Just Justin and me about bleeding edge complex-features, fm, cpx branches
feb 27 22:03:10 <groldan> jdeolive, hi
feb 27 22:13:47 <jdeolive> hi gabriel
feb 27 22:13:51 <jdeolive> groldan
feb 27 22:13:58 <groldan> hey
feb 27 22:14:09 <groldan> how is it going?
feb 27 22:14:22 <jdeolive> good
feb 27 22:14:41 <jdeolive> i am merging changes from complex-features onto fm
feb 27 22:14:47 <groldan> wen I wasl almost done.. had no chance but making a wide api change... are you porting complex-features to fm?
feb 27 22:14:54 <jdeolive> yes
feb 27 22:15:13 <jdeolive> although the new geoapi changes conflict, but i am working around the conflicts
feb 27 22:15:17 <groldan> ok, may be you should freeze for a sec. I need to agree on a topic
feb 27 22:15:24 <jdeolive> ok
feb 27 22:15:55 <groldan> the point is: Attribute.getType() is insufficient, we need Attribute.getDescriptor(): AttributeDescriptor
feb 27 22:15:56 <groldan> instead
feb 27 22:16:04 <groldan> rationale:
feb 27 22:16:06 <jdeolive> i thought thats what it was
feb 27 22:16:50 <groldan> nope, we had only getType(), which means with the separation of concerns between attribute name and type name isnt enough
feb 27 22:16:54 <jdeolive> at least the latest version of the geoapi interfaces have that
feb 27 22:16:58 <groldan> actually getType() alone has no sense?
feb 27 22:17:18 <groldan> cool, let me check
feb 27 22:17:25 <jdeolive> there is a convenience method on Attribute, called type()
feb 27 22:17:37 <jdeolive> which is i would imageine is just shorthand for getDescriptor().getType()
feb 27 22:18:14 <groldan> meaning I'm moving all methods from AttributeFactory to create(AttributeDescriptor,...) instead of (AttributeType,...)
feb 27 22:19:00 <jdeolive> cool, that makes sense
feb 27 22:19:05 <groldan> ok, I missed the new interface has it, thats so cool
feb 27 22:19:16 <jdeolive> phew, glad i already picked up this one
feb 27 22:19:26 <jdeolive> i am actually starting to get the hang of this new model
feb 27 22:19:29 <groldan> I'm making the change on complex-features, may be it'll be better for you to get it from there?
feb 27 22:19:42 <jdeolive> yeah, i will merge it on, i will probably get a conflict but that is ok
feb 27 22:19:45 <jdeolive> go ahed
feb 27 22:19:53 <jdeolive> also, the build for geotools-cpx seems to be failing
feb 27 22:20:15 <groldan> yeah, I was fixing it as soon as I received the cc report
feb 27 22:20:23 <jdeolive> cool
feb 27 22:20:38 <groldan> one of the times I had to make a maven clean to pick the error in my dev environment
feb 27 22:20:54 <jdeolive> yup, good old maven
feb 27 22:21:31 <groldan> is there a meeting today? I thought the time was an hour sooner than now
feb 27 22:21:46 <jdeolive> noone really showed up
feb 27 22:21:48 <jdeolive> so i guess not
feb 27 22:22:02 <jdeolive> however, this discussion could have been our meeting 
feb 27 22:22:04 <groldan> well, not so bad, may be it gives me time to finish this
feb 27 22:22:13 <groldan> sure 
feb 27 22:22:37 <groldan> well, I'm happy this is the only remaining change
feb 27 22:22:43 <groldan> report:
feb 27 22:22:43 <jdeolive> so where is geoserver on the complex-stuff
feb 27 22:23:10 <groldan> - I have a _complex_ schema being mapped from sde to complexds
feb 27 22:23:26 <jdeolive> woohooo
feb 27 22:23:43 <groldan> - the schema is broken into various xsd's, complexds loads them alright, type reusing is ensured
feb 27 22:23:55 <-- eighty has quit (Remote closed the connection)
feb 27 22:24:12 <groldan> (nasty side effect is that I'm using TypeFactory as type and attributedescriptor registry)
feb 27 22:24:28 <groldan> but that could be fixed once we decide how to treat that issue
feb 27 22:24:34 <jdeolive> yeah, sounds minor
feb 27 22:25:00 --> eighty (n=eighty@adsl-71-132-249-0.dsl.pltn13.pacbell.net) has joined #geotools
feb 27 22:25:05 <groldan> -arcsde plugin is already ported to separate concerns between types and descriptors
feb 27 22:25:42 <jdeolive> cool
feb 27 22:25:48 <groldan> as a workaround, all the other datastores should be working as they used, 'cause an attribute descriptor is automatically created with the same name as the type
feb 27 22:26:15 <groldan> this is meant as temporal behavior, while in the process of porting datastores to "reuse type definitions"
feb 27 22:26:21 <groldan> sounds reasonable?
