(I'm posting the logs here because apparently there is no more news space on the Geoserver.org confluence? Where did it go?)
---------------------------------------------------------------------- Summary: - Google soc - 1.6.3 release and related bugs - UI framework talk ---------------------------------------------------------------------- aaime: topics? arneke: I guess I groldan: not really arneke: 'm interested in whether we have any potential GSOC applicants? dwinslow: arneke: afaik we don't find out until the end of the application period dwinslow: or if they bug us in person arneke: right, but people who have contacted us seriously arneke: the latter aaime: no one contacted the list afaik aaime: (nor me in private) aaime: anyone else aaime: got an alien contact this week? arneke: ok, there's a 50% chance I have one aaime: nice, topic? arneke: he said the database backends for GeoServer aaime: ah ,sql server aaime: jdeolive, this would be for your new jdbc framework right? arneke: cholmes was inclined to push him to GWC since I know him, but that would be his choice jdeolive: aaime: i would think so... from what i coudl tell on the project description aaime: hum, whatever, but the student must be enthusiast about whatever he's doing aaime: so better have him on sql server if he loves to arneke: and have a US work permit <- there's where the 50% comes in aaime: (imho) aaime: Oh, why is the US permit so important? aaime: in person contact with the student is real nice arneke: otherwise Google doesnt give you the scholarship aaime: but lack of it did not stop successfull applications in the past aaime: scholarship == ?? arneke: the USD 5000 students get to participate? ;) arneke: stipend aaime: they don't give that to non US students? that's a first for me? arneke: good think you mentioned it, let me double check the eligibility page groldan: borsa di studio aaime: thanks groldan :) groldan: :P arneke: ah... crap. Thanks aaime: http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_student_eligibility sigq: Title: Google Summer of Code - Google Code (at code.google.com) arneke: "You must also be eligible to work in the country in which you'll reside throughout the duration of the program, e.g. if you are in the United States on an F-1 visa," aaime: so one can stay at home and be paid anyways, no? dwinslow: that's how I read it groldan: but you can't stay in Cuba aaime: the problem seems to be raised only if you're resident in the USA with the wrong visa no? aaime: I'd suggest to stay away from Tibet too aaime: thought it's not in the list ;) groldan: seems wise for the moment arneke: yes arneke: he's from romania jgarnett: so working anywhere in the EU should be good for him? arneke: well, he's a F1 student at Vassar, but will be graduating before that arneke: jgarnett: you'd think, but they put extra restrictions on the newcomers arneke: so it's back to a tiered society based on passport arneke: anyway, I'm emailing him regarding the good newsx aaime: cool aaime: so release is scheduled for the end of the next week aaime: any concern? groldan: just ammount of issues... aaime: yep, me too :) aaime: I'm also a little worried about pissing off some users groldan: how so aaime: since a few issues have been postponed like 2 times dwinslow: have the issues assigned to ivan been reassigned? i only see one with his name on it aaime: dunno. I have seen that one and I raised the question groldan: hmmm.. we could think of raising them up as blockers groldan: you know which ones? cholmes [EMAIL PROTECTED] è entrato nella stanza. groldan: "release directory contains jars" is giving grief to sde users, for example aaime: yeah, I guess it would be easier to decide which issues are harmless enough groldan: http://jira.codehaus.org/browse/GEOS-1682 sigq: Title: [#GEOS-1682] release directory contains jars - jira.codehaus.org (at jira.codehaus.org) aaime: that we can move them to 1.6.4 aaime: Hi Chris, feeling better= aaime: ? groldan: jdeolive: do you think you'll be able to at least give some insight on it? groldan: (I'm sure deleting that gt 2.0 arcsde jar would be fine for sde, but wonder about the others) aaime: deleting the release/extensions/arcsde/gt2-arcsde.jar right? groldan: for example, there's also a ojdbc14.jar aaime: seems the thing to do from where I stand jdeolive: groldan: hi ***mchua is trying to get a gsoc applicant in this channel atm, btw jdeolive: i tried to look into that jdeolive: but i awas not sure what the issue was groldan: aaime: right aaime: jdeolive, from what I can see jgarnett: a bit confused? the the gt2-oracle jar should not "do anything" unless an ojdbc14.jar is around? groldan: jdeolive: I commented on it after you told so jdeolive: ahh, ok, let me look