I am trying to port an xml parser to using lxml.etree. I can get the parser to work find and get it to validate properly and produce a parse tree in tuple form ie (tag, attrib, [contents...], extra) The standard xml error locations are find and provide both line number and column. However, the current parser allows for debugging during post processing of the parse and it has a tuple of information in place of extra above that looks like ((srcname,startline,startcolumn),(srcname,endline,endcolumn)). This extra information allows post analysis to determine if one tag starts after another. I can find no way to access the column information in the standard parsers. I believe that the information is present in the XMLReader that libxml2 provides, but no way to get access in lxml. I think I just need to determine a tag ordering ie does tag0 start before or after tag1 in the source. Is there an obvious way to do that? Currently (tag0.startline, tag0.startcolumn) is compared with (tag1.startline, tag1.startcolumn) and the latest tag is returned. I believe I could just add a tag sequence to determine order, but is there an easier way? -- Robin Becker