -0700 > @@ -48,7 +48,7 @@ endif > > FILES_java = \ > $(ORG_OMG_SENDINGCONTEXT_java) \ > - $(ORG_OMG_CORBA_2_3_java) \ > + $(org_omg_CORBA_2_3_java) \ > $(ORG_OMG_CORBA_java) > > # > > I'm testing it now but I still don't understand why it fails sometimes > and not others. The 2_3 sources are never explicitly listed to be > compiled so how they got compiled in the past is a puzzle.
In icedtea we have the following patches/icedtea-corba.patch: --- openjdk.old/corba/make/org/omg/CORBA/Makefile 2007-10-12 03:49:14.000000000 -0400 +++ openjdk/corba/make/org/omg/CORBA/Makefile 2007-11-06 13:57:24.000000000 -0500 @@ -49,7 +49,10 @@ FILES_java = \ $(ORG_OMG_SENDINGCONTEXT_java) \ $(ORG_OMG_CORBA_2_3_java) \ - $(ORG_OMG_CORBA_java) + $(ORG_OMG_CORBA_java) \ + $(org_omg_SENDINGCONTEXT_java) \ + $(org_omg_CORBA_2_3_java) \ + $(org_omg_CORBA_java) # It comes from a similar thread last year about corba failing to compile in some (bootstrap) situations with alternative compilers gcj/ecj in our case: http://thread.gmane.org/gmane.comp.java.openjdk.distro-packaging.devel/414 Cheers, Mark