This is the mail archive of the java@gcc.gnu.org mailing list for the Java project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: cygwin issues


>>>>> "Tony" == Tony Kimball <alk@pobox.com> writes:

Tony>   1) There are duplicate definitions of java::lang::Class::Class() in 
Tony>   natClass.o (inlined public constructor from Class.h) and Class.o
Tony>   (private constructor).  I removed the inline public definition.

If that works then it is fine.  Could you send a patch and ChangeLog
entry?

Tony>   2) The #pragma weak for JNI_OnLoad was ignored, so I #if 0'd out the
Tony>   calls. 

We'll need something better here.  I like Jeff's idea of a check for
weak symbol support.  If that is too hard to write a test for we could
make it platform-dependent by adding a new variable to configure.host.

Tony> - I applied a patch from the mailing list to fix vtable symbols:
Tony>   http://gcc.gnu.org/ml/java/2001-04/msg00349.html

This patch didn't seem to get reviewed.  Jeff, is it still needed?  I
didn't see it in the gcc3 ChangeLog.

Tony> - timezone definitions conflict.  The cygwin sys/time.h and time.h 
Tony>   are a bit messy.  I just #if 0'd the case of char *timezone(),
Tony>   and eliminated the '#define timezone _timezone' (to prevent
Tony>   conflicts with "struct timezone", and did s/timezone/_timezone/
Tony>   in java/lang/natSystem.cc and java/util/natGregorianCalendar.cc.

This won't be as much of a problem in the future, since we ditched
natGregorianCalendar.cc.  If you could come up with a clean patch for
natSystem.cc, that would be great.  Use the trunk for this though.

Tony> - alloca.  ok, this should be obvious to me, but its not.  There's
Tony>   seemingly no alloca.h for cygwin, so I just changed references to 
Tony>   __builtin_alloca. 

Could you make a separate patch to change alloca to __builtin_alloca?
I think we could just check that in.

Tony> - More JNI crud.  I did a bunch of stuff like this:
Tony>   +#ifdef ENABLE_JNI
Tony>      _Jv_JNI_Init ();
Tony>   +#endif

How extensive are the no-JNI changes?  I think this is actually
something we'd like.  I picture a `--disable-jni' configure flag.

Tom


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]