Thursday, February 12, 2009

IRC Logs February 14th

(7:10:58 AM) emily_g: morning jody
(7:12:14 AM) silli: morning all!
(7:12:27 AM) silli: unfortunately moovida is sick
(7:12:36 AM) silli: and can not join the IRC
(7:15:34 AM) emily_g: I have a question about udig libs & building udig (I'm getting a bunch of errors)
(7:15:50 AM) jgarnett: okay cool; lets set an agenda ...
(7:16:12 AM) jgarnett has changed the topic to: 0) what is up 1) udig and building
(7:16:45 AM) jgarnett: I was thinking of talking to moovida about the table view problem I fixed; and seeing if anyone wanted to back port it to 1.1.x.
(7:17:08 AM) jgarnett: anything else worth talking about? We should be doing something with SoC I guess.
(7:17:28 AM) jgarnett: Right now the best thing we can do is make sure we have time to act as sponsors
(7:17:45 AM) jgarnett: sorry mentors
(7:17:55 AM) jgarnett: emily_g are you interested in mentoring this year?
(7:18:25 AM) emily_g: I have no idea what is involved in mentoring ...
(7:18:58 AM) jgarnett: there is a nice page on it somewhere; but the actual work kind of depends on how dialed in the student you get is
(7:19:39 AM) jgarnett: I have gotten away with code reviews and email before; others have had failure even with spending lots of time
(7:19:47 AM) jgarnett: Suppose I will send it to the email list ...
(7:19:49 AM) jgarnett: 0) what is up
(7:20:02 AM) emily_g: I will do some reading and think about it
(7:20:21 AM) jgarnett: jgarnett - fixed table view; not sure if anyone wants that backported to 1.1.x. May get some time to hack this weekend.
(7:20:50 AM) emily_g: emily - still poking away at feature caching
(7:22:37 AM) jgarnett: 1) udig and building
(7:22:47 AM) emily_g: so I updated udig this morning
(7:23:00 AM) jgarnett: so where are you at emily? there was some issue with imageio-ext jars not being deployed to the osgeo repo; but that is fixed now?
(7:23:01 AM) emily_g: and libs gives me errors
(7:23:08 AM) jgarnett: (I have not done that yet; let me try now)
(7:23:12 AM) emily_g: package org.geotools.arcsde.gce.band does not exist in this plug-in
(7:23:29 AM) jgarnett: ah
(7:23:30 AM) emily_g: same for org.geotools.arcsde.gce.imageio & org.geotools.arcsde.gce.producter
(7:23:41 AM) jgarnett: gabriel has been hacking away at arcsde hard these last couple of days
(7:23:50 AM) jgarnett: and some of the package names may of changed
(7:24:00 AM) jgarnett: so here are the two ways it could break us
(7:24:23 AM) jgarnett: - gabriel may be testing in udig; and updating udig as he goes; but if he has not deployed his geotools changes we will have trouble
(7:24:38 AM) jgarnett: - or he may of deployed his geotools changes; and nobody has updated udig libs yet
(7:24:58 AM) emily_g: I think it may be the second
(7:25:19 AM) jgarnett: There is a single button "update" you can do; called organized manifests ... I tend to do it by hand so I can hide a few projects that we
(7:25:22 AM) emily_g: as I don't see any commits related to sde in the last few days
(7:25:50 AM) jgarnett: I am going to duck over to the geotools list and ask
(7:26:30 AM) groldan [n=groldan@host179.201-253-34.telecom.net.ar] entered the room.
(7:26:35 AM) groldan: hello
(7:26:35 AM) jgarnett: hi gabriel
(7:26:46 AM) groldan: hi there
(7:26:53 AM) jgarnett: emily_g reports our udig build broken; something about package changes in arcsde
(7:27:04 AM) jgarnett: did you deploy your changes?
