Hello,
Attached are further improvements to XML handling classes. I am sending the whole files instead of patches, because patches are larger than the files :)
During the time passed, I have got my hands on a test suite from w3.org and on a profiler. So the code now is much more "mature" than in the previous changes.
The current status is:
- xmlwrite.pp: Implemented built-in buffering with simultaneous UTF8 encoding, which made it faster about 3 times.
- dom.pp: Implemented a number of missing methods; wrote a skeleton for namespace support (not working yet); implemented timestamping for TDOMNodeList (so it is rebuilt only when underlying nodes are changed). The complete list of changes is in 'FIX:' and 'DONE:' comments in dom.pp itself.
- xmlread.pp: Numerous fixes in order to improve performance and comply with the testsuite.
