Feature suggestions and discussion [READ THE FIRST POST]

gallery_id:1 || gallery_id:2 || gallery_id:3 || gallery_id:4 || gallery_id:5, [search term]

gallery_id:(1 || 2 || 3 || 4 || 5), [search term]

1. By adding this to the grammar it creates an ambiguity of searching for a literal term with the same qualifiers.  
   Is description:(yes || no) a boolean OR for two different description keywords, or is it a search for images with the literal description “(yes || no)”?
    
2. If you answered yes to the former premise in 1, how do you search for the literal description? If you instead answered yes to the latter premise in 1, then how do you resolve the lexical ambiguity in the language? In the second case, for a numeric field, the lexer would see  
   \[SYM\_intfield, 'gallery\_id'\] \[SYM\_lparen, '\('\] \[SYM\_lit, '1'\] \[SYM\_space, ' '\] \[SYM\_or, '\|\|'\] \[SYM\_space, ' '\] \[SYM\_lit, '2'\] \[SYM\_rparen, '\)'\]  
   and parsed like this  
   if field = accept\(SYM\_intfield\)   if accept\(SYM\_lparen\)     node = int\_conj\(field\)     expect\(SYM\_rparen\)     node   else     IntegerNode.new\(int, field\)   end end
    
   which in theory is fine because it’s not ambiguous for integers, but when it gets to literals you have a problem, because you have multiple constructions based on the potential grammar you want to see implemented. One of them could be  
   \[SYM\_litfield, 'description'\] \[SYM\_lparen, '\('\] \[SYM\_lit, 'yes'\] \[SYM\_space, ' '\] \[SYM\_or, '\|\|'\] \[SYM\_space, ' '\] \[SYM\_lit, 'no'\] \[SYM\_rparen, '\)'\]  
   and the other would be  
   \[SYM\_litfield, 'description'\] \[SYM\_lit, '\(yes \|\| no\)'\]  
   which have obviously different parser reductions.
    
   The first tokenization, which is implemented in the manner you describe, can only be parsed in terms of the “boolean conjunction” approach. The second can only be parsed in terms of the “literal” approach.

([terms]), -my:uploads