Hi,

I think I found an issue with the XPath evaluation of REXML, as I am currently comparing multiple XPath libraries.

If we pass an XPath expression to REXML which should evaluate a condition on nodes, it evaluates it on every node. While this may seem intuitve, it differs from the XPath 1.0 Standard, as it states there:

If one object to be compared is a node-set and the other is a number, then the comparison will be true if and only if there is a node in the node-set such that the result of performing the comparison on the number to be compared and on the result of converting the string-value of that node to a number using the number function is true.
(Source: https://www.w3.org/TR/1999/REC-xpath-19991116/ Chapter 3.4 Booleans)

Below is a simple demonstration example tested on ruby 2.6. The output is falsetruetruefalse

require "rexml/document"
include REXML
string = <<EOF
  <?xml version="1.0"?>
  <bookstore>
      <price>123</price>
      <price>456</price>
      <price>789</price>
      <price>234</price>
  </bookstore>
EOF
doc = Document.new string
puts XPath.match( doc, "//price>400")

This is significantly different from the other libraries I investigated:

Testing library: nokogiri , xpath: //price>400
true
Testing library: xqilla , xpath: //price>400
true
Testing library: rexml , xpath: //price>400
falsetruetruefalse
Testing library: xalan-j , xpath: //price>400
true
Testing library: saxon , xpath: //price>400
true
Testing library: jaxen , xpath: //price>400
true
Testing library: lxml , xpath: //price>400
true
Testing library: VTD-Gen , xpath: //price>400
true

Basically I think to fix this, a further evaluation step is needed. Instead of returning the Results of the comparison directly, it should be evaluated if one of the results is true and then just return true.

I hope this helps and best regards,
Mirko