I was pleased to attend a Geotools presentation for EoGeo that James Macgill put together.
The presentation is available here, and was well recieved.
James also scored a bit of a write up here.
James and Jody will be available for a BOF at the Where 2.0 conference this week.
The Java Open Source geospatial community is thriving. Java GIS applications cover the full spectrum, from backend services (such as GeoServer, OpenMap and DeeGree) to client solutions (such as uDIG, JUMP).
The community is centered on a core set of libraries (GeoTools, JTS and DeeGree) and collaboration is facilitated by the use of common interfaces (GeoAPI) based on open standards (ISO and OGC).
The lead developers of GeoTools and uDIG will be there and all Java geospatial projects are more than welcome.
The GeoTools team is pleased to announce the second release candidate for the 2.1 edition.
This release, and with the switch to the 2.0 version of GeoAPI completed, only fixes to bugs lie between here and the final 2.1 release.
If you are developing aginst the 2.0 branch then now is a good time to start thinking about switching to 2.1 as all new (and thus experimental features) are being done on the 2.2 branch with only bug fixes and optimizations scheduled for the 2.1 branch.
For more information please see the 2.1.RC1 download page.
jmacgill Hi all
jgarnett Hi James!
jgarnett See the subject line, I am still near the conference and there is a weather warning out.
jmacgill who is awake here?
jgarnett I was going to ask if the 2.1.RC1 has been anounced yet (like on the website?)
ERROR Connection to irc://freenode/ (irc://irc.freenode.net/) reset.
INFO Connecting to irc://freenode/ (irc://irc.freenode.net/), attempt 1 of 5...
jgarnett Hi (again) see subject line, there is a good old storm warning and may cut out at time ...
jmacgill looks like quite a few people are logged in but akf
jgarnett Well I imagine the udig crew is flaked out
jmacgill head now full of ideas from conference
jgarnett (I know richard and myself are on holiday)
jgarnett 1) Ideas from the conference (aka Cholmes is right)
jgarnett 2) 2.1.RC1 can someone anounce on the website
jgarnett I could ask where the IRC logs are from last week as well?
jmacgill what happend at 'meet the developers' I had to run?
jgarnett It was pretty good
jgarnett Lots of agreement on "Users please download early download often"\
jgarnett lots of storeys of 10 downloads for a milestone 100 for a rlease candidate 1000 for a release.
jgarnett (aka the open source routine is not exactly as advertised)
jgarnett I was surprised to find that the geotools / geoserver / udig sceen is one of the largest
jgarnett in terms of developers.
jgarnett And that for the work that needs to be done everyone is looking our way.
jmacgill Yeah?
jgarnett Strange for a community that is so under represented, and is not on the list for the next conference.
jgarnett Everything from the C libs not being thread safe (they hope the java bindings help kick out the bugs), to the usual need for catalog.
jgarnett I also found that talk from Apache interesting, and think we shoudl hunt down some of the code in geotools and kick it out.
jmacgill Have you read WPS yet?
jgarnett (There is some that has different (c) at the top from fed employees that needs attention)
jgarnett Not yet (on holiday) ![]()
jmacgill yeah, I know - I 'think' we can re-licencse
jmacgill as we are narrowing the scope
jgarnett I think I am going to write down a list of projects I want to work on.
jgarnett And why I think they are cool, or important.
jmacgill me too, its quite long ![]()
jmacgill yeah, I'm about to start blogging (with embeded JIRA links)
jgarnett And raster, raster ,raster, raster. Catalog and raster were the two big holes in the community right now.
jgarnett Where were you going to do your blogs james?
jmacgill codehaus
jgarnett I was thinking of blogging on Java.net
jgarnett we we tag team in a couple of places we can increse visiability.
jmacgill you know, that might be better
jmacgill so spread out you mean?
jgarnett well lets hit both, the moment udig stepped off of code haus our hits went down a lot.
jmacgill ok, I'll stay on codehaus
jgarnett Yeah.
jgarnett And we need to get cholmes going somewhere (his emails are often better blogs), I know I have lost a couple I wanted to go back and read.
