Yes, it looks like `anoncvs' is hosed, since I saw similar problems
pulling from a mirror that rsync's against anoncvs.

> I don't have the first line of that happening, because screen can't scroll back
> far enough, so I just pasted what happened when cvs died.

Here's where it appeared to start, as displayed by my `cvsup' session:


 Edit syssrc/sys/arch/luna68k/include/db_machdep.h,v
  Add tag kqueue-nbase -> 1.2
 Edit syssrc/sys/arch/luna68k/include/disklabel.h,v
  Add tag kqueue-nbase -> 1.2
Server warning: RCS file error in "/usr/local/etc/cvsup/prefixes/netbsd/syssrc/s
ys/arch/luna68k/include/elf_machdep.h,v": 1: "head" expected
Server warning: RCS file error in "/usr/local/etc/cvsup/prefixes/netbsd/syssrc/s
ys/arch/luna68k/include/endian.h,v": 1: "head" expected
     [much more snipped]

Here's where the bad files seemed to end:
[snippage]
Server warning: RCS file error in "/usr/local/etc/cvsup/prefixes/netbsd/syssrc/s
ys/arch/m68k/fpsp/x_unimp.sa,v": 1: "head" expected
Server warning: RCS file error in "/usr/local/etc/cvsup/prefixes/netbsd/syssrc/s
ys/arch/m68k/fpsp/x_unsupp.sa,v": 1: "head" expected
Updating collection netbsd-xsrc/cvs


The affected directories seem to be the arch/luna68k and arch/m68k,
containing a bunch of empty files on my (norge) chosen mirror.


So, you are not alone, and thanks for confirming for me that the
problems aren't with the mirror I leech off of, but rather upstream
on the anoncvs machine from which it pulls.


I have another corrupt file I quoted from earlier, which apparently
is corrupt on the site I use, perhaps also on anoncvs, that b0rked my
build yesterday, and which hasn't been updated today to be correct.
Would you care to take a look at your basesrc/usr.sbin/makefs/ffs/mkfs.c
file and tell me if you find a line some 180 lines into it like
statectors/track including spares */
within it?  That will tell me if the file is equally corrupt on anoncvs,
or whether it's getting corrupted in the mirroring process, or somewhere
else...


thanks
barry bouwsma