Weekly IRC meeting: 5 may 2008
0) what is up
1) svn cleanup
2) 2.5-M2
3) inccubation
4) FeatureCollection aggregated functions
5) backport paging to 2.4.x

<jgarnett> okay lets go ... although I had hoped to hear from simboss...
<jgarnett> 0) what is up
ggesquiere (n=gesquier@arl13-3-88-169-136-131.fbx.proxad.net) has joined #geotools
<acuster> acuster — cleaning svn; looking into mercurial
<desruisseaux> Martin: MosaicImageReader and postgrid.
<jgarnett> jgarnett - looking at Process discussion from last week and updating the codebase, trying to rememeber what was going on w/ ArcSDE now that I can run tests, annoying Eclesia with geoapi feedback when it sounds like he is on a deadline.
<Eclesia> jsorel : removed some swing widget that will not be supported in the futur
<groldan> groldan: adding tests to geoserver wms, quite stuck on static stuff
<acuster> no deadline, merely a distinct urge to make progress
<acuster> anyone else?
<jgarnett> simboss ping
<jgarnett> 1) svn cleanup
<jgarnett> acuster and/or martin ?
<acuster> afabiani, Awp_ chorner dwins elizard ggesquiere hbullen jdeolive mgrant ?
aaime (n=aaime@host97-45-dynamic.3-87-r.retail.telecomitalia.it) has joined #geotools
<acuster> me
<afabiani> hi
<Awp_> it wasn't me
aaime has changed the topic to: 0) what is up 1) svn cleanup 2) 2.5-M2 3) inccubation 4) FeatureCollection aggregated functions
<acuster> any of you want to give us a quick what is up?
<acuster> going once...
<acuster> okay gone
<acuster> SVN cleanup
<acuster> hmm, I even wrote up a page on what I'm doing
<acuster> in summary
<acuster> we start with the repo, use svnadmin to get a dump
<acuster> run that through svndumpfilter a bunch of times
<acuster> that gets rid of 1) most of udig 2) big files 3) lots of stuff we don't care about, e.g. .cvsignore files
<acuster> we go from 3.0GB to 1.4GB or so
<jgarnett> (link to page?)
<acuster> then we run the dump through a java class to pick up all the files which were duplicates
<acuster> http://docs.codehaus.org/display/GEOTOOLS/Svn+Cleanup+2008
<acuster> all the files which were added as duplicates
<acuster> only we don't do anything if they were added in the same commit
<acuster> because fixing that would be hard and error prone
<acuster> anyhow,
<acuster> we are good to go
<acuster> we need from you all (1) a go ahead (2) a date or date range where we can do this work
<acuster> how does this week work for everyone?
<desruisseaux> Fine for me.
<aaime> yap
<groldan> for how long it will mean having no svn?
<jgarnett> how does it work with refractions sys admin?
<acuster> groldan, I'm guessing 24 hours probably less
<groldan> cool
<acuster> jgarnett, I need to clear it with them when I get a sense from the gt community what works here
<acuster> I have no idea of GS or uDig deadlines in the near future
<acuster> also I need to coordinate with you jody to work out the permission stuff
<acuster> okay, people seem willing to do it. I'll contact refractions and get a date from them, then send out a confirm email
<acuster> if it causes a time conflict for anyone at that point, yell loudly
<aaime> Hem... deadlines... groldan, what was the planned release date for gs 1.6.4?
<acuster> I'd like to do it wed/thrusday
<aaime> end of this week, or end of next week?
<groldan> I'm not really sure, guess end of week?
<groldan> this one I guess
<aaime> acuster, are you planning to do the work during the week or during weekend?
<acuster> week
<acuster> since it depends on refractions
<aaime> Hmm.... ok, then we cannot cut 2.4.3 on Friday I guess
<jgarnett> acuster I am away for the weekend+monday; refractions sys admin can also do permissions if needed.
<acuster> ok
<acuster> aaime, groldan is this week a bad idea for you?
<acuster> is next week better?
<groldan> hmmm I thought I could afford a day without gt svn, not sure about andrea
<aaime> ah, me too
<aaime> the problem is the timed gs release
<aaime> next week is probably going to work better for us, yes
<aaime> I'm asking our release manager just to make sure
<aaime> (crazy times at gs)
<groldan> acuster: waiting till next week would be killer for you?