jgarnett (see comment about cholmes being right) I am glad het usually attends these things for us.
jmacgill eogeo
jgarnett get him to blog on eogeo?
jmacgill yeah
jgarnett good plan.
jgarnett The other thing I want is pictures, such as the one josh drew us.
jgarnett Or that geotools diagram I drew that keeps appearing in everyones presentation.
jgarnett So James, we don't really have enough for a meeting.
jmacgill Don't think so no
jgarnett If dblasby was here I would thank him for representing the community so well
jgarnett He did a great job about answer questions of process and so on during the panel.
jmacgill Lets email, blog and chat this week to pick up some direction and motivation for the next meeting
jgarnett James can you punt the release anouncements out for 2.1.RC1 ?
jmacgill has anywhere been hit already?
jgarnett I would like to get a clear game plan going, so we can ask for help.
jgarnett The devel list, and the release is up (and on the download page).
jmacgill but not FM, java.net, SF etc
jgarnett But the anouncements have not been sent out.
jgarnett I see no evidence of it.
jmacgill ok, will do
jgarnett I think udig 1.0.0 is in the same situtation.
jgarnett (perhaps you could cross link as we did with the last release and geoserver)
jgarnett Dave's talk was very good as well.
jgarnett (did you catch any of it?)
jgarnett I initially thought it was too low level (several people left when they saw javascript at the start - several other talks were on the topic), and several more when he started carefully explaining WFS and WMS. However when he put it all together with SLD and doing things in a OGC standard way people paid attention.
jgarnett In fact one wonder why they should go to this effort to talk to GeoServer - they did not know WFS was a standard with Commerical implementations available.
jgarnett Some times I felt like we are in our own little world.
jgarnett So what was the take home message, catalog is needed, raster is needed (GDAL?), and ...
ERROR Connection to irc://freenode/ (irc://irc.freenode.net/) reset.
jgarnett .. one thing I did not find was a lot of developers.
INFO Connecting to irc://freenode/ (irc://irc.freenode.net/), attempt 1 of 5...
INFO This channel requires that you have registered and identified yourself with the network's nickname registration services (e.g. NickServ). Please see the documentation of this network's nickname registration services that should be found in the MOTD (/motd to display it).
jgarnett ping?
jmacgill lo again
jgarnett Well I don't know when i cut out ...
jmacgill WFS commercial implementations
jmacgill yeah, education and outreach seems important
jgarnett Also enthusasim, I am tempted to work through the users list after we write down the list of ideas.
jgarnett See if we can stir up some interest.
jmacgill WPS interests me as it adds the missing operations part that lets users 'do' something with the tools
jgarnett It would be fun to resolve the operations API based on that.
jgarnett (we do have a simple one in udig BTW.
jgarnett http://udig.refractions.net/docs/uDig-DataAccessGuide.pdf
jgarnett Please read - it contains the outline of the "Registgry" class you were asking about.
jgarnett More docs here: http://udig.refractions.net/confluence/display/UDIG/Documentation
jgarnett Kind of wish I printed them all out and gave you a copy when we met.
jmacgill oooh, catalog plugin looks fun
jmacgill acgill on #geotools (+n) * type /help for help
jmacgill acgill on #geotools (+n) * type /help for help
jmacgill erk
jgarnett That is the redo of the GeoServer FeatrueTypeInfo thing
jmacgill Object CatalogPlugin.open( File file )
jmacgill why does that return an Object?
jgarnett Because that is what we had in Jan.
jgarnett Let me link you to the improvement.
jgarnett However I thought you should check out the geotools (teir 0) or the Op intrefaces
jgarnett http://udig.refractions.net/confluence/display/DEV/Programmer%27s+Guide
jgarnett Links to GISPlatform, Catalog and Tracking Changes
jgarnett (ie section 4 Working with the GIS Platform)
jgarnett Is where the good stuff is.
jgarnett Man I need to update docs.
jgarnett So should we post this as the IRC log
jgarnett So should we post this as the IRC log?
jmacgill don't see why not
jgarnett Your connection has been up longer then mine, I think I will need to ask you to post.
jmacgill can't, im in a command line client on a remote box
jgarnett And lets try blogging a little while we still know something.
jmacgill lets say their was no meeting
jgarnett I will post what I got then.
