Agenda: 0) what is up 1) OSGeo 2) FactoryFinder Proposal 3) 2.4.0 release delayed 4) wfs update 5) shapefile update

jgarnett has changed the topic to: 0) what is up
jgarnett: okay floor is open for topics.
desruisseaux: OSGEO
jgarnett: I am hoping for an update after last Friday's OSGeo board meeting.
jgarnett has changed the topic to: 0) what is up 1) OSGeo
aaime [n=aaime@host218-168-dynamic.54-82-r.retail.telecomitalia.it] entered the room.
desruisseaux: 2) The FactoryFinder incompatible change proposal
jgarnett has changed the topic to: 0) what is up 1) OSGeo 2) FactoryFinder
jgarnett: Martin do we have a proposal page for that one yet?
jgarnett: Or you wanted to talk and then write one up...
aaime has changed the topic to: 0) what is up 1) OSGeo 2) FactoryFinder 3) 2.4.0 release probably delayed
desruisseaux: No. I will write one.
desruisseaux: (looking for the URL...)
jgarnett: Child page of http://docs.codehaus.org/display/GEOTOOLS/Proposals and there is a template to get you going.
jgarnett: okay looks like it is time to start ...
jgarnett: 0) what is up
desruisseaux: Thanks!
desruisseaux: Martin: splitted between IFREMER and some coverage stuff (looking for something that could help an issue I have)
jgarnett: jgarnett; looking at aggregate functions on trunk will need to write a bunch of test cases to get this stuff going; applying a customer supplied partch to ArcSDE and learning PHP
***groldan : working on WFS 1.1.0 client gt<->udig
***aaime still working on wcs 1.1.1, ready to fix some more gs 1.6.0-rc issues
jgarnett: a ha - so that is why wfs is not building for me!
groldan: uh? gt or udig?
jgarnett: gt
jgarnett: Jesse is still doing QA on his shapefile rewrite as far as I know.
***jdeolive is doing some odds and ends for ows5, playing with h2 datastore
jgarnett: nifty ... moving on.
jgarnett: 1) OSGeo
jgarnett: Anyone know how the board meeting went last friday?
groldan: jgarnett: svn updating and building, don't want to say just "it works for me"
jdeolive: well two build servers say it is fine also
desruisseaux: Vincent saw on an IRC Chris said that the proposed agreement has been accepted by OSGEO
jdeolive: one windows one linux
jgarnett: okay; doing the rebuild dance now.
jgarnett: So the proposed agreement has been accepted
groldan: also note I got rid of the wfs datastore and factory subclasses in udig
jgarnett: what do we need on our end; to vote on a proposal to change our headers to (c) OSGeo ?
jgarnett has changed the topic to: 0) what is up 1) OSGeo 2) FactoryFinder 3) 2.4.0 release probably delayed 4) wfs update
jdeolive: btw, we will soon have a build server on windows publicly visible and complaining to the list when we break the build on windows
jgarnett: So can we use this page: http://docs.codehaus.org/display/GEOTOOLS/Gradudate+from+OSGeo
jgarnett: we will need to figure out what exactly the new headers look like.
aaime: jgarnett, this seems good news to me
jgarnett: aaime++
jgarnett: cholmes ping? do you have details?
aaime: jgarnett, is there a log of the meetings?
jgarnett: (anyone know if cholmes is live?)
jgarnett: There should be; I am searching now.
cholmes: I'm here
cholmes: details on what?
jgarnett: 1) OSGeo
cholmes: I sent an email to geotools-admin
cholmes: But I guess it didn't go through
desruisseaux: Vincent told me that he saw on an IRC Chris saying that the proposed agreement has been accepted by OSGEO
cholmes: But yeah, proposed contributor agreement was accepted
aaime: cholmes, no, I did not see the mail on admin
arneke [n=ak@cpe-66-108-80-238.nyc.res.rr.com] entered the room.
jgarnett: so we are getting excited for nothing? What was the objection to the proposed agreement cholmes?
cholmes: No, my email didn't go through
cholmes: the agreement was accepted
jgarnett: http://wiki.osgeo.org/index.php/Thirty_Second_Board_Meeting
cholmes: OSGeo has approved the contributor agreement proposed by Adrian et al
jgarnett: sweet; nice work.
cholmes: So now we have to on the business of getting people to sign it and send it in
cholmes: and figure out who can act as a representative of the foundation
cholmes: to sign the foundation part
jgarnett: So I better write up or proposal page with "Option One"; and put in the steps for signing and mailing.