<acuster> okay, I will now aim to take svn down for 24 hours the 13,14, or 15th
<acuster> nope
<acuster> glad to work around your schedule
<acuster> that's all, next.
<jgarnett> 2) 2.5-M2
<groldan> that's reliefing so, thanks
<jgarnett> I tried to release this last week; and got stuck on some work simboss was doing; I would like to try again this week ...
<jgarnett> is there anything else that is going to hold me up?
<groldan> yes
<jgarnett> (the goal here is to make a milestone release so some uDig developers can evaulate switching to trunk...)
<groldan> I need to add unit tests for the paging + querycapabilities stuff
<jgarnett> okay; can you ping me when done?
<groldan> planning to do that tomorrow though
<acuster> why does that block a release?
<jgarnett> I don't really mind why (my guess is gabriel does not trust the code until unit tests are in place?)
<groldan> not sure if it should, because there might be a couple regressions/new bugs we don't know about
<acuster> if jody can wait, it's all good
<acuster> the uDig folk are chomping at the bit though
<jgarnett> It is only a milestone release; if it "works" then it is worthwhile me making "SDK" releases available for uDig developers (so they can migrate to trunk)
<groldan> anyway, if jody is planning to do it this week I can put a hardline myself of tomorrow
<acuster> great
<jgarnett> moving on ...
<groldan> okay, I would preffer you wait for it so
<groldan> since udig uses a lot of postgis
<groldan> which's what I touched
<jgarnett> understood; also that functionality would make a great tableview :-P
<jgarnett> 3) incubbation
<jgarnett> no progress to report; I see some (c) headers have been updated ...
<acuster> do we need to schedule a big push for that?
<jgarnett> this is really in the hands of module maintainers right now....
<jgarnett> we do
<jgarnett> a search and replace would be a good start.
<acuster> that sound scarily lazy
<acuster> desruisseaux, how are your modules on the review.txt front?
<acuster> anyone else know where they stand on their modules?
groldan has changed the topic to: 0) what is up 1) svn cleanup 2) 2.5-M2 3) inccubation 4) FeatureCollection aggregated functions 5) backport paging to 2.4.x
<desruisseaux> I don't think that anything changed since the review.txt files has been wrote.
<desruisseaux> (I means, it seems to me that the pool of developper in metadata, referencing and coverage has not changed)
<acuster> which means what? are you ready for graduation?
<desruisseaux> If graduation == changing headers, yes.
<acuster> that's pretty much all that left
<jgarnett> graduation == changing headers & review of code
<acuster> we can smell the OSGeo on the horizon
<acuster> anyone else?
<acuster> can we give ourself a deadline?
<acuster> ourselves
<jgarnett> I have not looked at any of my modules recently; a deadline would be fine.
<acuster> end of the month?
<acuster> June 21st (i.e. summer)?
<desruisseaux> Maybe: target end of the month, and see at that point where we are?
<jgarnett> let's go for summer; the smart thing would be to look when the next OSGeo meeting is .... and plan to finish our requirements two weeks previously; to give the iccubation committee a chance to review our work.
ggesquiere_ (n=gilles@arl13-3-88-169-136-131.fbx.proxad.net) has joined #geotools
<jgarnett> Let's try and write up our incubation stuff up at the end of the month; it will be a good test of where we are.
<acuster> good
<jgarnett> 4) FeatureCollection aggregation functions
<jgarnett> aaime ?
<aaime> Yes
<aaime> summary: feature collection has some aggregate functions
<aaime> like disticnt/min/max/average and so on
<aaime> they are implemented as fc visitors on the standard fc
<aaime> but for the jdbc fc, no, they are encoded as sql
<aaime> unfortunately that encoding is done once for all db in a way that breaks postgis (among other)
<aaime> and that breaks if the expression passed as an argument is not a simple PropertyName
<aaime> I need to fix it
<jgarnett> thinking ...
<jgarnett> can the different datastores; list PostGIS
ggesquier (n=gesquier@arl13-3-88-169-136-131.fbx.proxad.net) has joined #geotools
<jgarnett> implement their own FC? I think PostGIS already does ...
<aaime> it's empty and unused
<aaime> (but yes, it's there)
<aaime> moreover, that would not solve the problem of encoding an expression like "attribute + 10"
<jgarnett> ah; it origionally was the one that had the visitor stuff encoded as SQL
<jgarnett> and it was pulled up and shared.