Geotools 2.1.RC0 is available for download.
We are pleased to announce the welcome addition of the DB2 support to geotools.
Ongoing working on rendering issues and style support have resulted in many
subtle bug fixes and the refinement of our styling system.
This release brings us to our intended feature set for the 2.1.0 release.
One additional API change is expected, updating geotools to the forthcoming
GeoAPI 2.0 release.
The release is made in conjunction with GeoServer 1.3.0-RC1.
For more information please see the 2.1-RC0 download page.
Agenda: 1) meet sig 2) 2.1.RC0 release 3) updates to build process 4) commiter status for sig
cholmes lperpcich glad you could make it. You want to introduce yourself? We'll get started soon.
lpercich Hi all! First time in #geotools (as well as in IRC and mIRC)... hope I'll not do too many mistakes! :o)
jgarnett Meeting time!
jgarnett (see soon)
lpercich My name is Luca Sigfrido Percich (sig for friends), I'm working with GT in developing a road information system for the municipality of Milan (Italy).
jgarnett grabbing your fix james and trying again.
jgarnett Well met sig
lpercich Nice to meet you all :o)
jgarnett Usually we start with an agenda (otherwise the meetings go on too long).
jgarnett 1) Everyone meet sig ![]()
jgarnett 2) 2.1.RC0 release (tag team with GeoServer)
jgarnett 3) Updates to build process
cholmes 4) commit rights and module maintainer shit for sig on mapinfo
cholmes maintainership I mean
lpercich :oP
jmacgill ![]()
jgarnett Fair enough - either way - has sig done a code review? (ie knows how to run maven, jalopy and junit)
jmacgill power glitches here
ianT_ re: 4 - sorry for the delay on reviewing, but I'm recomending acceptance and commit rights
jmacgill lights flickering
lpercich well, just downloaded main trunk with SVN
lpercich and built with maven
ianT_ though I spent all day making my build work again so I didn't get a chance to look at today's new code
lpercich still don't know jalopy
lpercich already using Junit
cholmes You don't? Your code sure looks like it was jalopied.
jgarnett This is why we do the review
Jalopy formats the code so we don't argue about bracket placement.
lpercich maybe Eclipse auto-format?
jgarnett I need to set up one that matches Jalopy.
jgarnett Jalopy is available as an eclipse plugin. But yes the default eclipse layout is well neigh the same as what we use.
jgarnett (Different headers)
lpercich can configure it, where can I find the specs?
lpercich I mean headers, style and so on
jgarnett Developers Guide
cholmes he got the headers right...
cholmes But yes, running jalopy is the best.
lpercich Ok, I'll read it deeply
jgarnett Nifty.
lpercich I'll look for jalopy plugin, thanks
jgarnett So yeah it sounded like IanT was willing to vouch - I have not done a code review so I can take his recommendation.
cholmes http://www.geotools.org/Coding+Conventions
cholmes I did a basic code review, it looks solid.
jgarnett Well +1 for me then
ianT_ +1
cholmes ianT_ are you ok with making him co-module maintainer?
cholmes +1
jgarnett (we do voting in a semi binary way -1, +1 and 0 for don't care)
lpercich oh, thanks
cholmes and you passed - you just need 3 +1's from committers
ianT_ he can be whole maintainer if he wants ![]()
jmacgill +1
cholmes Cool - add yourself to the project.xml file as module maintainer.
lpercich thanks ian, if you don't have time for me it's ok
lpercich ok, I'll do it
jmacgill power glitched again... scary storm
cholmes And you should not have too many difficulties with maven, since there already is a module set up.
jgarnett In case you get too worried - these tools are run automagically - http://modules.geotools.org/main/checkstyle-report.html
cholmes Just kill the old DataStore.
jgarnett (we are not always doing that well ourselves)
lpercich ok
lpercich you mean, like a cvs remove?