aaime: so?
aaime: do we wait for the legal part to be completed and then we switch all of the headers?
jgarnett: I will update the page here; and then we can vote on it?
jgarnett: I think we could do both at once aaime; they each will take a while to do.
aaime: ah, right, we cannot mass update the headers...
aaime: (or else, we can pick the common ones and update them)
aaime: (and then manually update the funny ones)
jgarnett: actuall we need to update them by hand; and that work will complete our providence review.
aaime: Jody, for the ones that have been copied and pasted why not mass update?
jgarnett: It has been two years since I did it last; and the library has grown since then. At least this time we will not have as many questions.
jgarnett: please explain?
jgarnett: do a global search and replace?
aaime: yes?
aaime: at least for those headers that are really the "standard" one we use now
aaime: we have to manually adjust the remaining ones
jgarnett: The goal is to review each file; doing a global search and replace would not accomplish that. And most files are not boiler plate; they have a list of (c) holders in them by now (often using svn history)
aaime: In the header?
aaime: what I'm supposed to do when I look at a file?
aaime: throw it into google and see if there is copied code?
jgarnett: You are supposed to (for the files in module you maintain; or since you are in PMC for files you are cleaning up for others)
jgarnett: look at the header
jgarnett: do a sanity check that the (c) history is correct
jgarnett: update the header; adding (c) OSGeo to the top
aaime: Ok, what I'm saying is this... I've just opened 20 files in a row
jgarnett: Update the review.txt file with any questions; say for david zwier code.
jgarnett: commit.
aaime: in 19 out of 20 I've seen the same header (letter by letter)
jgarnett: understood; that either means a) it is a mistake you need to fix or b) the 19 files were all made at the same time.
aaime: then I need rules to understand how the history of copyright is to be interpreted
jgarnett: I don't mind how you update them; as long as you have looked at each of the 20 files.
desruisseaux: I guess we can split the work. I can take care of referencing, coverage and some unsupported/plugins.
desruisseaux: (and metadata too)
jgarnett: okay; perhaps the best way to do this would be a breakout IRC session? I am not trying to make this hard; but I am also not trying to make this easy andrea. Them's the breaks.
aaime: I'm just asking for guidance on how to update the headers
jgarnett: The review.txt files should offer a good guide for trouble spots; remember we don't need to fix it all; just know what is broken.
aaime: since evidently I don't understand the copyright history
jgarnett: understood; we can look at a couple files together. It is pretty easy. Often you find the (c) information dating back to james or martins projects, or udig; we need to maintain that.
desruisseaux: Can we do that for trunk only, or do we need to do the same for 2.3 and 2.4 branches? (would be very long...)
jgarnett: Example: this file lacks a header
jgarnett: http://svn.geotools.org/geotools/branches/2.4.x/demo/example/src/main/java/org/geotools/demo/ShapeLab.java
jgarnett: trunk only
aaime: No Jody, not tothether... can you write the rules for everyone participatin to apply?
aaime: or do you foresee that we'll update the files just you and me?
jgarnett: This file came from udig: http://svn.geotools.org/geotools/branches/2.4.x/modules/library/api/src/main/java/org/geotools/catalog/Catalog.java
jgarnett: we all will; but we are looking for a sanity check more than rules.
aaime: this is the tipical header is saw in a ton of files: http://svn.geotools.org/geotools/branches/2.4.x/modules/library/api/src/main/java/org/geotools/data/ALLQuery.java
jgarnett: ie read what is there; add a line for (c) OSGeo and move on.
aaime: grrr... I need to know what is sane
aaime: is that one sane?
aaime: 2007 is not covered
jgarnett: It is sane except for three things
jgarnett: (c) 2008 OSGeo
jgarnett: (c) 2005-2007 GeoTools Project Management Committee
jgarnett: (is that spelled correctly?)
aaime: see? not sane, but I would haven't noticed
aaime: that's why I'd like some simple rules
jgarnett: And the rest looks okay; however the LGPL boiler plate has been changed on us - we should check again.
aaime: hmm... wait a second, we're not switching to LGPL 3 right?
desruisseaux: But I though that we would remove the old copyright notice (at least those who accept to sign the agreement) since we are giving the copyright to OSGEO...
jgarnett: And I think we are doing that for the year 2008 are we not?
desruisseaux: (typo: "would remove the old copyright holders...)
jgarnett: question aaime: did you look at the example of the proposal page? That should cover the rules ...
