dev-libs/libffi is now optional via the ffi use flag so you should be able to add your keywords back if jamvm works on ppc64 without libffi (meaning that it has it's own code for native calling on ppc64).
unfortunately I'm not able to comile gnu-classpath-0.93 (DEPEND of jamvm-1.4.5)
(In reply to comment #1) > unfortunately I'm not able to comile gnu-classpath-0.93 (DEPEND of jamvm-1.4.5) > As a stop gap measure you can mask the nsplugin use flag. The plugin is unsafe any way but I will try looking into this.
(In reply to comment #1) > unfortunately I'm not able to comile gnu-classpath-0.93 (DEPEND of jamvm-1.4.5) > Should be fixed now.
hmm.. I'm not sure what this program does ;-) I've done this to test this application: $ cat helloworld.java class Hello { public static void main(String args[]) { System.out.println("Hello World!"); } } $ javac helloworld.java $ jamvm Hello.class the result is that my CPU is being used 100%, but nothing else happens.. is this good or bad? what does this "virtual machine" do? have I used it completely wrong?
(In reply to comment #4) > > the result is that my CPU is being used 100%, but nothing else happens.. is > this good or bad? what does this "virtual machine" do? have I used it > completely wrong? > Ok some Java basics: javac is used with .java like and the vm with just the class name: betelgeuse@pena ~/test/java $ javac Hello.java betelgeuse@pena ~/test/java $ jamvm Hello Hello World! /tmp But it certainly should not just go into using 100% CPU.
(In reply to comment #5) > } > $ javac helloworld.java > $ jamvm Hello.class > > What javac did you use to compile it (eselect java-vm list)?
(In reply to comment #5) > betelgeuse@pena ~/test/java $ jamvm Hello ok. used that command and I'm getting the same result (100% CPU use, no output) $ eselect java-vm list Available Java Virtual Machines: [1] ibm-jdk-bin-1.4 [2] ibm-jdk-bin-1.5 system-vm [3] jamvm
(In reply to comment #7) > (In reply to comment #5) > > betelgeuse@pena ~/test/java $ jamvm Hello > > ok. used that command and I'm getting the same result (100% CPU use, no output) > Did you try with ffi use flag on and off?
Good point... I have only tried with USE="-ffi". Will try with that useflag enabled.
(In reply to comment #7) > (In reply to comment #5) > > betelgeuse@pena ~/test/java $ jamvm Hello > $ eselect java-vm list > Available Java Virtual Machines: > [1] ibm-jdk-bin-1.4 > [2] ibm-jdk-bin-1.5 system-vm > [3] jamvm This means your javac produced 1.5 bytecode, while jamvm can only handle 1.4 (I think)?
(In reply to comment #10) > > This means your javac produced 1.5 bytecode, while jamvm can only handle 1.4 (I > think)? > False: betelgeuse@pena ~/test/java $ javac Hello.java betelgeuse@pena ~/test/java $ javac -version javac 1.6.0 betelgeuse@pena ~/test/java $ jamvm Hello Hello World! /tmp betelgeuse@pena ~/test/java $
Hello ppc64 people. Time to test jamvm-1.5.0.
same as comment #4
still same; thoughts?
Can we try this newer version?
Make it 1.5.4-r1 Please check both with USE=libffi and USE="-libffi"
The other option is to drop the ppc64 keyword - no reverse dependencies - I doubt the current stable (1.4.2) still works - no user seemed to care so far So I ask whether you agree to - drop current stable and keep this open - drop and close - keep open
(In reply to comment #17) > The other option is to drop the ppc64 keyword > - no reverse dependencies > - I doubt the current stable (1.4.2) still works > - no user seemed to care so far > > So I ask whether you agree to > - drop current stable and keep this open > - drop and close > - keep open > + 05 Jul 2010; Samuli Suominen <ssuominen@gentoo.org> jamvm-1.4.2.ebuild: + Drop ppc64 wrt #168736, Comment #17
Created attachment 904127 [details, diff] PPC64 locking primitive fix This patch fixes JamVM on PPC64. Just in case anyone else is trying to bootstrap Java on a PPC64 in the year 2024.