jgarnett I was going to point out "clover" but I dont see it
cholmes In Jody's words (from the comments) 'This is the worst possilble datastore you can imagine...'
jgarnett (it is the tool that tells you how much of your code your unit tests cover)
jgarnett James what happened to clover?
cholmes No, it has the same name as yours, so yours will just be the next revision
lpercich fine then
jmacgill used a lot of memory - versions changed - ran out of personal time to keep it up and running
lpercich ooops... I would call it MIFDataStore, 'cause it uses the MIF format
cholmes But you can get rid of the DataSource when you do that - cvs remove.
lpercich what about keeping MapInfoDataStore for a future support of native format?
cholmes Oh, ok, that's fine. In that case you can cvs remove the old one.
cholmes That sounds good - remove it for now, and then if we do the native format we can bring it back.
lpercich Fine, I'll stick with MIF* then.
ianT_ sounds good to mee
cholmes jgarnett - what's the procedure for svn access - still send to pramsey?
jgarnett yep
jgarnett So welcome sig ![]()
jgarnett Hi justin just getting into the swing of things ( I think you forgot some Postgis commits)
jgarnett 2) 2.1.RC0 release - real soon now
jgarnett The build is currently annoying me with images, it has the patches dblasby wants for geoserver.
jdeolive_ i am in the process of merging my postgis changes onto 2.1.x
jgarnett An announcement email will be sent out as part of the geoserver release.
lpercich Thanks jody!
jgarnett Is there anymore questions about 2.1.RC0 - the Jira release notes are fairly clear about what changed.
jmacgill given that it has RC in its name, are we sure there will be no more API changes?
jgarnett Good point, there are two on the books.
jgarnett Martin will update us to the stable GeoAPI later this week (3-4 changes of Array to Collection)
jgarnett And Dzwiers would like to upgrade XDO - but lacks volunteer time himself, and documentation for others.
jgarnett 2.1.RC1 will go out with udig next week.
cholmes Are you doing udig 1.0.0 ?
jgarnett Yes.
dblasby meeting?
dblasby meeting?
jgarnett Is on right now?
jgarnett 3) Updates to build process
jgarnett James want to tell us what is new and improved?
jmacgill Sure...
jmacgill ok, for a while now the auto Cruise Control build has been using the MultiProject plugin to control the build
jmacgill this is far less ad-hock than the old system
jmacgill I have now extended that to all builds (so expect tonights nightly release to be different too)
jmacgill the major chance is in how the list of things to build is created
jmacgill specificaly excluded modules will NO LONGER build built in a normal build (or nightly or cc)
cholmes Could you possibly change the email subject a bit? In my email client all I see is F... which could mean fixed or failed.
jmacgill or 'createRelease' eather'
jmacgill sure
CIA-10 dadler * r14135 geotools/gt/module/sample-data/src/org/geotools/sampleData/ (6 files): GEOT-541 add .prj for shapefiles
jmacgill the main result is that the folowing modules will not build by default....
jmacgill pickle, db2, gtopo30, mapinfo and geometryless
CIA-10 dadler 2.1.x * r14136 geotools/module/sample-data/src/org/geotools/sampleData/ (6 files): GEOT-541 add .prj for shapefiles
jmacgill I saw an email re geometryless today that may get it off the excludes list and I guess mapinfo is about to die
cholmes db2 should be good to go, email today
jmacgill module maintainers are in control they should add/remove themselves from the maven.multiproject.excludes list
jgarnett Quick annoying interruption:
jgarnett zip Building zip: C:\java\geotools\gt\target\distributions\geotools-bin-2.1.RC0.zip
zip Building zip: C:\java\geotools\gt\target\distributions\geotools-src-2.1.RC0.zip
BUILD SUCCESSFUL
jgarnett Please resume.
jmacgill no javadoc?
jgarnett It made a javadoc.
