The IRC channel is #sakai on the freenode network, you can join using a web client at http://webchat.freenode.net/?channels=sakai and you can view the complete logs at http://sakai.iriscouch.com/irc/_design/viewer/index.html
00:08 <GitHub157> [nakamura] arwhyte pushed 6 new commits to 1.4.1: http://git.io/IKpxSw 00:08 <GitHub157> [nakamura/1.4.1] NOJIRA revert pom versions to 1.4.1-SNAPSHOT - Anthony Whyte 00:08 <GitHub157> [nakamura/1.4.1] NOJIRA revert nakamura dependencies with funky versions to *1.4.1-SNAPSHOT - Anthony Whyte 00:08 <GitHub157> [nakamura/1.4.1] NOJIRA update poms missed by script update - Anthony Whyte 02:18 <GitHub93> [nakamura] arwhyte tagged org.sakaiproject.nakamura-1.4.1 at a48d676: http://git.io/O1FjDQ 02:18 <GitHub93> [nakamura/org.sakaiproject.nakamura-1.4.1] [maven-release-plugin] prepare release 1.4.1-release-tag - Anthony Whyte 02:18 <GitHub93> [nakamura/org.sakaiproject.nakamura-1.4.1] NOJIRA missed version updates to 1.4.1 - Anthony Whyte 02:44 <GitHub57> [nakamura] arwhyte tagged org.sakaiproject.nakamura-1.4.1 at c807e31: http://git.io/zaW7qw 02:44 <GitHub57> [nakamura/org.sakaiproject.nakamura-1.4.1] [maven-release-plugin] prepare release 1.4.1-release-tag - Anthony Whyte 02:44 <GitHub57> [nakamura/org.sakaiproject.nakamura-1.4.1] NOJIRA update configuration, webdav and integration poms to 1.4.1 - Anthony Whyte 02:44 <GitHub57> [nakamura/org.sakaiproject.nakamura-1.4.1] NOJIRA missed version updates to 1.4.1 - Anthony Whyte 03:04 <GitHub170> [nakamura] arwhyte created 1.4.1 from org.sakaiproject.nakamura-1.4.1 (+9 new commits): http://git.io/ou4z4Q 03:04 <GitHub170> [nakamura/1.4.1] NOJIRA revert pom versions to 1.4.1-SNAPSHOT - Anthony Whyte 03:04 <GitHub170> [nakamura/1.4.1] NOJIRA revert nakamura dependencies with funky versions to *1.4.1-SNAPSHOT - Anthony Whyte 03:04 <GitHub170> [nakamura/1.4.1] NOJIRA update poms missed by script update - Anthony Whyte 09:45 <SakaiGithub> [3akai-ux] bp323 pushed 2 new commits to 1.4.2: http://git.io/oUsVKw 09:45 <SakaiGithub> [3akai-ux/1.4.2] SAKIII-6189 Fix the 'add to' button data in contacts - Bert Pareyn 09:45 <SakaiGithub> [3akai-ux/1.4.2] Merge pull request #2491 from bp323/SAKIII-6189 - Bert Pareyn 09:54 <SakaiGithub> [3akai-ux] bp323 pushed 1 new commit to 1.4.2: http://git.io/u2ToFQ 09:54 <SakaiGithub> [3akai-ux/1.4.2] SAKIII-6196 fix lastmodified timestamp by sending a random value for the "forceupdate" parameter - Duffy Gillman 13:21 <stuartf> PhysX: ping 13:21 <PhysX> stuartf pong 13:22 <stuartf> hey, I was looking at the Hilary project yesterday, how tied are we to the idea of each tenant running on its own port? 13:23 <PhysX> stuartf I think pretty tied 13:23 <PhysX> stuartf did you have some thoughts on it? 13:24 <nicolaas> it definitely helps development 13:24 <nicolaas> it's possible that something else could be done for production 13:24 <stuartf> yeah, I was thinking it would be more flexible/scalable if it all ran on one port and we did the mapping based on the hostname in the request 13:26 <stuartf> for instance I was going to demo how easy it would be to cloud deploy it on heroku, but heroku only lets you open one port 13:26 <mrvisser> stuartf: hostname was the original thought, but becomes a problem if you aren't running a reverse proxy in development 13:27 <nicolaas> nodejutsu does allow you to open up all ports 13:27 <mrvisser> stuartf: also, with reverse proxying, more work is required to pass along some kind of env variable from the actual virtual host, since to the app server everything looks like "localhost", or the local host in general 13:28 <mrvisser> stuartf: we fell back to port-based, fully expecting to expand on the strategies in the future. 13:28 <mrvisser> but definitely didn't know there may port restrictions in some cloud environments, hmmm.. 13:29 <stuartf> I get the issue with proxies showing as localhost, but I'm not sure I understand why dev would need a reverse proxy 13:30 <nicolaas> ec2 seems to allow you to open up ports 13:30 <mrvisser> stuartf: just to narrow the gap between "production" and "dev" environments. 13:31 <mrvisser> stuartf: I suppose you could map hosts to 'localhost' on your local machine 13:31 <nicolaas> I don't think it's a particularly hard problem to solve 13:33 <mrvisser> I'd be open to revise a new strategy if there are clear benefits to not being constrained to ports 13:33 <mrvisser> and it sounds like there may be 13:39 <stuartf> I was thinking there might be a mapping like "*" that just catches anything not mapped so it would grab localhost etc 15:09 <nicolaas> ping stuartf 15:09 <stuartf> nicolaas: pong 15:09 <nicolaas> Have you had a chance to look at the 1.4.2 ticket that's assigned to you ? 