This was fun :-) 1.6M i386, compiled current yesterday, kernel paniced whilst
in X (no trace, sorry), destdir looked complete, compiled kernel, make one
run with ./build.sh -u distribution, installed into /, booted and:

jarkko@sd161:/crap/sbin > file * |grep data
clri:            DBase 3 data file with memo(s)
dump_lfs:        data
fsck_lfs:        data
fsck_msdos:      data
mount_ados:      data
mount_ext2fs:    data
mount_fdesc:     data
mount_ffs:       data
mount_filecore:  data
mount_kernfs:    data
mount_lfs:       data
mount_msdos:     data
mount_nfs:       data
mount_ntfs:      data
mount_null:      data
mount_overlay:   data
mount_portal:    data
mount_procfs:    data
mount_ufs:       data
mount_umap:      data
mount_umap:      data
mount_union:     data
newfs_lfs:       data
newfs_msdos:     data
ping6:           setuid data
rdump_lfs:       data
rtsol:           data
setkey:          data
tunefs:          data
jarkko@sd161:/crap/sbin > file mount_mfs
mount_mfs: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for
NetBSD, dynamically linked (uses shared libs), stripped

/rescue/mount_ffs told me that the kernel process table was full
(in single user) and instructed me to fiddle with NPROCS. This was
with GENERIC.

Went back to boot floppies and got a working system. I'll try again
with a fresh build today.

It was fun :-) (Note: this is not a bug report and I'm not complaining.
Just goes to show how badly one can hurt one's leg when you're not
carefull)
-- 
jht