(7:27:16 AM) jgarnett: (if so we need to update udig libs)
(7:27:17 AM) emily_g: package org.geotools.arcsde.gce.band does not exist in this plug-in
(7:27:23 AM) groldan: hum, I only changed package names in the gce stuff
(7:27:28 AM) emily_g: same for org.geotools.arcsde.gce.imageio & org.geotools.arcsde.gce.producer
(7:27:38 AM) groldan: ok, so its just a matter of removing those packages from the manifest
(7:27:55 AM) groldan: yup, there's a single org.geotools.arcsde.gce package now
(7:28:00 AM) groldan: (we're doing more with less code)
(7:28:27 AM) emily_g: sounds good; I will remove them from the udig manifest and commit the fix
(7:28:29 AM) emily_g: thanks
(7:28:31 AM) groldan: so please just remove those packages from the manifest, I don't have a udig dev env at hand
(7:28:38 AM) groldan: thanks emily_g
(7:29:15 AM) groldan: FYI all this fuss is to have something like this: http://arcy.opengeo.org:8080/geoserver/wms?bbox=33001.0,774001.0,333001.0,961999.0&styles=&Format=application/openlayers&request=GetMap&version=1.1.1&layers=massgis:SDE.IMG_USGSQUADM&width=800&height=470&srs=EPSG:26986
(7:29:17 AM) groldan: :)
(7:29:49 AM) emily_g: nice :)
(7:30:17 AM) jgarnett: fun; performance seems good too
(7:31:06 AM) groldan: yup, thanks to JAI :)
(7:31:09 AM) jgarnett: gabriel your email on this subject was so dense I have not made it through yet
(7:31:20 AM) jgarnett: I trust Saul is giving you most of the feedback you need
(7:31:27 AM) groldan: ouch, I liked the rant
(7:31:29 AM) groldan: :/
(7:31:42 AM) jgarnett: is there anything else fun that you would like to talk about?
(7:31:46 AM) groldan: still, most of it shall go end up to the wiky though
(7:32:01 AM) jgarnett: (not a rant; you do not send email - and then you send one big one)
(7:32:16 AM) groldan: he, bad groldan
(7:32:45 AM) jgarnett: does arcsde support nD grids?
(7:33:07 AM) groldan: hum, supports N-band ones
(7:33:17 AM) jgarnett: yeah
(7:33:21 AM) dassouki [n=a_d@fctnnbsc15w-156034075039.pppoe-dynamic.nb.aliant.net] entered the room.
(7:33:24 AM) groldan: but I don't think it matches the nD concept
(7:33:47 AM) groldan: they're _rasters_, 2D
(7:34:00 AM) jgarnett: It kind of can; if you make your own color model based on the number of bands etc...
(7:34:02 AM) jgarnett: oh okay
(7:34:06 AM) jgarnett: that is fine; just asking
(7:34:19 AM) groldan: and yeah, I'm having a lot of fun to support raster catalogs now
(7:34:37 AM) jgarnett: reading your email about "fail if there is no support for the raster's format and number of bands" - I think you do not need to fail is all
(7:34:41 AM) groldan: see my previous chat with simboss at #geoserver
(7:34:54 AM) jgarnett: thinking
(7:35:02 AM) jgarnett: so what do you need to hook this up in uDig?
(7:35:13 AM) groldan: a little bit of time
(7:35:18 AM) groldan: (the spare one)
(7:35:19 AM) jgarnett: we need to be able to discover what grid coverages are available; how is that handled on the arcsde side?
(7:35:31 AM) groldan: but its something I really want to do
(7:35:41 AM) groldan: raster_columns table
(7:36:00 AM) emily_g: jgarentt - i removed those packages and now I get another error - libs is missing library lib/beam-1.0.jar; does this mean anything to you?
(7:36:08 AM) jgarnett: cool
(7:36:11 AM) jgarnett: thinking
(7:36:48 AM) groldan: yeah I can add an ArcSDERasterFormat.getAvaliableRasters():List method with ease
(7:36:52 AM) groldan: or such
(7:37:05 AM) silli left the room.
(7:37:17 AM) jgarnett: I have it in mine emily_g; but I can run again. What it would mean (if it is not there and you did your libs from scratch) is that beam is no longer a dependency
(7:37:33 AM) jgarnett: (we get random jars dropped on us via maven dependency; sometimes they go away as well)
(7:37:56 AM) jgarnett: groldan++ if you have a method like that we can make a list of raster children
(7:38:04 AM) jgarnett: um but are we likely to have a list of thousands?