15:10 <stuartf> the one about ie8? 15:10 <nicolaas> it's the only remaining one, but we need multiple eyes on it as it's hard to track down 15:10 <nicolaas> yes, although it might not be strictly limited to ie8 15:10 <nicolaas> there have been safari reports as well, but not nearly as widespread 15:10 <stuartf> I'm going to look at it after lunch 15:10 <nicolaas> thx stuartf 15:11 <stuartf> it's been a while since I booted up my XP vm, should be interesting 15:11 <nicolaas> :) 15:31 <mrvisser> stuartf: can we have a skype call sometime this afternoon re: performance testing ? 15:32 <mrvisser> stuartf: If you're heads-down on the 1.4.2 ticket, I don't mind pushing to a later time. 15:48 <stuartf> mrvisser: yeah, my skype is broken, but I can call a landline or give you my office number 15:49 <mrvisser> stuartf: office number should be fine 19:56 <lancespeelmon> ctweney: what is the general wisdom around eTags and returning collections (i.e. Lists) of entities? 20:04 <ctweney> lancespeelmon: it depends on what you're listing, I think… ETags are good when you don't know in advance when your cache should be invalidated 20:05 <ctweney> if you're listing multiple entities it can be harder (but not impossible) to invalidate the cache 20:05 <ctweney> lancespeelmon: are you speaking in the context of OAE's Dynamic response cache? 20:05 <lancespeelmon> no more generically 20:06 <ctweney> ETags force the client to make a request every time… but shortcut the delivery of the body content if the client presents a fresh ETag 20:06 <ctweney> expiration dates tell the client not to make a request again until the expiration time 20:06 <ctweney> so it's a balancing decision 20:07 <lancespeelmon> ok that helps - thanks Chris 20:07 <ctweney> usually etags are right for dynamic content that you control 20:43 <lancespeelmon> mrvisser: what are you using to generate javadoc for jax-rs annotations? 20:44 <lancespeelmon> lunatech or enunciate? 20:44 <mrvisser> lancespeelmon: lunatech 20:45 <lancespeelmon> have you used their javadoc tags? http://fromage.github.com/jax-doclets/docs/0.10.0/html/index.html#d0e538 20:45 <lancespeelmon> I cannot get @returnWrapped fq-classname doc to work without throwing NPE 20:45 <mrvisser> lancespeelmon: if you're using maven3+, you won't be able to get Javadoc, JAXB docs, JAX-RS docs all in one "mvn javadoc:javadoc" command, which is why I mushed them together with the NakamuraDoclet 20:45 <lancespeelmon> oh cool 20:45 <mrvisser> lancespeelmon: don't believe I have 20:45 <lancespeelmon> yeah I know mvn3 has a problem 20:46 <lancespeelmon> mrvisser: where can I find NakamuraDoclet? 20:47 <mrvisser> https://github.com/sakaiproject/nakamura-doclet 20:47 <lancespeelmon> ah 20:47 <mrvisser> I can grab an example of using it, one sec 20:47 <mrvisser> https://github.com/sakaiproject/nakamura/blob/master/pom.xml#L718 20:48 <mrvisser> in the pom.xml file, you basically just replace the default doclet with NakamuraDoclet. NakamuraDoclet includes the Javadocs. 20:48 <mrvisser> lancespeelmon: so, if you add that and run: "mvn javadoc:aggregate" from the root of the project, it will aggregate one giant JAX-RS doc website for all the bundles combined. 20:49 <lancespeelmon> very cool thanks 20:49 <mrvisser> np 20:52 <lancespeelmon> mrvisser: do you have any advice say for a getUser() method - should it return a User object or a jars Response ? 20:53 <lancespeelmon> s/jars/jaxrs 20:55 <mrvisser> lancespeelmon: I would recommend returning a User object and let Jackson do it's thing. But there are types when you'll need to just get the Response object and do something with it. In that case, if your REST endpoint still returns a "User" object to the client, you can do so with Response.entity(user) 20:55 <mrvisser> s/types/times 20:55 <lancespeelmon> right 20:56 <lancespeelmon> ok cool thx 20:56 <mrvisser> lancespeelmon: but in the spirit of re-usability, if your method is returning or accepting JAX-RS / HTTP specific entities, try and delegate all business logic to a service 20:56 <mrvisser> np 20:56 <lancespeelmon> sure _______________________________________________ oae-dev mailing list oae-dev@collab.sakaiproject.org http://collab.sakaiproject.org/mailman/listinfo/oae-dev
