But in "SORT" for linux redhat

[bhiksha+]

Consider the following file:

>cat tmpsortfile

ABALONE        AE B AH L OW N IY
AB        AE B
AB        EY B IY
ABEYANCE        AH B EY AH N S



>sort tmpsortfile

What I *should* get


AB        AE B
AB        EY B IY
ABALONE        AE B AH L OW N IY
ABEYANCE        AH B EY AH N S


What I *do* get

AB        AE B
ABALONE        AE B AH L OW N IY
ABEYANCE        AH B EY AH N S
AB        EY B IY


Clearly a bug. No way "AB " is both lesser than and greater than ABALONE.
It seems to be sorting by the second field, for some reson..