jgarnett: understood; I think we need to keep the old (c) holders in the mix
jgarnett: and for some cases; like david zwiers we will not assign to OSGeo; just have his name and the LGPL stuff. Although for a lot of that work a signature from his employer will do the trick
arneke left the room (quit: Remote closed the connection).
jgarnett: Is that okay martin?
jgarnett: Andrea can we work out some rules you are happy with and add them to that page; when we are done we can ask for a vote.
jgarnett: 2) Factory Finder
jgarnett: Martin this one is all yours.
desruisseaux: I don't have a strong feeling, but (I may surprise peoples ) sometime I like simple thing, and I though that a single OSGEO copyright holder was simple if all other (C) holders accepted to sign the assignment.
aaime: jganett, these rules shoudl cover some examples of Zwier code
jgarnett: okay will add that.
desruisseaux: FactoryFinder
desruisseaux: Copied my email there:
desruisseaux: http://docs.codehaus.org/display/GEOTOOLS/ReferencingFactoryFinder+incompatible+change
desruisseaux: Note that it is basically what Jody suggested one or two years ago (I don't remember)
aaime: ah
desruisseaux: I have only two questions:
desruisseaux: 1) Any objections?
aaime: nice we have some delayed agreement
desruisseaux: 2) Should we do the same for other FactoryFinder?
aaime: Not from me... the current situation is killing me (can't unit test wcs with lon/lat orientation)
aaime: Hmmm... in fact I had a question spinning in the back of my mind...
aaime: could'nt the same problem be solved by chaning the interpretation of the hints?
aaime: Now it seems the orientatio hints are taken as an order
aaime: "gimme lon/lat or die!"
aaime: what about treating them as... hints insetad?
aaime: "gimme lon/lat if you can"
jgarnett: Hi martin; I sent and email this morning outlining my origional thought (from a couple years ago). Did you get a chance to read yet?
jgarnett: the difference is placing the hints in the constructor; not the indvidual calls to create.
desruisseaux: It is kind of... For a given hint ("give me lon/lat") a factory can have 3 answers: 1) "I'm lon/lat so I'm fine", 2) "I'm lat/long so I die", or 3) "I don't care about lon/lat so take me as I'm"...
jgarnett: I also had two ideas on how you could have your strict functionality.
desruisseaux: Yes Jody I have read your email and replied to it.
desruisseaux: It seems to me that the ideas are very close together
desruisseaux: The main difference between your proposal and mine was that I didn't suggested to make the FactoryFinder(Hints) constructor public, but I have no strong objection if you wish.
aaime: Well, with my work what I would really like is just a way to say: EPSG facroy and http factory -> lon/lat, urn factory -> lat/lon
aaime: in a way that makes CRS methods respect it
aaime: since all of the gt2 code (minus a few modules) are really doing their business using the CRS object
aaime: hello?
desruisseaux: I will wait a few days in case peoples want to comment on the above-cited page or propose alternative.
desruisseaux: If nobody object and suggest ans other approach, I will try to apply it by the end of this week.
aaime: I'm fine with that
jgarnett: I also
aaime: which brings me to the next topic
desruisseaux: One of my question was "do we apply the same approach on other FactoryFinder". I saw no comment on this one, so could it means "peoples have no opinion so do it if you wish"?
jgarnett: Lets write up a code example of each technique on the email list.
desruisseaux: All right Jody
jgarnett: I like the idea of Hints.isStrict() so we dont have to mess with existing factory finders.
jgarnett: aaime; we should hit the next topic.
jgarnett: 3) 2.4.0 release probably delayed
jgarnett: Is it being delayed over this issue? Or is there something more...
aaime: Nope, it's delayed because we most probably need to delay the geoserver release
aaime: and if we release 2.4.0 this week we may end up releasing 2.4.1 the week after
aaime: (if just one critical bug pops in)
aaime: we already did this mistake with gt 2.3.x something
aaime: two minor releases in one week because we did release gt2 before gs
desruisseaux: A delay is fine for me.
aaime: Yeah, I was about to say, given this change proposal, releasing this week would be bad for you too
desruisseaux: True
jgarnett: sounds fine
jgarnett: moving on.
jgarnett: 4) wfs update
jgarnett: Hi Gabriel what is up
groldan: hey
groldan: so I'm doing wfs 1.1 (read only) support
groldan: extending the wfs plugin
groldan: got a common interface and separate datastore implementations for 1.0 and 1.1
groldan: the interface exposes some service metadata so udig don't need to subclass
groldan: the long term plan is to get them sharing more code where appropriate
groldan: so I decided to extend the wfs plugin instead of cooking a new one
groldan: yet so far both are based on different xml handling tech
groldan: so that more or less is it
groldan: in the end it should be a lot easier to maintain
groldan: and I would like to be at least a co-maintainer
groldan: but
groldan: here comes an actual question
groldan: how do you feel about getting emf dependencies in geotools?
