• cheeser
  • ernimril
  • joed
  • kinabalu
  • lunk
  • ojacobson
  • r0bby
  • ThaDon
  • ricky_clarkson
  • topriddy

« 2017-08-03


2017-08-05 »

Nick Message Date
pualj [pualj!~pualj@] has joined #glassfish [12:03]
pualj [pualj!~pualj@] has joined #glassfish [04:01]
djinni` [djinni`!] has joined #glassfish [04:32]
acidjnk22 [acidjnk22!] has joined #glassfish [04:45]
acidjnk hello [05:08]
acidjnk I'm using a plain embedded payara with plain JUnit. When I do a lot of testing, long methods, many methods etc., injections in EJBs start to fail occasionally. [05:09]
acidjnk Any workaround? Maybe I can prevent it from using more than one @Stateless instance for testing? [05:10]
pualj_ [pualj_!~pualj@] has joined #glassfish [05:13]
pualj_ [pualj_!~pualj@] has joined #glassfish [05:21]
sarri [sarri!] has joined #glassfish [06:33]
sarri [sarri!~sari@unaffiliated/sarri] has joined #glassfish [06:33]
yeitijem [yeitijem!] has joined #glassfish [06:37]
pualj [pualj!~pualj@] has joined #glassfish [06:53]
pualj [pualj!~pualj@] has joined #glassfish [07:15]
pualj_ [pualj_!~pualj@] has joined #glassfish [09:20]
pualj [pualj!~pualj@] has joined #glassfish [11:46]
pualj_ [pualj_!~pualj@] has joined #glassfish [12:10]
pualj [pualj!~pualj@] has joined #glassfish [12:26]
pualj_ [pualj_!~pualj@] has joined #glassfish [12:39]
pualj [pualj!~pualj@] has joined #glassfish [01:11]
pualj_ [pualj_!~pualj@] has joined #glassfish [01:56]
_Tenchi_ [_Tenchi_!] has joined #glassfish [02:21]
acidjnk wow, had two bad days due to that. But I think the problem only occurs when using the remote interfaces a lot from within the same JVM on the embedded payara. [02:46]
whartung people still use remote interfaces? [02:47]
jdlee remote interfaces in the same JVM? eek :) [02:52]
acidjnk They have this really strange design. The lookup method takes for example lookupLogicBean(MyService.class), then with a substring calculates the JNDI string, replacing Service by ...Bean. But not even their naming is consistent. [02:55]
acidjnk Sometimes ...Service is a local interface, at other times it is remote. and ...ServiceLocal would be the local one. [02:56]
acidjnk But then, the substring method does not work :-) [02:56]
acidjnk The remote interfaces would be used when it is run in a cluster though, right? That might have been the intend of putting them there in the first place. [02:58]
pualj [pualj!~pualj@] has joined #glassfish [02:59]
jdlee i'm not an expert in clustering, but i think if you used a local interface and let the container handle state replication, you'd be fine [03:02]
jdlee ??\_(???)_/?? [03:02]
acidjnk anyway strange that the embedded server gets messed up from using the remote interface a lot [03:04]
acidjnk I tried everything from glassfish 4.1 over oldest payara to latest payara [03:04]
jdlee yeah, that's odd [03:10]
pualj [pualj!~pualj@] has joined #glassfish [03:20]
pdurbin jdlee: see you at JavaOne? [04:02]
whartung when is javaone? [04:24]
whartung (Isn???t it Oracle One or something now, with a booth labeled ???Java????) [04:25]
jdlee pdurbin: i don't know yet. i have a hotel room reserved, but I'm still trying to get approval to purchase air and conf tickets [04:38]
jdlee whartung: JavaOne and Oracle OpenWorld [04:38]
jdlee October 1???5 [04:43]
whartung ah ok [04:43]
whartung you going pdurbin ? [04:43]
jdlee and I just realized that falls right in the middle of a potential business trip out of the country :( [04:44]
whartung alert Oracle, maybe they???ll move the conference for you [04:44]
jdlee haha [04:46]
pdurbin jdlee: if you submit a talk and it's accepted, they waive the conference fee. That's what I did. Just booked my hotel. [05:29]
whartung what???s your talk about pdurbin ? [05:30]