aaime: the 2.4.x jar is being pulled from the dependencies aaime: but there is also an old jar that is stuck in the release module aaime: statically, as a file aaime: and it gets included as well aaime: probably we want to remove it groldan: ok, given the oracle license is there I guess we can distribute? just wondering aaime: Yes, we can redistribute now jdeolive: ok... i understand now... aaime: that's what we're doing at least aaime: there are quite a bit of jars in the release/extensions dirs aaime: that seems to be something we can pull from the dependencies aaime: instead of having them in the relase directory aaime: ojdbc14 is maybe the only one that we can't put in a repo groldan: ok, makes sense aaime: since it has to be accompained with a license aaime: (can we put stuff with a license you have to read on a maven repo? I guess not?) aaime: hum, maybe yes, look here: http://repo1.maven.org/maven2/com/oracle/ojdbc14/ sigq: Title: Index of /maven2/com/oracle/ojdbc14 (at repo1.maven.org) aaime: or maybe not... just md5 files tehre? groldan: indeed. Its only the pom pointing to the online license aaime: yep: http://jira.codehaus.org/browse/MEV-496 sigq: Title: [#MEV-496] No jar for Oracle POM - jira.codehaus.org (at jira.codehaus.org) Kred ha abbandonato la stanza (quit: "ta ta"). aaime: groldan, I guess you can kill the arcsde and mysql ones thought aaime: and see what happens when you build the extension datastores zip files jdeolive: i am trying that now aaime: ah btw, not much feedback on the UI stuff groldan: ok, actually I'm sure I have to delete that one. The issue were more due to not knowing how exactly the extensions are built aaime: (and sorry for misundrestanding your mail jdeolive) aaime: what other developers think? aaime: people, you're gonna use whatever comes out of the analisys, so better participate actively ;) jdeolive: aaime: no worries, it was more or less just me ranting... not much thought behind it :) arneke: ah, UI aaime: As we say it in italy, I put my hands forward peter [EMAIL PROTECTED] è entrato nella stanza. groldan: anyways, back to the topic, aaime do you know what those long delayed issues are arneke: yeah, one sidenote, that I had started replying to aaime: scripting? no thank you :) (but if I have to, oh well ok) arneke: I think the current UI looks like it was designed by a Java developer arneke: no offense ;) aaime: no offense taken, you're really right arneke: and if you need to understand Java in order to make the UI arneke: then that will never change aaime: but that's one of the points in my reasoning. A java developer will have to mantain it aaime: when we do new services aaime: new datastores arneke: ... true aaime: new configurations aaime: will we have a UI wizard doing all the changes? arneke: or we can enlist people like tschaub... but I see your point aaime: If we have the UI wiz ok, but we need to be sure we have a resident one aaime: otherwise we would end up like the current UI aaime: it's there, sort of works, but nobody can really touch it jdeolive: aaime: from what i can see, a lot of the new frameworks (liek struts2) have "plugins" for doing ajax stuff, and other scripting stuff jdeolive: might be useful to get us the best of both worlds jgarnett: finding it hard to participate in this meeting; andrea I have a preference for GWT or Wicket (everything else looks like a compromise) aaime: jgarnett, it's because it's not an organised meeting arneke: ok, I retract my point, I also see you covered it more in your email than i first picked up aaime: we're really chit-chatting jdeolive: groldan: confirmed, we need to kill that jar in svn jdeolive: the arcsde one groldan: ok, only that one? aaime: mysql too imho aaime: jdeolive, ajax stuff can be done in two ways jdeolive: so there is another connector that gets dragged in as a dependency? aaime: I believe so, there should be aaime: I mean, the mysql datastore needs one groldan: repeat: I'm sure the arcsde needs to go, wondering about the others and if its a matter of just removing it/them or some config file needs to be touched aaime: and it's open source aaime: Oh hum, jgarnett si right, better get organised... jdeolive: aaime: looking at the mysql artifact... i only see the datastore, and the connector [EMAIL PROTECTED]: aaime ha scelto come argomento: 0) What's up 1) Extensions extra jars 2) UI aaime: and we're not at topic 1) aaime: the connector being the jdbc driver jdeolive: right aaime: but mysql datastore is using a newer driver than the one included in the extensions directory jdeolive: ahh... ok jdeolive: so jdeolive: we should delete teh old one jdeolive: and add the new one as a dependency so it gets dragged in properly ... ( i think ) aaime: Yeah, that's what I think