GitLab

The full (unihan-included) database can be loaded by specifying the -H
option.

- No longer depends on Perl's Unicode::UCD for information: instead a
  new script builds a unicode.data file from the UCD XML file.
- Character aliases and abbreviations are now supported for searching
  and are displayed
- Searching now accepts multiple search terms (or regexes), all of which
  must match (any of which if using new --or argument)
- Control character abbreviations now come out of the unicode data
  instead of being hard coded.
- Added scripts to download and extract the data from the latest unicode
  specification.
- Various character attributes are now reported
- The new data file (unicode.data.gz) built by extract.pl is designed to
  work compressed, reducing the storage space required.
- Added the unicode version in which a character was added
- Small cosmetic tweaks
- Rewrote --help output
- Fixed bugs:
  - UTF-16 display was flat out wrong for 4 byte UTF-16 characters
  - --details didn't do anything when searching
  - searching never found control characters

The previous commit removed binary literals, but didn't properly update
the code for literal character mode.

"0b1" is a valid hex value, and was interpreted as such, so the binary
input value regular expression subgroup never matched.

Perl 5.14 no longer installs UnicodeData.txt, so include a copy (from Unicode 6.0.0) alongside the script and, using FindBin, load it instead of Perl's internal version.

Added -d for --details
Added -l and --list as aliases for --nodetails
Made -X (and --HEX) actually work (it was mentioned in --help, but not handled in the code)
Added the long option names for format aliases to --help output

- Added decomposition, when it exists, including looking up the decomposed character names, and printing out
	the decomposed version (which should be visually indistinguishable, but can be copied)
- Added unicode version to --help output
- Allowed a single character to be prefixed with _ so that _e can be used to show U+0065 instead of U+000E
- Changed the regexp matching a single character to \X instead of . so that it matches a single grapheme
	cluster instead of just a single codepoint, thus allow decomposed character encodings to work properly
- Print out an error if you try to do x-y where one of x or y is a multi-codepoint composition
	(i.e. a grapheme cluster), since a range makes no sense in that case.

[[:xdigit:]] doesn't work properly under 'i' in a regex (perlbug #64838); changed to 0-9a-fA-F which does work.

Added support for spaces (i.e. shell-escaped) in addition to underscores

Added the shortened special characters again to make the list mode show up nicely; changed 0-255 default to 0-ff (since arguments are now hex)