groldan: afaik justin were trying to keep them as isolated as possible
groldan: but real use of gtxml<-->emf carries out emf on plugins
jdeolive: yeah that was another part that made me originally want to do a seperate module
groldan: yet I'm not sure why that'd be a problem=?
groldan: (among the extra sutff to download)
jdeolive: that was the problem for people
jdeolive: its about an extra 1.1M
groldan: do they complain about the size of google earth?
jdeolive: or maybe 1.5... not sure
jgarnett: we aquired a xerces dependency in much the same way
jgarnett: anyone know if we are still needing that one?
jdeolive: yes
jdeolive: gtxml uses xerces as well
groldan: well, yes that's something that would need to be fixed imho
jgarnett: darn; I was hoping to have one or the other (as long as the point of them is to represent the schema?)
jdeolive: groldan: what do you mean?
groldan: the dependency on xerces
jdeolive: well with java5 we might be able to kill it
groldan: I guess its just because of the use of xmlserializer (or so)?
jdeolive: nope, there is validation stuff too
jdeolive: but those are the two reasons
groldan: validation like in schema validation?
jdeolive: yeah... i could never get schema validation to work with another impl other than xerces
groldan: could those two things be done using pure api anyways?
jgarnett: I am going to have to go soon guys; can someone post the logs please.
jdeolive: so that might just be a problem with my "know how"
groldan: jgarnett: I can
groldan: jdeolive: bah, neither me would know how to do that off the top of my head... but anyways
jdeolive: groldan: with java 5 i think its possible... but with java 4 no
groldan: cool, we're java5 now
jdeolive: yup... i just need to try it out
jdeolive: java5 validation is much different
groldan: still we didn't heard opinions about carring out emf dependencies?
jdeolive: aaime was against it
groldan: is the concern just the extra download size?
aaime: both the download size and a personal distaste for everything that carries an Eclipse stigma (api wise)
aaime: but that's just me
groldan: ah, note taken
Jesse_Eichar__: 5) Shapefile
groldan: so what do we do? provided there's quite a lot of stuff already commited to emf
Jesse_Eichar__: I'm now ready to commit.
jdeolive: groldan: that was the point of the seperate module
jdeolive: keep it seperate and an optional thing for people
jdeolive: but i admit
jdeolive: it was a bandaid
groldan: yeah
aaime: well, with gtxml emf is going everywhere xml is so...
aaime: I guess trying to keep it away is a lost cause
jdeolive: not necessarily
jdeolive: i mean gtxml is not used by any core api
jdeolive: and as long as only datastores use it there should be no problems with it creeping in
jdeolive: all the core stuff is over in extension
jgarnett: back
jgarnett: Jesse_Eichar++
jgarnett: cool
jgarnett: shall we move on?
jgarnett: 5) Shapefile
jdeolive: `i dont think 4 is finished yet
jgarnett: doh
jdeolive: but i guess we are on overtime now
jgarnett: are we ready for jesse to be ready to commit?
jgarnett: I have not reviewed feature event handling yet jesse; want to do that when we debug in uDig?
Jesse_Eichar__: I haven't changed anything about the functionality for this release
Jesse_Eichar__: just changed how locking is handled.
Jesse_Eichar__: release === commit
Jesse_Eichar__: Things like events and alternative geometry handlers will come later
aaime: well, trunk is trunk
aaime: I guess you can commit and have someone review it
aaime: and curse if things stop wroking all of a sudden?
aaime: :=
Jesse_Eichar__: ok. Maybe I'll see if Jody has a second before I actually commit. I've got uDig working with it. So that's some indication that it works. Editing zooming panning selection.
CIA-21: jdeolive * r28747 geotools/gt/modules/extension/xsd/wfs/src/test/java/org/geotools/wfs/WFSParsingTest.java: added some tests for testing online resource urls
aaime: wow, looks promising then
jgarnett: excited
jgarnett: (aside: http://docs.codehaus.org/display/GEOTOOLS/Gradudate+from+OSGeo updated with different examples)
jgarnett: going over @author dzwiers showed nothing that was not as a refractions employee; so the usual update header rules apply
jgarnett: cool
jgarnett: thanks for the meeting everyone.
desruisseaux: thanks for driving it