Hi
,
I have been
profiling an application that parses many thousand, ver small xml
documents.
I have noticed that
the impl,entation of the SAXReader consumes more CPU and generates more garbage
form the initialisation of the XMLReader and the ContentHandler that it takes in
the parsing
some small
modifications to the SAXreader so that the XMLReader and ContentHandler are not
is not re-initialised, yealded a four times less garbage and 3 fold less
CPU usage
I am happy to
contribute the changes if anyone is interested
profiling was from
jprobe, against dom4j 1.6.1, xerces 2.7.1 and java 1.4.2
Mike
