[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [udig-devel] Which branch to choose?
|
Good morning Mauro:
Hello everybody,
I'm approaching Udig development to build customized GIS Viewers.
We have previously developed a GeoTools DataStore to use in GeoServer.
The DataStore is written for GeoTools 2.3.x (to use it in GeoServer
1.5.x) and we are currently migrating it to GeoTools 2.4.x (to use it
in GeoServer 1.6.x).
Right; we don't have a version of uDig that works against either of
those releases; Frankly two things prevent us from using GeoTools 2.3 or
2.4:
- without a client application like uDig around much of the
functionality we need simply does not work
- you cannot edit shapefiles without deadlock (a complete show stopper)
- the transition to filter 1.1 is incomplete something we make use of a lot
- feature event notifications as datastores are edited are not always sent
- and some of the "Aggregate" functions we use for the style themer do
not work
Obviously we'd like to reuse our DataStore, implementing a new input
source for Udig as well, but unfortunately:
- current version (1.1.x branch?) is using GeoTools 2.2.x (a little
bit old for us, if we wish to reuse the DataStore as is). I managed to
build this branch locally following your documentation, so we could
start "developing" in a short time.
- trunk is using GeoTools 2.5.x (too new :-) ). I tried to checkout
trunk, but I didn't get to a succesful build. Maybe someone an help me
to accomplish it, but firstly I need to understand if developing on
trunk is feasible if we need to build a "stable" product.
Did you have a problem with geotools trunk or uDig trunk? I would be
happy to talk you through building either one of these on IRC or Skype
if needed.
How soon does your product go out? I am working on two projects right
now that will be completed in the next six months; both of them on
"trunk". Although I would like to use a "stable" geotools as a base one
simply is not avaialble that is up to the needs of a desktop
application. As such I need to make it myself; and am contributing my
time to GeoTools trunk to make it happen.
The good news is the "Scary" stuff for GeoTools 2.5 is done; ie the
transition to the new feature model. There are no further planned API
changes on GeoTools trunk; just clean up work.
And here is an update on the clean up work:
- shapefile support has been rewritten; it no longer deadlocks and has
suffered a complete code review
- the transition to filter 1.1 is complete from the perspective of a
client application (you no longer have to use org.geotools.filter.Filter
anywhere in your own code). Internally we have a 40% of the
"FilterVisitors" to update
- FetureEvent problems are treated on a case by case basis as they cause
redraw problems in uDig
- Aggregate functions still need to be fixed up; mostly methodical test
cases need to be written
So, we're in the middle of the "way" (2.3.x - 2.4.x) of what Udig is
currently using (2.2.x - 2.5.x). Either choice would mean to branch
our DataStore.
I would not worry about branching your datastore forward to geotools
trunk; you will want to move to the latest geoserver eventually and the
more bug reports you can provide them the better.
Now, my question is: have you got any suggestion on the branch to
choose? Do you intend, in a short time, to evolve 1.1.x to a more
recent geotools version?
Not at all; the only thing intended for 1.1.x is a "mac os x" installer,
and a French translation. Nobody is paying us to work on 1.1.x; and we
have work in hand for 1.2.x.
I don't know if we can give you any help in this direction (I have
just started "studying" Udig), but I'll be glad to do what I can.
There really is nothing that can be done; what is needed is bug fixes
for GeoTools 2.4.x and I think that effort can be better applied to
GeoTools trunk.
Jody
