Regex design pattern for Textmate language grammars. An absolutely cool feature of PCRE (even supported by Oniguruma) are named subexpression calls. Allowing for recursive matching while preserving (atomic) contexts at every recursion level in form of named match results. However, there is no light without shadow: When no multiline matches can b…

demo-grammar.tmProperties

myPattern = {
    name = 'meta.spec.triples';
    begin = '(?x)
    	(?<subject> \g<iri> | \g<BlankNode> | \g<collection> ){0}
    	(?<iri> \g<IRIREF> | \g<PrefixedName> ){0}
    	(?<IRIREF> \< (?:[^\x00-\x20\<\>\\\"\{\}\|\^`] | (?:\\u[0-9A-Fa-f]{4}|\\U[0-9A-Fa-f]{8}))* \> ){0}
    	(?<PrefixedName> \g<PNAME_LN> | \g<PNAME_NS> ){0}
    	(?<PNAME_LN> (\g<PNAME_NS>) ( (?: [\p{L}\p{M}_] | [:0-9] | %[0-9A-Fa-f]{2} | [\\_~.\-!$&''\(\)*+,;=/?#@%] ) (?:(?:[\p{L}\p{M}_\-0-9\u00B7\u0300-\u036F\u203F-\u2040] | [:.] | %[0-9A-Fa-f]{2} | [\\_~.\-!$&''\(\)*+,;=/?#@%])* (?:[\p{L}\p{M}_\-0-9\u00B7\u0300-\u036F\u203F-\u2040] | : | %[0-9A-Fa-f]{2} | [\\_~.\-!$&''\(\)*+,;=/?#@%])) )? ){0}
    	(?<PNAME_NS> (?: [\p{L}\p{M}] (?:(?:[\p{L}\p{M}_.\-0-9\u00B7\u0300-\u036F\u203F-\u2040])* [\p{L}\p{M}_\-0-9\u00B7\u0300-\u036F\u203F-\u2040] )? )? \: ){0}
    	(?<BlankNode> \g<BLANK_NODE_LABEL> | \g<ANON> ){0}
    	(?<BLANK_NODE_LABEL> _: ([\p{L}\p{M}_0-9]) ([\p{L}\p{M}_\-0-9\u00B7\u0300-\u036F\u203F-\u2040.]* [\p{L}\p{M}_\-0-9\u00B7\u0300-\u036F\u203F-\u2040])? ){0}
    	(?<ANON> \[[\x20\x9\xD\xA]*\] ){0}
    	(?<collection> \(\g<object>*\) ){0}
    	(?<object> \g<iri> | \g<BlankNode> | \g<collection> | \g<blankNodePropertyList> | \g<literal> ){0}
    	(?<blankNodePropertyList> \[\g<predicateObjectList>\] ){0}
    	(?<predicateObjectList> \g<verb>\s+\g<objectList> (?:[ ]*;\s+\g<verb>\s+\g<objectList> )* ){0}
    	(?<verb>\s\g<predicate>\s|\sa\s){0}
    	(?<predicate> \g<iri> ){0}
    	(?<objectList> \g<object> (?:,\s\g<object>)* ){0}
    	(?<literal> "([^\x22\x5C\xA\xD] | \\[tbnrf"''] | (?:\\u[0-9A-Fa-f]{4}|\\U[0-9A-Fa-f]{8}))*" ){0}
    (?#Here starts the show)
    ^(?i)(?!@|\#|PREFIX|BASE)(?-i)(?:[ ]*)(\g<subject>)\b
    ';
    end = '(?x)(?#Here starts the show)';
    captures = {
    	ANON = { name = 'meta.tmp.ANON'; };
    	BlankNode = { name = 'meta.tmp.BlankNode'; };
    	IRIREF = { name = 'meta.tmp.IRIREF'; };
    	PNAME_LN = { name = 'meta.tmp.PNAME_LN'; };
    	PNAME_NS = { name = 'meta.tmp.PNAME_NS'; };
    	PrefixedName = { name = 'meta.tmp.PrefixedName'; };
    	blankNodePropertyList = { name = 'meta.tmp.blankNodePropertyList'; };
    	collection = { name = 'meta.tmp.collection'; };
    	iri = { name = 'meta.tmp.iri'; };
    	literal = { name = 'meta.tmp.literal'; };
    	object = { name = 'meta.tmp.object'; };
    	objectList = { name = 'meta.tmp.objectList'; };
    	predicate = { name = 'meta.tmp.predicate'; };
    	predicateObjectList = { name = 'meta.tmp.predicateObjectList'; };
    	subject = { name = 'meta.tmp.subject'; };
    	verb = { name = 'meta.tmp.verb'; };
    };
}