pdurbin whartung: heh, yes it's basically the Oracle show these days with some stuff on Java. At least we get our own space, typically. [05:30]
pdurbin whartung: maybe you and others can help me write this talk. _Tenchi_ jdlee whoever else. Give me ideas anyway. It's called "BOF2805: How to Run a Successful Open Source Java EE Project". [05:32]
whartung When you show up tell them ???Oh, this was supposed to be a talk? I thought it was a fact finding mission??? :) [05:33]
pdurbin heh [05:33]
_Tenchi_ "successful" is too much of an objective term [05:33]
whartung yea [05:33]
whartung what does that mean [05:33]
pdurbin hmm, what would be better? [05:34]
whartung well, simply, you ahve an OS JEE project yes? [05:34]
pdurbin hes [05:34]
pdurbin yes* [05:34]
pdurbin hell yes [05:34]
whartung what challenges have you had and how did you meet them? [05:34]
whartung that???s what you should talk about [05:34]
_Tenchi_ "How to define critierias for success in today's opensource java EE projects" [05:34]
whartung do you have a lot of comtters? [05:34]
pdurbin one problem is that no one has heard of glassfish :) [05:35]
jdlee pdurbin: are you an Oracle employee? [05:35]
pdurbin challenge* [05:35]
pdurbin jdlee: nope. I work at Harvard. [05:35]
_Tenchi_ do you have only 3 committers to your project but the project generates 2.5 billion dollars of service revenue for you? [05:35]
_Tenchi_ i'd say that's successful [05:35]
jdlee i actually went last year as a speaker, but that was before the oracle acquisition [05:35]
whartung well that???s the thing, not that I went googling, but yours is the only OSS project I know of that uses EJBs. [05:35]
_Tenchi_ but others would say no, that's not success [05:35]
jdlee pdurbin: ok. i didn't think so :) [05:35]
whartung I don???t think servlet apps with JSPs are Java EE. [05:36]
pdurbin github says we have 47 contributors [05:36]
whartung what problems have you had? [05:36]
whartung hae you have any problems with the contributors? [05:36]
jdlee whartung: i think they are, technically. both of those techs are Java EE technologies [05:36]
pdurbin it does not generate billions of dollars of service revenue [05:36]
jdlee you can, though, deploy that to a simple "servlet container" like Tomcat, and not a full EE app server [05:37]
whartung Oh yea, they are. But that???s like saying running Tomcat on Linux is a ???Unix application" [05:37]
pdurbin jdlee: that's not how I'm defining Java EE. Our app doesn't run on Tomcat. [05:37]
whartung JEE is pretty broad, and, yea, while using one or two of the technologies is ???JEE??? it???s not, well, ???JEE???. [05:38]
jdlee pdurbin: with EJB, i would think not (TomEE, though... :) i tend to look at servlet+JSP as EE, though. maybe kind of a precursor to the profiles [05:38]
whartung I wouldn???t call that developer or designer an JEE develpoer. [05:38]
whartung you can talk about portability issues (or not since you ???just run glassfish???) pdurbin [05:39]
whartung ???how important is portability today??? kind of thing with an OSS JEE app [05:39]
jdlee i would be ok with calling that person a Java EE developer [05:39]
pdurbin whartung: we just run glassfish. I don't think our app would run on wildfly or tomee without a fair amount of work. [05:39]
whartung right, so that could be an aspect that you explore in your talk. [05:39]
jdlee pdurbin: do to API usage or app configuration? [05:40]
whartung the importance or lack of that [05:40]
jdlee due to... good grief :P [05:40]
whartung doo da [05:40]
pdurbin jdlee: app configuration. Our installer, at least, is very glassfish-specific. [05:40]
jdlee nods [05:40]
whartung To me JEE is about managing the transactional state as manifested by the EJB tier ??? that???s the heart of it. Servlet programming doesn???t teach you any of that. [05:41]