jmacgill if a module braks the cc build for too long I will add it to the excludes list (and thus it will not be in the nightly or any releases)
jmacgill you can find the magic line in build.properties file at the top level
jmacgill they look like this:
jmacgill maven.multiproject.includes=module//project.xml,ext//project.xml,plugin//project.xml,demo//project.xml
jmacgill maven.multiproject.excludes=*/pickle/,/db2/,/gtopo30/,/mapinfo/,/geometryless/*
jmacgill so new modules will be picked up automaticaly unless they are added to the exclude list explicitly
jgarnett James is mockrunner needed?
jmacgill no idea
jmacgill my guess would be no
jgarnett I would like to trim our download size at some point.
jmacgill batik is used for SVG symbol support in the renderers
jgarnett Uploading now - if I can ask:
jgarnett 1) dblasby to try this with geoserver
jgarnett 2) anyone to try a build from the src download
cholmes darn, my dateTime fix won't be in there...
jgarnett 3) And I will try it with udig
jgarnett What is your dateTime fix chris?
jmacgill anyway, thats the upshot of the new build system (use it the same way as before)
jmacgill any questions?
cholmes So temporal attributes toString returns a better formatting.
cholmes I made the fix between when you notified of the branch and when you sent the email.
cholmes And now my svn is f'ed.
cholmes nothing will check out, anywhere.
jgarnett And I can run the script again, dblasby is the person we are slowing up.
lpercich So as soon as I'll put MIFDatastore on main svn trunk it will be built up? Should I work on a separate branch?
jgarnett Do you want to email me the fix?
cholmes Sure.
cholmes Or you can probably just do a diff against the feature.type package.
jgarnett jdeolive (or jdeolive_) ? I know you were working on getting maven to build eclipse .project and .classpath
dblasby I just build jars from my checkout. My checkout is 2.1.x so I'm just going to be using them.
jmacgill as long as it builds it can be on trunk
jgarnett (as a replacement to the gt2eclipse script)
cholmes New DateUtils class, and modifications to the temporalAttributeType
jmacgill if it does not build it can still be on trunk provied you put an entry against the excludes list
jgarnett what revision was you checkout dave?
jdeolive_ yeah, maven has a built in plugin for eclipse
jgarnett (can find out with svn info)
jdeolive_ it builds .classpath and .project files for each module
cholmes lpercich - trunk is not the stable branch, so if you want it in 2.1.x we should talk about getting it in. You'll have to commit it there too.
cholmes Better to make sure all is well on trunk before considering it...
cholmes It's not an api change, so it should be fine.
jdeolive_ at one point i modified the root maven.xml and added a reactor goal to generate them for all modules
jdeolive_ with the new multi project stuff, i am not sure how that changes
cholmes It would be very nice to have that. In the past few weeks I've had some good success tempting people to make gt2 fixes, but then they get so discouraged by trying for half a day to get the source to build...
lpercich I see. Probably it could work in 2.1.x, just tell me what you think is better
cholmes start on trunk.
lpercich I'll do.
cholmes then we'll see. I mean, if it was on 2.1.x then it'd be in both - it's just a question of if we get it in the 2.1.0 release or in 2.2.0
jgarnett James this is a big enough issue that I am tempted to commit .classpath and .project files for people. I have spent several days is support over the last month.
jmacgill how are the files generated?
jmacgill by hand?
jgarnett I have attached .classpath and .project files to the "Eclipse Developers Guide" page. These were only generated with gt2eclipse.
jmacgill or a script?
jgarnett My bad script.
jgarnett I would much rather have a decent maven aware script.
cholmes But we'd better not have too many api changes to 2.2.0 that one can't easily use datastores from 2.2.0 on 2.1.0, or I will be very upset. Which will mean udig and geoserver should be able to use mapinfo as a plug-in
cholmes I'm a big +1 on committing .classpath and .project files
cholmes And I don't even use them...
jgarnett Well I suspect FeatureType will need a pretty big smack.
jmacgill where would they live?
jgarnett depends what you want to do.
cholmes Shouldn't affect the pluggability of datastore.
cholmes Unlike that freaking URI change.
