duff

Release Notes: This release improves performance for both small and large file sets, prefixes warning and error messages with the program name, changes sampling to use the first n bytes of a file, changes the default sampling limit to zero (sample all files), fixes the man page examples, and removes the length limit on path names read from stdin.

Release Notes: This release fixes the crash bug which occurred when the cluster header in thorough mode used a digest and adds a new default cluster header for thorough mode.

Release Notes: This release adds selection of the entire SHA family of hashes, reading and writing null-terminated file names, i18n via gettext, a translation to Swedish, support for large files on more platforms, and a number of bugfixes.

Release Notes: This release adds file name reading from stdin, primitive directory loop detection, compatibility with more systems, physical mode, an argument symbolic link flag, reporting of access-protected empty files, and an example script for hard-linking duplicate files. The man page has been updated and improved. Several bugs have been fixed. The collection of symbolic links to files has been removed.

Release Notes: A typo in the help for the -t option was fixed. A missing "t" was added to the getopt string. A bug in ordering of comparison phases was fixed. The -z option for ignoring empty files was added. The primary driver logic was moved into a separate file.