MgWktReaderWrite performs horrendously on large Polygons

[stevedang]

Attempts to process the large polygons result in the the Web Tier consuming 97% of the cpu process when trying to read the polygon either in SDF format using an MgAgfReaderWriter object or in WKT format using an MgWktReaderWriter object. The largest polygon appears to process for more than an hour or an hour and a half. The process has never been left to run to completion as the API testharness calling the script always times out before the execution of the script completes.

The ReaderWriters appear to process fine on small polygons, but the performance is unacceptable on the large ones.

Note the priority on this is high because of the usability/performance aspect of the defect. Technically the server does not crash during the operation, but if something like this is executed first then the MapGuide Web Tier may drop requests and become unusable for clients.

[stevedang]

Fixed by submission <<​http://trac.osgeo.org/mapguide/changeset/3234>>