jgarnett 1) a single .classpath and .project file in the top level "gt" folder (no respect for module dependencies)
jgarnett 2) .classpath and .project files in each module
jgarnett #1 I can generate via brute force and the gt2eclipse script
jgarnett #2 Justin could generate using maven
jgarnett Maven is still required to download everything into the maven repository in either solution (and to run testcases)
jgarnett Please try out http://lists.refractions.net/geotools/gt2
jgarnett james something is wrong
jgarnett the bin download is 23.9 MB
CIA-10 jdeolive 2.1.x * r14137 geotools/plugin/postgis/ (4 files in 2 dirs): migrated postgis changes to 2.1.x branch
jmacgill bound to be glitches in a new build system ![]()
jgarnett We are including xalan-2.5.1.jar multiple times in the bin download.
jgarnett I will remove duplicates by hand and upload.
jmacgill odd its in the shared folder
jmacgill share test must have broken down
cholmes sweet Geoff Adams said my oracle fix worked.
jgarnett nice!
jmacgill cool
jmacgill dam
jgarnett james justin was telling me that he needs you to rename build.properties as project.properties or something.
jmacgill shared jar test is screwed
jmacgill almost all jars are in there twice
jmacgill ok, will do the rename v.soon
cholmes jdeolive - what changes did you end up doing to postgis?
jgarnett Apparently he had to leave.
jgarnett I think you asked him not optimize the getBounds
jgarnett So the work probably only exists in a udig subclass of PostgisDataStore ...
jgarnett .. I would suspect as an additional method.
jgarnett So even after a diet the bin download is bigger then I remember.
jgarnett Did we forget to turn on compression james?
cholmes Jody, just sent you an email, with two fixes...
cholmes anyone have any clue why my svn is messed?
jgarnett And they need to go on trunk and on 2.1.x?
cholmes I get 400 Bad request (http://svn.geotools.org)
jgarnett No I don't know why you svn is messed. svn cleanup did not work?
cholmes one needs to go on both, one is on trunk, you just need to move it over.
cholmes details in the email.
cholmes no, cleanup didn't work.
jgarnett oh perhaps, when did it stop working for you? You may have mapped it explicitly in your hosts file back when the domain registration process was causing us all grief.
cholmes I'll check...I think I did fix that, though maybe I just changed it.
cholmes Where is my hosts file again?
cholmes Hrm, I def. changed back because I couldn't even see it in a web browser. Now I can see it from a browser, but I can't svn anything
jgarnett C:\WINDOWS\system32\drivers\etc
jgarnett Okay the new improved 15 mb download is up.
cholmes nope, it's clean ![]()
jgarnett I think the increase in size may just be due to new modules (hsql and friends)
jgarnett james should I wait for a patch for the shared jars thing?
jgarnett I assume you have tried a fresh checkout in a empty directory already.
jgarnett That was the last agenda item btw - thanks for the meeting everyone.
jgarnett I got your email. I will apply the patches after 2.1.RC0 is out. I didn't want to give dblasby anything else to worry about.
jgarnett Posting logs.
Version 6.7 of the EPSG geodetic parameter dataset has been released and may be downloaded without charge from www.epsg.org. This release contains a number of corrections to data in the previous v6.6 (October 2004) release together with additional geodetic data for Antarctica, Australia, Brazil, Colombia, Finland, Libya, Netherlands, Nigeria, Philippines, Syria and the UK Channel Islands. Additional cross references to data obtained from EuroGeographics and the US NGA Technical Report 8350.2 have also been included. As in previous releases the v6.7 dataset is available in two forms: an MS Access database which also includes a data reporting capability, and as SQL scripts for populating an alternative relational database application. The v6.7 release is accompanied by updated versions of Guidance Notes 7 part 1 and 7 part 2.
Please note that since the previous dataset release, the European Petroleum Survey Group has reformed as the Surveying and Positiong Committee of the International Association of Oil and Gas Producers (OGP). The EPSG Geodesy Working Group has reformed as the Geodesy Subcommittee of the OGP Surveying and Positioning Committee. This Geodesy Subcommittee continues with the same members as the previous EPSG Geodesy Working Group. This OGP Geodesy Subcommittee will continue to maintain the dataset. The dataset will continue to be known as the EPSG (geodetic parameter) dataset.