[DAFFODIL-1917] Support for new character classes

• Type: New Feature
• Status: Closed
• Priority: Minor
• Resolution: Not A Problem
• Component/s: Back End, Front End, Middle "End"
• Assignee: Unassigned
• Reporter: Steve Lawrence
• Created: 13/Mar/18 14:43
• Updated: 21/Jan/21 21:27
• Resolved: 21/Jan/21 21:27

Description

Daffodil v2.0 wishlist includes new character class entities LSP, LSP* and LSP+, which are the same as WSP except they do not match line endings. This is useful as currently it is impossible to match an empty line followed by a newline via something like %WSP*;%NL; due to the greediness of WSP* and the inability for our delimiter scanner to backtrack of forward lookahead.

Related, it might be useful to also add NL+ and NL* to match an unbounded number of newlines.

Additionally, it might be a good idea to cause an SDE if a character class that contains * or
+ is immediately followed by a character inside that character class. The reason being that such sequences will always fail without lookahead or backtracking, which probalby isn't worth implementing.

For security reasons, we should also implement a tunable to limit the number or repetitions of these character entities.