feb 27 22:26:27 <jdeolive> yup
feb 27 22:26:31 <jdeolive> sounds great
feb 27 22:26:54 <groldan> cool, we'll have to make a "datastore porting guide" on the wiki
feb 27 22:27:02 <jdeolive> definitley
feb 27 22:27:05 <jdeolive> right on on fm
feb 27 22:27:13 <jdeolive> i am going to ignore the datastore plugins
feb 27 22:27:14 <groldan> fortunately the change, though somewhat "wide", should be trivial
feb 27 22:27:23 <groldan> sure
feb 27 22:27:32 <jdeolive> until i get the core modules going
feb 27 22:27:46 <jdeolive> then we can call for help for the datastores and write up some docs
feb 27 22:28:18 <groldan> cool. After I get this "proof of concept" release out, I'll start helping you with the merge
feb 27 22:28:40 <jdeolive> good good, that release is definitley priority
feb 27 22:28:52 <jdeolive> so i am going to update, and get your latest changes
feb 27 22:28:52 <groldan> say, wednsday?
feb 27 22:29:07 <jdeolive> for release?
feb 27 22:29:10 <groldan> wai, have almost 200 files to commit
feb 27 22:29:19 <jdeolive> oh, ok
feb 27 22:29:22 <groldan> if god helps, release tomorrow
feb 27 22:30:18 <jdeolive> sure just let me know when you want to say go, and we can do it
feb 27 22:30:26 <jdeolive> although i dont have a release process set up for 1.4.x stuff yet
feb 27 22:30:33 <jdeolive> but it wont take me long to script it up
feb 27 22:30:48 <jdeolive> definitley be done by wednesday
feb 27 22:31:26 <groldan> well, I first will send bgs guys a self contained project demostrating their needs in unit tests
feb 27 22:32:10 <groldan> then will make it working on cpx and tell them to svn checkout and test with provided config.zip
feb 27 22:32:24 <groldan> not sure if a proper release is needed
feb 27 22:32:43 <groldan> I think the _release_ should go out from fm
feb 27 22:32:48 <jdeolive> alrighty
feb 27 22:33:03 <jdeolive> yeah that makes sense, complex-features is still a bit rough around the edges
feb 27 22:33:23 <groldan> exactly
feb 27 22:33:32 <CIA-10> jeichar * r18298 udig/plugins/ (9 files in 8 dirs): working on changing Layers View kind of squished now though
feb 27 22:33:39 <groldan> well, back to work, so you'll have the code to merge sooner.
feb 27 22:33:44 <groldan> ping me for any questions
feb 27 22:33:58 <jdeolive> sounds good, i am taking off for a couple of hours
feb 27 22:34:04 <groldan> also, could you upload this chatting as today's meeting?
feb 27 22:34:05 <jdeolive> later
feb 27 22:34:06 <groldan> ok
feb 27 22:34:10 <jdeolive> yes i will