add errata README file

The commands in this section will produce an incorrect result if any

image filenames contain space characters. This is because awk treats

spaces as input separators. When md5sum produces its two column

output, it places two space characters between the checksum and the

filename, and awk properly treats them as a separator. But if the

filenames also contain space characters, awk incorrectly treats them

as separators too. As a result, the second field, $2, is only a prefix

of the filename (up to the space), and it is wrongly stored in the

names array and printed in the end.

A quick fix is to replace the term $2 with substr($0, 35, length),

which means "all text from character 35 to the end of the line." The

output of md5sum is predictable because the checksums are always 32

characters long. That means the filename begins at character 35.

A correct, final command:

md5sum *.jpg \

  | awk '{counts[$1]++; names[$1]=names[$1] " " substr($0, 35, length) } \

        END {for (key in counts) print counts[key] " " key ":" names[key] }' \

  | grep -v '^1 ' \

  | sort -nr