This is very exciting stuff Dave, I can't wait to see it slip into
the uDig stream (v. v. soon, if we're on nightlies :)
One comment, on:
* changed WKB parser to optimized WKB parser (slightly modified
JTS version)
Have you sent your optimization back to Martin?
P
On Apr 18, 2006, at 1:31 PM, David Blasby wrote:
Changes
---------
NOTE: test cases are mostly turned off for "maven 2" and non-
functional in "maven 1". I ran the tests in eclipse and they pass.
NOTE: I had awful trouble moving my changes from 2.2.x to trunk.
As I mentioned last week, I made a bunch of changes to improve
performance and fix bugs. I'm now commiting the changes, so I'm
telling you again.
Everything should work as before, but be faster. I made a few
changes to the labeling algorithm so if you dont like the new way
let me know whats different and why it looks "bad". The new
labeling algorithm is faster.
1. main
* added LiteCoordinateSequence & ...Factory to org.geotools.util
so it can be shared.
* added StyleAttributeExtractorTruncated which is a subclass of
StyleAttributeExtractor but doesnt pickup 'things' that are
directly grabbed from the
feature during rendering -- ie. Geometry, TextSymbolizer's
label, TextSymbolizer's
priority. See the javadoc. This fixed a bug in the renderer
and gave a "free"
performance boost.
* StyleAttributeExtractor - added javadocs
* CompareFilterImpl - fixed several bugs & increased performance
(unable to apply to trunk)
* JDBC1DataStore -- changed default fetch size from 200 to 1000
2. PostGIS
* added (optimized) Base64.java class. This is a public domain
library.
* changed WKB parser to optimized WKB parser (slightly modified
JTS version)
* WKB parser uses ObjectInputStream.readDoubles() native method
to parse
doubles. Falls back to a 100% pure java method if there's
problems.
* uses the LiteCoordinateSequenceFactory to make Geometries
instead of the
default JTS version.
3. Renderer
* moved LiteCoordinateSequence & ...Factory to org.geotools.util
so it can be shared.
* changed the line labeler so it networks using an approx O(n)
algorithm instead of the old O(N*N)
* line labeler now chooses the rotation of the line based on the
instantaneous slope of the line
at the label point instead of the slope between the 1st and
last point. Some cases this
is better, some cases its slight worse. Its easy to change
back if there's feedback.
* new line iterator "LiteIterator2"
* decimation changed so its more efficient.
* use LiteCoordinateSequences
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting
language
that extends applications into web and mobile media. Attend the
live webcast
and join the prime developer group breaking into this new coding
territory!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
