On Thu, Oct 21, 2010 at 8:35 PM, Julian Reschke <julian.resc...@gmx.de> wrote: > "The purpose of the exsl:node-set function is to return a node-set from a > result tree fragment. If the argument is a node-set already, it is simply > returned as is..."
If you've a node-set already then what's the use of passing that again to node-set function? Why can't the original node-set be used as it is? The most important aspect of node-set function is to convert RTF to node-set, and that works fine with Xalan. Do you really need an identity conversion of a node-set to a similar node-set for practical needs? If yes, then is the type of input to node-set function known statically, before compilation of stylesheet -- if yes then I believe, another call to node-set with node-set as argument is not really needed; I think then we can just use the original node-set reference. -- Regards, Mukul Gandhi