<aaime> My point is that we can keep it shared
<jgarnett> okay
<aaime> we alrady have factored out ds differences in SqlBuilder/FilterToSql
<aaime> but sqlbuilder does not expose that functinality in a useful way
ggesquiere_ has quit (Client Quit)
<aaime> I would just need a encode(Expression) method in SQLBuilder and I could implement it the right way
<aaime> so that the col names are escaped according to the db needs and
<aaime> so that complex expressions are encoded properly as well
<aaime> but to do so, I need to add a method to an interface
<desruisseaux> I need to go before the rain... Bye all!
desruisseaux has quit ("ChatZilla 0.9.81 [Firefox 2.0.0.14/2008042015]")
<aaime> Now, SQLBuilder is one of those interfaces that are public
<aaime> but not really "published"
<aaime> it's not like our end user docs speak about them
<aaime> so I was wondering, can I add the encode(Expression) method to the interface
<aaime> without going thru a proposal?
<aaime> (and I need it for the 2.4.x series)
aaime listens to this resounding silence
<acuster> you could probably persuade jody if you promised him some fodder for the user docs
<aaime> fodder? what's that?
<acuster> stuff
<acuster> food technically
<aaime> http://en.wikipedia.org/wiki/Fodder
<acuster> alegorically, material
<acuster> cannon fodder ---soldiers to get killed by cannons
<jgarnett> thinking ...
<jgarnett> aaime; this JDBC stuff is a contract between you and the subclasses of JDBCDataStore
<jgarnett> it is not really user facing code is it?
<aaime> no
<jgarnett> then personally I don't care too much.
<aaime> it's not meant to be, yet of course the interface of SQLbuilder is public
<aaime> So it's ok for me to add that method?
<aaime> +1 for me
<aaime> jgarnett, jdeolive? vote?
<jdeolive> +0 I have not followed this issue
<acuster> sounds good
<simboss> +1
<jgarnett> +1
<acuster> in your email, you spoke of not having time to do the right thing.
<jgarnett> aaime; there is another way - volunteer to be the jdbc module maintainer
<acuster> if that's true, any chance you could lay out what that would be for future reference
<aaime> acuster, this way I can do it right
<acuster> great
<aaime> there were other "right" options that were more expensive than this one
<aaime> jgarnett, not ready to give up the last bit of my spare time
<jgarnett> (hey I gotta try)
<jgarnett> moving on ...
<jgarnett> 5) backporting paging to 2.4.x
<jgarnett> Now this one is user facing groldan
<groldan> more jdeolive's actually
<groldan> but anyways
<jgarnett> so tell me why we are considering this? (hint it involves you being paid and 2.5.x being a long way away)
<acuster> lol
<groldan> sort of
<jdeolive> jgarnett: yes
<acuster> do we have a sense of where the OGC is with their various proposals on this issue?
Eclesia bye ++
<jdeolive> jgarnett: and its only adding api... which we have allowed in teh past on a stable branch
Eclesia (n=sorel@mtd203.teledetection.fr) has left #Geotools
<aaime> acuster, OGC side stepped the issue
<jgarnett> I have tired to limit us to adding functionality (additional plug-ins etc...) and not api in the past.
<jgarnett> but your point is taken.
jdeolive laughs when people talk abotu enforcing backwards compatability in geotools
<jdeolive> but your point is taken as well
<jgarnett> 2.4.x does not have an user docs; but I would like to ask that we behave a bit better once 2.5.x series goes out.
<jgarnett> so how about this gabriel; you do the work; and you write a nice note about it in the release notes; explaining that we are very bad etc...
<groldan> I can do the work and add some doco, won't say we're bad though
<jgarnett> lol
<jgarnett> well we are at least treading in a gray area
<jgarnett> possibly at night.
<acuster> the puritan vs the latin
<acuster> groldan, you have what you need?
<acuster> are we done?
<groldan> I have what I need, if that means constantly moving on a gray area
ggesquiere has quit (No route to host)
<groldan> okay thanks for the exception on behalf of the ones doing this paid work
<acuster> money talks, eh?
<groldan> and there are three minutes left for the end of the meeting yet
<groldan> acuster: that you already know
<groldan> that's how geotools goes forward I guess
<groldan> </meeting>?
<acuster> indeed