> it's not libxml2. you will see that libxml2 by itself on such a > box should parse your 36MB of XML in a mere second. Try with xmllint --stream > for example. The potential for improvements by solving the python string > import would have been huge if you did want to look at it, but it seems > obvious you didn't want. > Daniel, But we're not talking about libxml2 in C, we're talking about using it in python. I agree with you, we could have worked on the python string handling and it's interface to libxml2. But time is not free and Icon nor I had the inclination to work on it. Icon saw elementTree, it proved to do the same job and do it fairly well. It was low-hanging fruit. If you or anyone else would like to implement the functions in libxml2 and make the import work just as quickly and the api look just as pretty and I'll be glad to take a look at the results. But I just don't have the time to do it. Remember, some of us are volunteers. -sv
