Support for patterns with regex-like character classes for Latin-1

[shreepads]

The current pattern support for character classes is limited to characters from the basic ASCII characters, A-Z, a-z, 0-9 and some punctuation chars.

Apart from the known issue re the chars {, }, [ and ](see Issue #12) there is also the issue that up to PDF versions 1.6 and 1.7, passwords are allowed with the full Latin-1 character set.

See:

• Max password length and encoding for PDF versions: http://www.pdflib.com/knowledge-base/pdf-security/encryption/
• Latin-1 character set: https://en.wikipedia.org/wiki/ISO/IEC_8859-1

In order to handle this better, add the following character classes

• :anylat: - :any: plus all remaining Latin-1 chars
• :alnumlat: - :alnum: plus all remaining Latin-1 alphabet e.g. Å and å
• :alphalat: - :alpha: plus all remaining Latin-1 alphabet e.g. Å and å
• :digitlat: - the same as :digit:, doesn't include fractions like ¼ from Latin-1
• :lowerlat: - :lower: plus all remaining lowercase Latin-1 alphabet e.g. å but not Å
• :upperlat: - :upper: plus all remaining uppercase Latin-1 alphabet e.g. Å but not å
• :punctlat: - :punct: plus all remaining Latin-1 non alphanumeric chars e.g.¥, », ¿, ½

Also see Issue re support for longer Unicode character patterns.