
"Yitzchak Gale"
In some usage patterns, it can reduce the cost of processing from linear in the size of the document, to a constant (the distance into the document to find a particular element).
Oh oh - does that mean that Ketil's original case (an element containing a large quantity of CDATA) could still be a problem?
Not necessarily. If the CDATA is not actually needed, it is possible that it would simply be discarded automatically by the lazy demand pattern. That does depend very much on how the consumer is written however. HaXml will still require a rather large amount of space to _lex_ the 25k line text element into a single token of course (I estimate no bigger than about 3Mb though). I have been considering moving the lexer to use ByteString instead of String, which would neatly solve that problem too. Regards, Malcolm