Darwin's uniq

[Akim Demaille]

Tracking down some weird behavior on my Mac, I just noticed messages
such as:

(cd .libs/_fr.lax/libhw.a && ar x /tmp/c++-libtool-swig-python-0.01a/_bulid/swig
/../.libs/libhw.a)
libtool: link: warning: object name conflicts; renaming object files
libtool: link: warning: to ensure that they will not overwrite
usage: uniq [-c | -du] [-f fields] [-s chars] [input [output]]
ar cru .libs/_fr.a  fr_wrap.o .libs/_fr.lax/libhw.a/hw.o

this is 1.5.10.

It turns out there are many uniq -cd, but:

sulaco% uniq -cd
usage: uniq [-c | -du] [-f fields] [-s chars] [input [output]]
sulaco% echo $?
1

The man page pretends -d is supported, but it is not.  I don't know
what -du stands for.  Note that -c and -du are exclusive: you can't
use them both.

sulaco% for i in 1 2 3 3 4 4 5 5 5
for> do
for> echo $i
for> done | uniq -c -du
usage: uniq [-c | -du] [-f fields] [-s chars] [input [output]]
sulaco% for i in 1 2 3 3 4 4 5 5 5
do
echo $i
done | uniq -c   
   1 1
   1 2
   2 3
   2 4
   3 5
sulaco% for i in 1 2 3 3 4 4 5 5 5
do
echo $i
done | uniq -du
1
2
3
4
5