as well aaime: groldan, yes, I think some xml files have to be modfiifed jdeolive: ok cool, and same goes for oracle? aaime: in order to have the dependency be pulled from the repo? aaime: oracle not, since the jar is not in the repos? groldan: cool, so that issue is being worked out groldan: andrea aaime: yes? groldan: you were worried about other ones aaime: ah yes groldan: for the release [EMAIL PROTECTED]: aaime ha scelto come argomento: 0) What's up 1) Extensions extra jars and other 1.6.2 nasty issues 2) UI aaime: just a sec aaime: well, they are easy to spot aaime: sort by key aaime: http://jira.codehaus.org/secure/IssueNavigator.jspa?sorter/field=issuekey&sorter/order=ASC sigq: Title: Issue Navigator - jira.codehaus.org (at jira.codehaus.org) aaime: the frist ones have been moved over a few times aaime: the one I was thinking about is http://jira.codehaus.org/browse/GEOS-1580 sigq: Title: [#GEOS-1580] Data properties unavailable after updating a shapefile and reloading the catalog. - jira.codehaus.org (at jira.codehaus.org) aaime: thought there is also this one that has been marked critical: http://jira.codehaus.org/browse/GEOS-1373 sigq: Title: [#GEOS-1373] Postgis datastore queried with a fid filter, returns features with the same number but different fid - jira.codehaus.org (at jira.codehaus.org) aaime: and I never really understood why aaime: seems one that has little return and lots of work behind it cholmes ha abbandonato la stanza (quit: ). aaime: I saw Chris comment but seems like a problem in the analisys of the issue they found aaime: or else I'm not understanding it groldan: hmmm aaime: as usaul the geoserver-wrap-o-matic solution is easy but ugly groldan: I'm having a hard time trying to understand what exactly it is aaime: well, the original issue there is that aaime: when you ask for fid=states.15 aaime: the datastore will kill whatever is before the period groldan: seems to be that if you send a fid filter over ft a, but the filter is b.1, b.2 you'll get a.1 and a.2 anyways? aaime: so if you filter using hoolabaloola.15 you get the same result aaime: yes groldan: ok aaime: but it's also a limitation of the fidmapper api aaime: that is, the fid mapper does not have a way to say "I don't recognized this FID as valid, so don't include it in the sql" groldan: aha jdeolive: groldan: artifact issue should be fixed aaime: an exception is over-reacting, blah.15 is a legitimate fid, should just not return anything groldan: hmm... put arcsde on the list, though its not jdbc ***jdeolive is happy when he gets to remove binary jars from svn aaime: yeah, everything that's using the typed fid mapper is affected groldan: cool, thanks a lot aaime: shapefile too aaime: but as I noticed in the comments, I don't see an easy cure groldan: jdeolive: you gonna commit? shall I close the issue? jdeolive: groldan: done, and done aaime: (port to trunk?) groldan: great jdeolive: aaime: done :) TeddyBouch [EMAIL PROTECTED] è entrato nella stanza. aaime: excellent :) aaime: guys, time is flowing aaime: can anyone of you make up a list of things you believe should be moved away to 1.6.4 aaime: and post it to the thread on the mailing list? aaime: or else, make a list of the ones you absolutely want to keep aaime: 38 issues are too many for the time we have left aaime: imho jdeolive: agreed groldan: yeah, the second option better, so we have more freedom on what to trhow out jdeolive: maybe we can get cholmes to go through and filter out jdeolive: but having us pre-filter wont be a bad idea either aaime: btw, jdoelive, isn't http://jira.codehaus.org/browse/GEOS-1618 already fixed? sigq: Title: [#GEOS-1618] pass describe feature type cite tests - jira.codehaus.org (at jira.codehaus.org) aaime: yeah, we can build a list of suggestions nickyg ha abbandonato la stanza (quit: Read error: 104 (Connection reset by peer)). nickyg [EMAIL PROTECTED] è entrato nella stanza. jdeolive: aaime: yes i thin it is jdeolive: closing it groldan: you resolved to get some testing from my side or mistakenly? aaime: so coming back to the UI) aaime: is there anything you'd love to see there? aaime: any favourite web framework aaime: ? groldan: not flash aaime: or favourite approach to make web apps independent of the framework= aaime: what would you like to see? aaime: don't tell me you don't have an opinions on web apps ;) jdeolive: aaime: to be honest it think your requirements are good jdeolive: what i am looking toward now jdeolive: is the tech evaluation, the meat and bones stuff groldan: not sure what to add to your email andrea mchua: dwinslow: TeddyBouch is the gsoc student btw jdeolive: i want to see what this stuff will actually be like to work with dwinslow: mchua: ah, thanks dwinslow: hi TeddyBouch