(7:38:15 AM) jgarnett: like more than the number of tables
(7:38:38 AM) jgarnett: do they have any kind of organization to the raster? Or perhaps the rasters can be very big ... not really sure of the arcsde model
(7:40:52 AM) groldan: catching up...
(7:41:24 AM) groldan: list of thousands: you'll have a list as big as the number of rasters are registered, just like for FeatureTypes
(7:41:35 AM) jgarnett: fair enough
(7:41:41 AM) groldan: I don't quite get why you say more than the number of tables?
(7:41:56 AM) jgarnett: hopefully we can get some good descriptive information out of them as well; so users can find what they want
(7:42:11 AM) groldan: hehe, I'm your friend here
(7:42:21 AM) jgarnett: I have seen rasters represented as values in a database table before
(7:42:26 AM) jgarnett: a big old blob
(7:42:30 AM) groldan: check out http://arcy.opengeo.org:8080/geoserver/ows?service=WCS&request=GetCapabilities
(7:42:59 AM) groldan: look at the abstract info for the ones in the massgis namespace
(7:43:11 AM) groldan: I'm building quite a nice ResourceInfo here
(7:43:21 AM) jgarnett: SDE.IMG_COQ2001_CLIP_BOS_1 is an ArcSDE Raster
(7:43:25 AM) groldan: and made the geoserver config to get the default description from ResourceInfo
(7:43:32 AM) jgarnett: really? SDE.IMG_COQ2001.....
(7:43:38 AM) jgarnett: for the title :-)
(7:43:47 AM) groldan: uh?
(7:44:24 AM) groldan: what tile?
(7:44:27 AM) jgarnett: this looks like the ID has been turned into a tilte
(7:44:52 AM) jgarnett: for massgis:SDE.IMG_BWORTHOS_CLIP_BOS the title is SDE.IMG_COQ2001_CLIP_BOS_1 is an ArcSDE Raster
(7:45:24 AM) jgarnett: a lot of the titles have the " is an ArcSDE Raster"
(7:45:41 AM) groldan: agreed, bad
(7:46:00 AM) groldan: just the old geoserver way "XXX is a YYY Raster"
(7:46:34 AM) groldan: I can get something more useful from the table metadata but found they hardly set it
(7:47:01 AM) groldan: still, fallbacking to the table name seems like the only solution as for what ResourceInfo concerns?
(7:47:22 AM) jgarnett: agreed
(7:47:36 AM) jgarnett: I should check if ResourceInfo is allowed to have a null title
(7:47:46 AM) jgarnett: we are allowed to point out to users when their title is bad :-)
(7:47:46 AM) groldan: I realize you're thinking in udig where there's not an admin setting nice titles for you like in geoserver
(7:48:13 AM) groldan: we are
(7:48:21 AM) jgarnett: well if you give me the API
(7:48:33 AM) jgarnett: we can let users update the metadata from udig
(7:49:03 AM) jgarnett: aside: our idea of introducing "info" to geotools has kind of failed
(7:49:17 AM) jgarnett: we did not push it through the other datastore implemntations and thus it has not been picked up
(7:49:50 AM) jgarnett: emily_g how goes the build? still stuck on beam?
(7:50:05 AM) emily_g: I deleted it from the class path and everything works fine for me
(7:50:26 AM) emily_g: should I commit that change?
(7:51:07 AM) groldan: ok, deal: http://jira.codehaus.org/browse/GEOT-2337
(7:51:20 AM) jgarnett: yep
(7:51:38 AM) jgarnett: I am doing a fresh build here to make sure beam does not get dragged in from refresh.xml
(7:51:46 AM) groldan: I gotta go back trying to figure out the JAI Mosaic operation
(7:51:54 AM) jgarnett: it could be that beam is new; and the dependency has not been deployed to the repo yet
(7:51:57 AM) groldan: thanks for the chat, see you
(7:52:00 AM) jgarnett: ciao
(7:52:51 AM) CIA-8: UDig: egouge * r31109 udig/plugins/net.refractions.udig.libs/ (.classpath META-INF/MANIFEST.MF): fixes for updates to sde geotools jars & removal of beam from classpath
(7:53:10 AM) emily_g: I have another more geotools related question for you - I would like to add a bit of documentation about this caching stuff I've done.
