[Home] [By Thread] [By Date] [Recent Entries]
At 2012-09-20 13:56 +0100, Andrew Welch wrote:
> I saw a lot of comments talking about "[1]" and "[last()]" talking about Granted, Andrew, thank you. Of course you are right that it is based on the context list which is dynamically changing all the time. But I understood the discussion to be oriented around "@*[1]" and "@*[last()]", which when the abbreviation is expanded is "attribute::*[1]" and "attribute::*[last()]", which itself is oriented around nodes found along the attribute axis attached to an element. The predicates are not being applied to the artefacts found as markup in start tags. For all we know an implementation might have alphabetized the nodes along the attribute axis in order to stop looking for an attribute when the name being looked for comes before the name of a given node on the axis! So the specs say there is no order to attribute specifications and defaulted attributes, and the building of the XDM model imposes an order of attribute nodes on the attribute axis any way it wants or needs from the attribute specifications and defaults. . . . . . . . . . Ken
|

Cart



