[08:06] <jbailey> doko: Around?
[08:07] <doko> jbailey: yes
[08:08] <jbailey> doko: Mind work-related chatter, or would you rather relax?
[08:08] <doko> just found the cause for the binutils test failures, see my posting to the binutils ML
[08:08] <jbailey> nice!  I'll look in the archives, I'm not subscribed.
[08:09] <doko> the dynamic linker name was substitured by ^A (expanded from \1 on amd64 by dash) :-(
[08:09] <doko> so what's up?
[08:11] <jbailey> Hah, nice.  Glad I was useful on that. =)
[08:12] <jbailey> I'm just building gij on hppa to see what's causing it to puke.  sablevm works, so I know that it's not classpath.
[08:12] <jbailey> And part of me was wondering about the transition to openjdk for arches where it's available.
[08:13] <doko> jbailey: you need gcj to build it, and a port to parisc. bdale told me at debconf that it's unlikely to get the parisc/hpux specific bits from HP
[08:14] <jbailey> from what I read in tromey's blog, it looks like porting the non-hotspot jvm is not that hard.
[08:15] <jbailey> But I had more been thinking that more and more arches are going to get openjdk bits, and I suspect that it should become the default where possible.
[08:16] <jbailey> At that point, I'm not sure there's any point in building the -gcj bits (except where we just inherit that love from Debian)
[08:16] <jbailey> The gij-default decision was from when we thought Sun was joking about a free license. =)
[08:22] <doko> sure, whatever we can get faster, would be better.
[08:57] <jbailey> doko: Hey, is there a trick to make it so that I just build gij with the system compile and not bother building gcc first?
[09:00] <doko> jbailey: you need to build gcc/java/jc1 first, maybe configure with --disable-bootstrap to avoid building the bootstrap compiler? didn't try this with 4.2 ...
[09:01] <jbailey> Cool, thanks.
[09:02] <jbailey> I was just hoping to make ccache useful.
[09:02] <jbailey> =)
[09:14] <doko> jbailey: do you have the gutsy hppa repository somewhere online?