[kaffe] Kaffe JIT on Alpha
robilad at yahoo.com
Tue Apr 22 06:20:02 PDT 2003
--- "Kevin D. Kissell" <kevink at mips.com> wrote:
> Just to remind Kapil, the MIPS JIT, if built from 1.0.7 sources
> with a few patches I posted last year, will build, but fails 5 out
> of 116 regression tests. I managed to figure out and fix one
> failure - and yes, I know I owe the group the necessary patch.
Woof, woof ;)
On a more serious note: the switch to compiling the class library has exposed
some bugs in the jitter on many platforms other than i386 that were not that
obvious before. I'm getting some weird exceptions on ARM (with jit) as well,
during class library compilation, while interpreter works fine.
The weird thing is that jit works fine for compiling the class library on i386,
Class library build failures on IRIX seem to be unrelated to the jitter (as I
get the same error using the interpreter): kjc isn't able to parse
java.lang.Double.MIN_VALUE value properly. I'm inclined to blame it on their C
library's strtod and use a strtod replacement function on that platform. Of
course, once that's fixed, the cross-platform jitter-bugs could kick in as well
I'm wondering if compiling the jitter with some extra gcc warnings could give a
hint at what's fishy, or using splint ...
> Note that Dalibor has been able to get the MIPS JIT under IRIX
> from CVS sources past the @essential files build failure by using
> Jikes instead of kaffe to build the class libraries. If you do that,
> you should be able to build with --enable-debug, set KAFFE_DEBUG
> to "gdb" (or whatever), install the crippled kaffe, and invoke it with
> --vmdebug JIT,MORE_JIT under the debugger, piping the output
> to "tee" to get a trace fule, to get some idea of what's going on...
That's very good advice. I'll put it into FAQ.debugging.
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo
More information about the kaffe