Helmer Krämer ([EMAIL PROTECTED]) wrote: > this is actually my bad. Kaffe's BigInteger.longValue() gave wrong > results for negative values. While trying to fix that, I managed > to get some errors in it that I didn't notice because my testcases > succeeded :(((
Kaffe CVS ([EMAIL PROTECTED]) wrote: > * libraries/javalib/java/math/BigInteger.java: > (intValue, longValue): fixed my broken patch I think it's actually working! I've got gmp 3.1.1 installed (the old OpenBSD 3.2 port, since the early-May Kaffe I was using is still dynamically linked against it). I built today's CVS Kaffe using this configuration, and it hasn't crashed or gone comatose yet! PID USERNAME PRI NICE SIZE RES STATE WAIT TIME CPU COMMAND 15010 freenet 59 0 91M 93M run - 2:45 65.38% kaffe-bin 19688 greg 104 20 752K 616K run - 19.9H 29.30% dnetc The unusually high %CPU on the Kaffe process is because I'm hammering on it, loading several high-number-of-images pages plus an FEC splitfile download that's already pretty well cached. Here is it a bit later, closer to "normal": PID USERNAME PRI NICE SIZE RES STATE WAIT TIME CPU COMMAND 19688 greg 103 20 752K 616K run - 20.0H 92.33% dnetc 15010 freenet 2 0 108M 103M run - 5:09 3.91% kaffe-bin At some point I'll probably update back to the newer gmp just to keep things up to date, then rebuild Kaffe again with that. But as far as I'm concerned, it's working! Ship it! ;-) -- Greg Wooledge | "Truth belongs to everybody." [EMAIL PROTECTED] | - The Red Hot Chili Peppers http://wooledge.org/~greg/ |
pgp00000.pgp
Description: PGP signature
