Hi,
the shapefile-ng work is nearing completion.
I have all the tests passing now (at least, once the transaction stuff
for content data
store gets in) and the code seems to work, in test harness, for both indexed and
non indexed shapefiles.

In particular thanks to content data store this one should support
sorting for good.

General highlights:
- the basic underpinnings (low level readers and writers) are the same
as shapefile
  data store (did not get around to try a cleanup there, and bug
reports are piling up,
  have stop doing R&D and fix at least some of them).
- the top level is rewritten to use content data store
- there is a single configurable store instead of two (indexed or not)
- the non indexed version should be somewhat faster

Issues I see in a migration:
- not nearly as real world tested as the old code
- the old code has a public interface large from here to the moon, with a ton
  of methods that should have never been made public (lots of methods that
  were not in the DataStore interface to grab readers and writers for example)
- the old code had a billion constructors with all the possible options, the
  new one has _one_ constructor taking a URL, and then has getters/setters
  for all the various options (turn off indexing, timezone, charset)
- the code is sitting, at least for the moment, in a
org.geootools.data.shapefile.ng
  package

Of course migration won't be instant for all that code that used directly the
extended api the old store was providing.

In any case, I was wondering how to handle the upgrade.
Moving the classes in the same package might help the migration, but at
the very least I would like to get rid of the "ng" since that's going to become
the default thing (if we do another rewrite, are we going to mark it as
"aab"? (above and beyond?)).
Maybe be consistent with jdbc and move it to org.geotools.shapefile

And then what... have the new code pick up the old params in case it's the
only one in the classpath, otherwise redirect to the old store?

We also need to test it out quite a bit in all the GeoTools code that depends on
it (test, image mosaic), in GeoServer and uDig. Still quite a bit of work ahead.

Cheers
Andrea

-- 
-------------------------------------------------------
Ing. Andrea Aime
GeoSolutions S.A.S.
Tech lead

Via Poggio alle Viti 1187
55054  Massarosa (LU)
Italy

phone: +39 0584 962313
fax:      +39 0584 962313

http://www.geo-solutions.it
http://geo-solutions.blogspot.com/
http://www.youtube.com/user/GeoSolutionsIT
http://www.linkedin.com/in/andreaaime
http://twitter.com/geowolf

-------------------------------------------------------

------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure 
contains a definitive record of customers, application performance, 
security threats, fraudulent activity, and more. Splunk takes this 
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to