(7:54:20 AM) emily_g: I think this is the original page - http://docs.codehaus.org/display/GEOTOOLS/SoC+2007+Caching+data
(7:54:41 AM) emily_g: should I modify this page , add a new one, or ask the geotools list?
(7:55:57 AM) jgarnett: question
(7:56:01 AM) jgarnett: is this thing its own module?
(7:56:23 AM) emily_g: ya it's an unsupported module
(7:56:25 AM) jgarnett: make a page here (http://docs.codehaus.org/display/GEOTOOLS/Module+Matrix) and link back to the SOC page
(7:56:25 AM) emily_g: gt-caching
(7:56:34 AM) jgarnett: we try and have just one developer notes page
(7:56:49 AM) jgarnett: this SOC page was kind of project notes; most students used it to report on their timeline and what they were doing
(7:57:26 AM) jgarnett: does that help?
(7:57:39 AM) jgarnett: Hopefully your hard work will leave this thing in a state it can be used
(7:57:46 AM) jgarnett: I never even asked for user documentation for this one
(7:58:43 AM) emily_g: i hope so too but I"m not use; I read through aaime list of requirements of a cache and this cache isn't that sophisticated
(7:59:10 AM) jgarnett: aaime has been talking about a cache
(7:59:16 AM) jgarnett: for 4 years
(7:59:20 AM) jgarnett: there are two good ways to do it
(7:59:28 AM) jgarnett: and nobody really knows which is great
(7:59:32 AM) jgarnett: until we build and profile
(7:59:39 AM) emily_g: So I'll add a link from the modules matrix page to the soc page and update the soc page with the modifications I've made
(7:59:41 AM) jgarnett: so I would not worry; it is faster then remote WFS access :-)
(7:59:53 AM) emily_g: yup :)
(7:59:55 AM) jgarnett: um
(8:00:05 AM) jgarnett: I would rather see you update the model matrics page
(8:00:10 AM) jgarnett: talk about the code
(8:00:51 AM) jgarnett: the SoC 2007 is over right?
(8:00:57 AM) jgarnett: what you are doing now is about the module
(8:00:59 AM) emily_g: yup
(8:01:07 AM) jgarnett: yup
(8:01:13 AM) emily_g: okay I'll make a new page for the module
(8:01:52 AM) jgarnett: I am running refresh.libs now
(8:02:12 AM) emily_g: what do the different colored stars mean?
(8:02:13 AM) jgarnett: if beam is not around then I you can commit your change. If not we have a mystery to solve first
(8:02:35 AM) jgarnett: At the bottom of the module matrics page it tells you
(8:02:46 AM) jgarnett: http://docs.codehaus.org/display/GEOT/Gold+Star+Quality+Assurance+Check
(8:03:01 AM) jgarnett: really primitive metric to report to users the state of maturity of the module
(8:03:09 AM) emily_g: umm I already committed my change to remove beam ...
(8:03:12 AM) jgarnett: idea stolen from the maven plugin system
(8:03:22 AM) jgarnett: ha ha; okay
(8:03:30 AM) emily_g: ya I got that; I was wondering what the difference was between red and yellow
(8:04:44 AM) kartben [n=kartben@pur31-3-82-240-189-250.fbx.proxad.net] entered the room.
(8:04:58 AM) jgarnett: yellow is good
(8:05:00 AM) jgarnett: red is bad
(8:05:07 AM) jgarnett: pass or fail
(8:05:28 AM) jgarnett: I would of just listed the number of stars but that is me
(8:05:34 AM) emily_g: ok I though no star meant fail -
(8:05:42 AM) emily_g: thanks
(8:06:19 AM) jgarnett: yeah I did as well
(8:09:46 AM) jgarnett: okay thanks for the meeting

No comments: