Logs for the RingoJS IRC channel,
as logged by ringostarr.
2010-11-24
[07:37] <avi_> what's the idiomatic way to include extra jars in a ringojs app?
[07:48] <avi_> hannesw_: what's the idiomatic way to include extra jars in a ringojs app?
[08:24] <avi_> ah.. dropping them into lib/ works. ok.
[09:44] <hannesw_> hey robertg!
[09:44] <robertg> hi hannes!
[09:45] <hannesw> wanted to email you for meeting
[09:46] <hannesw> was wondering if we could invite outside developers to our next meeting?
[09:46] <robertg> ah ok. i'm on "holiday" this week :)
[09:46] <hannesw> i see.
[09:46] <hannesw> was thinking next week?
[09:47] <robertg> next week is fine with me
[09:48] <hannesw> what do you think about the outside developer idea? (just for the next meeting)
[09:48] <hannesw> (as long as we still fit into the meeting room)
[09:48] <robertg> sure, but shouldn't we then do the meeting in a more central location?
[09:49] <hannesw> he :)
[09:49] <hannesw> we could, but that requires some organization :)
[09:49] <hannesw> so that would be the simple solution (for me at least)
[09:49] <hannesw> ;)
[09:49] <hannesw> but ok to find a more central place
[09:51] <hannesw> oberhamsi: I changed that req.env.lookup() to req.link()
[09:51] <hannesw> still not sure about naming
[09:52] <robertg> well, i'm fine with meeting at our place too, but let me check that on monday, ok?
[09:53] <hannesw> robertg: cool! but monday may be late for inviting for a same-week meetup
[09:54] <hannesw> i'll think about other possible locations
[09:54] <robertg> ok, i'll try to check faster :)
[10:05] <hannesw> would tuesday be ok for an informal developer meet-up in vienna?
[10:12] <hannesw> robertg: just saw your sqlstore commits
[10:12] <hannesw> cool!
[10:12] <robertg> yes, ringo-storable just works fine
[10:13] <robertg> had to change a few things, because the constructor methods are anonymous now
[10:15] <hannesw> hm. you relied on the name property to be set on the constructor?
[10:16] <hannesw> i guess we could define the name property
[10:16] <robertg> i tried, but i couldn't find a way to do that
[10:16] <robertg> it's readonly, afaik
[10:16] <hannesw> right :(
[10:17] <robertg> but it's not much of a problem, i changed to instanceof instead
[10:28] <robertg> reg. unit testing: i've been thinking of changing run() to accept the name of a test method too
[10:28] <robertg> this would allow to run a single test method
[10:29] <robertg> what do you think?
[10:37] <hannesw> robertg: very good idea!
[10:37] <robertg> reg. meeting: thursday 13:00/14:00 would be fine for oberhamsi, marius and me, meeting at our place is fine too
[10:38] <hannesw> robertg: great!
[10:39] <hannesw> so do you think is it ok if I announce it @ your place?
[10:40] <robertg> sure, unless we all fit in our meeting room (8-10 people max.)
[10:54] <oberhamsi> sweet, real life meet :)
[11:02] <hannesw> ok i'll ask people to to respond if they're coming so we can settle for some other location if it's > 10
[13:20] <earl> hannesw: wouldn't `app.link` be a nice place for the url building function?
[13:21] <earl> i confess i haven't looked into how `.link` works atm to see if that's feasible at all
[13:22] <hannesw> earl: what do you mean by "url building function"?
[13:23] <earl> isn't that what req.link does currently?
[13:23] <hannesw> ah, ok
[13:23] <hannesw> well the problem is how does a nested app get its parent app(s)
[13:23] <hannesw> that's currently not defined
[13:23] <hannesw> it probably should be
[13:24] <earl> how would i call req.link currently?
[13:24] <hannesw> but even then, traversing manually over the app chain is not what you want to do
[13:24] <hannesw> req.link("module/id", "additional", "path", "elements")
[13:24] <earl> what if i don't have a module/id?
[13:25] <earl> app.mount('/foo', function (req) ...);
[13:25] <hannesw> you'd need something else
[13:25] <earl> ok
[13:25] <hannesw> we could introduce some abstract name argument
[13:26] <earl> not sure that's needed
[13:27] <earl> does it work with "local" ids?
[13:27] <earl> i.e. if i do app.mount('/foo', function foo(req) ...);
[13:27] <earl> now foo is a named function
[13:27] <earl> does req.link('foo') or req.link('./foo') work?
[13:28] <earl> well, i guess i gotta play around a bit :)
[13:34] <hannesw> no, only works with module ids
[13:34] <hannesw> would be possible to make it work with function names, though that might be a bit erratic
[13:35] <hannesw> imagine you mount some function somewhere and half your app gets derailed...
[13:35] <hannesw> working on router middleware right now...
[13:40] <earl> btw, ok if i push that minor commit mainline: https://github.com/earl/ringojs/commit/9d0eec2
[13:40] <earl> adds a log.info output like the following to webapp startup: "469 [main] INFO ringo.httpserver - Server on http://localhost:8080 started.
[13:44] <earl> a simple nicety to guide users who do not know (or keep forgetting) which ports their apps listen on :)
[13:51] <hannesw> hm, don't you get that message from jetty already?
[13:51] <earl> nope
[13:52] <hannesw> hm, strange, i do
[13:52] <earl> not as a clickable link
[13:52] <hannesw> ah, i see
[14:41] <hannesw> robertg, earl: can you push your changes to the master repo?
[14:42] <hannesw> i don't have the capacity to do this right now
[14:44] <robertg> ah ok, no hurry. am i allowed to push to master repo? not that i would know ;)
[14:56] <hannesw> yes you are :)
[14:56] <hannesw> does github show you a SSH git URL? https://github.com/ringo/ringojs
[14:58] <robertg> ah, yes! didn't see that :)
[15:06] <earl> pushed
[15:08] <earl> there we go, now even the commit hooks work again
[15:10] <hannesw> thanks earl!
[15:10] <hannesw> miracles :)
[15:11] <earl> robertg: do you use your modification of test.js to really run only a single test method in a testsuite?
[15:11] <earl> or rather to run a single "sub-suite" of all ringo tests?
[15:14] <robertg> if i'm executing a "sub-suite" it restricts the runner to a single method, if executing eg. "all.js" it runs all tests of a sub-suite
[15:15] <earl> do you use the former?
[15:17] <robertg> currently i'm using both ways. why?
[15:18] <earl> curiosity :)
[15:19] <earl> i regularly use e.g. `ringo test test/ringo/skin_test.js` to run only the skin tests
[15:19] <earl> and looking at your pull request i wondered if we should update the CLI of test.js as well
[15:19] <earl> it currently is:
[15:19] <earl> Usage: bin/ringo test test/file1 test/file2
[15:22] <earl> but that's really a separate question, i guess
[15:26] <robertg> hmm, i see. this could be a problem, yes
[15:29] <robertg> all testfiles of sqlstore can be executed directly, therefor i never ran single testfiles using "ringo test test/file"
[15:30] <earl> don't think it's a real problem
[15:30] <earl> test.run() for those who want to write their own runners
[15:31] <earl> and the very simplisting "main" of test.js to have one default runner bundled
[15:31] <earl> simplistic*
[15:42] <robertg> yes. the only other possibility i can think of would be to change the CLI of test.js to accept only a single file and an optional method name.
[15:43] <robertg> but i don't think that would be a good move
[17:36] <hannesw_> I like this quite a lot:
[17:36] <hannesw_> app.get("/:id/:action?", function(req, id, action) {...});
[17:36] <hannesw_> action may be undefined, id not