jdlee whartung: but Java EE these days seems to have move past EJB, largely, so that defintion is a bit limiting, no? [05:41]
pdurbin I'm with whartung. A Java EE app requires a Java EE application server. If it'll run on tomcat it's just a java web app. [05:41]
jdlee meh :P [05:41]
whartung well, they???ve blurred the lines between the tiers as manifest entities of the application, but the transaction barrier is still there. [05:42]
pdurbin jdlee: come to my talk and we can arm wrestle about it :) [05:42]
jdlee what about a Java EE microservice? say, JAX-RS, CDI, JPA? definitiely EE techs, but no app server [05:42]
jdlee pdurbin: haha. i would love to, if I can get oracle to fly me out there :) [05:42]
pdurbin it's oracle's conference! [05:42]
_Tenchi_ i consider anything using a JEE api to be JEE [05:42]
_Tenchi_ servlets... JEE [05:43]
_Tenchi_ jsps... JEE [05:43]
jdlee s/JEE/Java EE/ ;) [05:43]
whartung but JEE is about how those invididual aspects of the platform work together [05:43]
pdurbin _Tenchi_: so what do you and jdlee call what my app is. "fat Java EE"? [05:43]
jdlee whartung: i guess, but not all EE apps need all of that complexity [05:43]
jdlee pdurbin: "Java EE" [05:43]
acidjnk If I have to reduce it to two ideas as takeaways for Java EE, I usually say: The container provides commonly needed things such as pooling, transactions and other things. [05:43]
_Tenchi_ i either say "JEE" or "serverside java" [05:43]
whartung no they don???t. But then those that don???t, don???t need JEE either, right? [05:43]
jdlee "Oh, what specs/libs?" "EJB, JTA, etc" [05:44]
pdurbin Java EE has web profile and uh... full profile? I could say that ours is a Full Profile Java EE Project. [05:44]
jdlee whartung: i think you're conflating "Java EE" with transactional, clustered applications, perhaps :) [05:44]
jdlee pdurbin: there you go [05:44]
_Tenchi_ unless i'm fighting with lawyers and trademark police... I don't really care, as long as people can understand me [05:45]
pdurbin phew! I didn't think I'd have to define what kind of app I'm talking about at the beginning but I guess I will. "requires glassfish" "won't run on tomcat" [05:45]
jdlee and, of course, there's the microprofile, which isn't technically part of the official EE spec, but is another valid flavor of Java EE [05:45]
_Tenchi_ but if people want to debate me, i just say... call it whatever you want [05:45]
pdurbin our app won't run on microprofile either [05:46]
jdlee pdurbin: "Requires a [micro|web|full] profile environment" [05:46]
whartung Early on, JEE was about the component model, and that???s has faded for some time. But it???s truly distinguishing capability, the ONE thing that makes it stand out above everything else on the market (save Spring with JEE-with-another-name) is the transaction layer. Nothing else has it as a first class pervasive attribute. And understanding that is critical to developing an JEE application. [05:46]
jdlee _Tenchi_: yeah. 'cept for this discussion, i try not to get too worked up about it. not that I am here. I'm just putting off doing real work :P [05:47]
_Tenchi_ jdlee: you and me both haha [05:47]
jdlee hehe [05:47]
jdlee _Tenchi_: there's this guy that keeps nagging me for a book review... ;) [05:47]
whartung understanding how that transaction layer interacts with JPA, CDI, Servlets, EJB, MDB, JCA, etc. etc. etc. is pretty fundamental [05:48]
pdurbin whartung: and that layer is called JTA, right? [05:49]
whartung partly. You can???t jsut do it with JTA ??? that???s why we have the containers to vita-mix it all up. [05:49]
whartung And the transaction management is the one thing that keeps me from abandoning the platform. [05:52]
pdurbin don't leave us, whartung! [05:53]
pdurbin acidjnk: those takeaways you mentioned make sense [05:55]
pualj [pualj!~pualj@] has joined #glassfish [06:04]