[ Apologies if this doesn't thread correctly - I'm not subscribed, so can't
just hit reply. Please CC any replies. ]
Regarding this backtrace:
http://lists.gnu.org/archive/html/bug-coreutils/2006-09/msg00017.html
FWIW, my backup script was failing, so I looked into this today:
From around line 550 of dd.c:
fprintf (stderr,
ngettext ("1 byte (1 B) copied",
"%"PRIuMAX" bytes (%s) copied",
MIN (w_bytes, ULONG_MAX)),
w_bytes,
human_readable (w_bytes, hbuf, human_opts, 1, 1));
The problem is that, if the locale matching $LANG has been built
(for me that's en_GB.UTF-8), then ngettext appears to return something of
the form:
%<PRIuMAX> bytes (%s) copied
(To be honest, although I see what's causing the crash, I'm not sure
why the string looks exactly like that)
This string is then passed as the format to fprintf - where %<PRIuMAX>
should be something like %llu.
I'm working around it by using "unset LANG" before the backup
script is run.
Cheers,
Martin