aaime: mchua, thanks, TeddyBouch, can you wait 10 mintues? jdeolive: my decision will most likely be based on that mchua: dwinslow: (for when you folks are ready) aaime: eh, the thing that bothers me TeddyBouch: Absolutely aaime: is that the list I presented in the mail aaime: is basically the same, point by point, that was used one year ago aaime: and I alrady did some tech evaluations aaime: but they weren't deemed satisfactory so aaime: I'm a little desperate as to what I should do jdeolive: right... but at that point jdeolive: the ui was still far away... so i am nto sure many people gave your findings due course groldan: you told you know how jira does its plugins, right? jdeolive: now that the ui work is coming up... we will have to review it closely jdeolive: also aaime: groldan, xwork + freemarker groldan: (I don't know how, but I've seen them being simple jars?) jdeolive: a year old evaluation in the java web framework world is ancient :) jdeolive: probably want to dust it off and shine it up aaime: which brings me to the point aaime: does anyone want to pick this up? jdeolive: doing the evaluation? aaime: yes aaime: I've lost steam on this topic jdeolive: what i think would work aaime: it's becoming frustrating jdeolive: is everyone chooses a technology.. perhaps the one they think they might like to use jdeolive: and then each does their own litle mock application aaime: If no one else wants to do it ok, no problem jdeolive: witha set of criteria layed out jdeolive: i would like to do an evalutation yes aaime: jdeolive, yes aaime: sounds like a good idea jdeolive: cool jdeolive: so perhaps youa re done andrea jdeolive: since the wicket one is complete aaime: lmorandinini is cooking up a CooCoon prototype similar to the one I did with Wicket groldan: ah groldan: one thing for the ui aaime: jdeolive, I studied Struts2 for the last 2 days jdeolive: cool, someone can probably do struts2 aaime: but the lack of tools support is really killing me groldan: I've got lots of browser timeouts waiting for the config to be applied jdeolive: perhaps i can eat my own dog food and do a groovy one groldan: wonder if more async jobs would be needed aaime: groldan, I would say there is something wrong in a datastore if that happens? jdeolive: agreed groldan: it doesn't matter ***dwinslow points out he's had that problem in the REST config api as well aaime: going anynch requires non trivial code? jdeolive: i think the ui needs to be more careful about how it does io, when processing arequest groldan: in this case is the sde server being hit through a sdl connection groldan: so its slow groldan: but can be anything else, having 250 ftypes, etc aaime: uh, so the apply and save stuff should be handled like a bugzilla search page jdeolive: groldan: perhaps this will make good criteria for an evaulation jdeolive: see how the various frameworks handle these types of situations groldan: yup, could be good jdeolive: so stepping back...perhaps we should set up teh wiki page to start doign teh evaluations? aaime: yes, why not aaime: one is already there jdeolive: cool aaime: each one championing a different web framework aaime: sounds fun! :) jdeolive: let the games begin!! :) ***jdeolive feels like he is stepping into the colusseum aaime: so is modularity still a strict requirement? Or do we allow frameworks that do not allow it too? ***dwinslow has 0 experience with web frameworks outside the current struts config stuff dwinslow: but i'd be up for one if you guys have another framework in mind jdeolive: aaime: i think so... aaime: dwinslow, the first time is not always that good, this is the proof jdeolive: or at least that is my opinion milovanderlinde1 ha abbandonato la stanza (quit: Read error: 113 (No route to host)). aaime: jdeolive, your opinion being? :) jdeolive: that for choosing candidate technologies we drop that requirement for the time being jdeolive: and consider it more in the final choice aaime: hum, then GWT becomes a candidate aaime: it was my second choice besides Wicket, but scapped it because it could not do pluggable UI jdeolive: if soemone wants to do it :) jdeolive: anyways, folks, i have to run, lets set up the wiki page and continue on teh mailing list aaime: dwinslow, eager to try that one? ;) dwinslow: it's supposed to work like swing right? groldan: loosing modularity like in having single, separate module for the whole ui? dwinslow: i've done swing before aaime: yes, should be similar dwinslow: i just found http://tapestry.apache.org/ as well, any impressions about that one? sigq: Title: Tapestry - Welcome to Tapestry (at tapestry.apache.org) aaime: that the api changes a lot more than the gt2 one from release to release aaime: this is the old UI page: