This is the mail archive of the java-discuss@sourceware.cygnus.com mailing list for the GCJ project. See the GCJ home page for more information.
> From the assembler output of the Java compiler, it appears that > gcj treats every method as if it has a variable length arglist. I don't believe that is the case. On the other hand, we've only tried it on machines where the varargs ABI is the same as the non-varargs one. > Note the following output from an arm-gcj cross assembler: I don't know the ARM ABI or assembler language, so I can't help. Are you compiling with -O? > For many architectures with alot of registers like sparc or a29k, this > would incur a big overhead... I don't beleive the overhead you fear exists, at least on the sparc. (Of course, it has register windows.) --Per Bothner Cygnus Solutions bothner@cygnus.com http://www.cygnus.com/~bothner