Regular Expression Classes

SharpMUSH uses .NET character classes and Unicode categories. POSIX bracket classes such as [[:digit:]], [[:alpha:]], and [[:^word:]] are not supported. Convert them when porting PennMUSH patterns; .NET can accept some of this text as a different pattern without reporting an error.

The following are replacement choices, not exact equivalences for every PCRE locale or option:

POSIX name	ASCII pattern	Unicode pattern
`digit`	`[0-9]`	`\d` or `\p{Nd}`
`alpha`	`[A-Za-z]`	`\p{L}`
`alnum`	`[A-Za-z0-9]`	`[\p{L}\p{Nd}]`
`lower`	`[a-z]`	`\p{Ll}`
`upper`	`[A-Z]`	`\p{Lu}`
`word`	`[A-Za-z0-9_]`	`\w`
`space`	`[\t\n\v\f\r ]`	`\s`
`ascii`	`[\x00-\x7F]`	Same ASCII range.
`cntrl`	`[\x00-\x1F\x7F]`	`\p{Cc}`
`graph`	`[\x21-\x7E]`	Choose the required categories explicitly.
`print`	`[\x20-\x7E]`	Choose the required categories explicitly.
`punct`	`[\x21-\x2F\x3A-\x40\x5B-\x60\x7B-\x7E]`	`\p{P}` (Unicode punctuation excludes many symbols).
`xdigit`	`[A-Fa-f0-9]`	Same range for ASCII hexadecimal input.

Negate a whole class with a leading ^, for example [^0-9]. Use \D, \S, \W, or \P{L} for the complements of their corresponding Unicode classes. To combine letters and digits, write [\p{L}\p{Nd}], not [[:alnum:]].

Portable character classes

Choose ASCII or Unicode deliberately. [A-Za-z] excludes accented letters; \p{L} includes letters from other scripts. \d includes non-ASCII decimal digits, while [0-9] is appropriate for an ASCII-only numeric format. \w includes more than ASCII letters, digits, and underscore.

In evaluated MUSH function arguments, lit() keeps the character-class brackets and regex backslashes literal. Below, accent(cafe,___') is “cafe” with an acute accent on the e, and chr(1635) is the Arabic-Indic digit three:

> think regmatch(foo_bar,lit(\A\w+\z))
1
> think regmatch(foo bar,lit(\A\w+\z))
0
> think regmatch(accent(cafe,___'),lit(\A\p{L}+\z))
1
> think regmatch(accent(cafe,___'),lit(\A[A-Za-z]+\z))
0
> think regmatch(chr(1635),lit(\A\d\z))
1
> think regmatch(chr(1635),lit(\A[0-9]\z))
0

See regexp syntax for anchors, escaping, and the other differences from PCRE.