Re: Build failure with CVS build

[Roland McGrath]

The direct thing to do is use gdb on ext2fs.static when booting a subhurd.
Then you can step through the dir_lookup stub that's failing and see what
the reply message it got was.  In this instance, the filesystem is talking
to itself (one thread making an RPC to another), so you can also step
through the server side in the same gdb.

If you haven't tried this already, you could also try using the new
binaries in a non-boot situation.  If there are any problems there,
that will be easier to debug than the boot case.

The kernel panic at the end is especially exciting.  I suspect that will go
away when you fix the underlying hurd problems, but who knows.