ia64 "hello world" success; crfsd receives SEGV under heavier load

[Zach Brown]

> 
> Unfortunately, it looks like there *might* be some data
> corruption (haven't gotten around to verifying it yet).

Yeah, that isn't entirely unlikely.  There were a few places in the
client item manipulation that weren't as careful as they should have
been.  Things like not dealing with directory entries that hash to the
same item, that kind of thing.

  http://oss.oracle.com/mercurial/zab/crfs/rev/57067590d58d

I just landed a big change to the way that the client maps file system
operations to multiple item changes.  It should help things quite a bit,
I hope.  Over here kernel builds (make, make clean, make, etc) have
gotten a lot more solid.

> I say *might* based on this experiment:

I'd use find and md5sum to dig into this.

  tests/complex/untar-md5-around-remount.sh

That script in the source tree does something like this.

> I'm also seeing these messages from the log file:
> 
> Apr 03,2008|14:39:49|29576|send_final_neg_errno: assertion 'wire' failed: errno 4294967274

There's a few things that could be.  The debug log would tell us more.

- z