http://docs.codehaus.org/display/GEOS/User+interface+modular+technology sigq: Title: User interface modular technology - Codehaus (at docs.codehaus.org) dwinslow: i could give GWT a try aaime: it seems bowens already did quite a review: http://geoserver.org/display/GEOS/Google+Web+Toolkit sigq: Title: Google Web Toolkit - Confluence (at geoserver.org) aaime: but I guess you can refresh a lot that page aaime: since GWT went throw one or two major relases in the meantime dwinslow: that does look thorough aaime: yeah, I did not remember that page being there at all! :) jgarnett: aaime as per email; we just did a round of evaulation and ended up with GWT over Wicket. jgarnett: But they were dead even; decision came down to team experience. dwinslow: all right, so does that cover the UI discussion for now? aaime: I guess so aaime: jgarnett, yes, saw the mail, thank you aaime: so some people in the team already used gwt in the past? arneke: neva aaime: ? aaime: neva? arneke: never dwinslow: will there be a requirements page for the webapp 'colosseum' or are we just going to mimic the wicket demo? aaime: ah, good one aaime: I can try to make up one but aaime: if we remove the modularity I'm not sure a demo is needed aaime: thoght some impressions on how it did go in practice would be nice anyways aaime: Ok, I'll try to come up with some requirements for tomorrow aaime: ok? dwinslow: sounds fine to me dwinslow: jdeolive and groldan might have something to say? groldan: I'm ok with it too groldan: I mean, that modularity would be good, and also being able to run the service without ui, so the service and the service ui could well be two modules aaime: yeah, sure aaime: as long as the UI is pluggable I don't see the problem aaime: the backend is already pluggable as is aaime: only config and UI are non pluggable these days no? groldan: and I'm ok with whatever fits both the java developer and our ui wizards (they'll make it pretty and functional after all, we'll add a couple combos from time to time) aaime: groldan, I'm not sure there is such a thing ;) jgarnett: back; sorry andrea - yes they had already used GWT previously and had learn the tricks of debugging etc.. aaime: but we can try to look for the UI graal :) groldan: as a combo? or as a ui wiz? aaime: something that is ok for both the java developer and the UI wiz aaime: arneke, this reminds me of your comment groldan: boh arneke: yes :) aaime: We could try to have a UI wiz help us making the initial cut of the ui aaime: as a framework of css, layout and hints, suggestions aaime: that we can play within when adding new UI pages aaime: so that at least for some time we don't screw up completely arneke: yeah,, but also have them a bit in on the worflow aaime: yeah, thought the backing object model is what it is aaime: so we can improve a lot for sure aaime: but not completely change everything I guess aaime: what I mean that usually the UI exposes all or some of the object model structures aaime: so reconceptualizing the UI requires deep changes in the object model as well (usually, at least) arneke: often, but sometimes it's indifferent arneke: like the submit -> apply -> save aaime: aaagh aaime: horrible arneke: it pretty nonsensical if you don't know how the stuff is stored ;) aaime: but that requires changes in the underlying model aaime: we ditch the submit apply save cycle arneke: not really, the submit can be skipped aaime: we also remove two levels of indirection dwinslow: we can fake it out like i do for the REST api aaime: in the model arneke: and have apply + save dwinslow: where 'fake it out' == 'apply and save in one step' arneke: or that one aaime: ok, but that does not make sense to me :) aaime: it's a temporary solution, a hack arneke: anyway, I'm just saying, we can always argue with the interaction person afterwards aaime: in the long term if you remove it from the UI you really want to remove it from the object model as well arneke: but knowing the program inside and out isnt always great for making UIs aaime: I agree aaime: what I'm saying is that ugly hacks excluded aaime: this will bring a change in the object model ;) arneke: ok :) dwinslow: oh, well, if you want to exclude ugly hacks dwinslow: that's different :) aaime: grr... maintentace ==> no ugly hacks please ;) aaime: I dream of one day when powering up the computer I won't feel sick by looking at the geoserver source code ;) dwinslow: aaime: set your eclipse editor background color to black :) aaime: and leave the foreground black too :) aaime: (damn syntax highlight, I'll see something in any case) dwinslow: turn your monitor off then }:] ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Geoserver-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/geoserver-devel
