In brief, uniq seems to treat lines containing only spaces differently when given the -f 1 option (compared to when given -f 0 or no -f option). My question is: Is this behavior intentional (or is it a bug in the implementation or docs)? I find it difficult to reconcile with my understanding of the docs.
In more detail, consider a file in.txt with the following contents (which can be reconstructed based on the descriptive line if mangled by mailers): Next 5 lines have, resp., 1, 2, 1, 2, and 1, blanks: Last line Given this input, the output of 'uniq -u -f 1 in.txt' is different from that of 'uniq -u in.txt' and 'uniq -u -f 0 in.txt'. (With -f 1, the blanks-only lines are all removed, but not so with the others.) I tested the above originally on the uniq from coreutils 8.32 but later also on 9.1.42 (built from the git sources I just pulled a short while ago) and both versions exhibit the same behavior. Regards, -chaw
