atom feed5 messages in com.marklogic.developer.general[MarkLogic Dev General] bounty hunting
FromSent OnAttachments
Patrick WiedSep 27, 2010 4:27 pm 
Mike BrevoortSep 28, 2010 7:20 am 
Patrick WiedSep 29, 2010 1:20 pm 
Gunther RademacherSep 30, 2010 1:18 pm 
Eric BlochSep 30, 2010 2:35 pm 
Subject:[MarkLogic Dev General] bounty hunting
From:Gunther Rademacher (gr@gmx.net)
Date:Sep 30, 2010 1:18:03 pm
List:com.marklogic.developer.general

Hi all,

what a pity - I only noticed this challenge recently, way too short before the close of this month.

Now I hope that the acceptance tests are hard enough to get the competition prolonged.

A problem that definitely should be solved is the proper distinction of keywords and identifiers, in the absence of reserved words. Sorry to say, and I may be missing something, but I have not seen a solution here addressing this satisfactorily.

Is it necessary to point out that the distinction highly depends on a token's context? Repetitive application of regular expressions, without proper context detection, will not be sufficient, and enumerating some special cases does not provide a complete solution either. In fact I believe that the best way to tokenize XQuery is feeding back information from a full-fledged parser into the lexer.

Of course the environment must provide appropriate interfaces for integrating that. From a quick check it seems to me that CodeMirror would be my first choice to go for.

So if there is any chance to extend the timeframe of this challenge, I would gladly attempt a CodeMirror approach.

Best regards Gunther