Zotonic
Zotonic
zotonic@conference.zotonic.com
Monday, 14 October 2013< ^ >
Arjan has set the subject to: Zotonic - The Erlang Web Framework & CMS
Room Configuration

GMT+2
[00:37:31] Steffen Hanikel leaves the room
[07:24:08] Maas joins the room
[08:04:24] jeff.5nines joins the room
[08:53:57] Maas14916 joins the room
[08:53:58] Maas14916 leaves the room
[08:53:58] Maas66224 joins the room
[08:54:09] Maas leaves the room
[09:11:04] <Maas66224> This looks like a very interesting service. http://www.stathat.com
[09:27:43] arjan joins the room
[10:12:12] jarimatti joins the room
[11:22:18] <Maas66224> Would like to make it easier to run the unit tests. When I now run: eunit:test({dir, "ebin"}) it fails miserably...
[11:22:21] <Maas66224> =======================================================
Failed: 73. Skipped: 0. Passed: 25.
[11:23:19] <Maas66224> Lots of things fail because the trans server is not started. Not that it that is needed to run the test.
[11:25:00] <Maas66224> Thinking about adding a z:mt() function (make and test) which compiles the changed modules with the TEST flag, and then runs all tests it can find.
[11:29:32] <arjan> +1
[11:31:18] <Maas66224> hehe.. shot myself in the foot too many times.
[11:31:43] <jarimatti> "Me too" :)
[11:32:47] <jarimatti> I'd like to see more tests. And yes, if time permits I'll volunteer writing them.
[11:33:59] <Maas66224> Awesome :-) I'll add the function later today..
[11:37:24] Maas66224 leaves the room
[12:16:04] jarimatti leaves the room
[12:23:11] Maas joins the room
[12:38:27] <Maas> more comet woes... 2013-10-14 08:33:12.781 [error] <0.30711.257> application: mochiweb, "Request failed", "flush"
[12:39:09] <Maas> mochiweb request process receives messages from zotonic's controller comet
[12:39:32] <Maas> And then closes the process with a bad request response.
[12:39:52] <Maas> also: 2013-10-14 08:59:25.492 [error] <0.13852.20> application: mochiweb, "Request failed", "{'DOWN',#Ref<0.0.119.87808>,process,<0.23883.20>,normal}"
[12:40:14] <Maas> when the session page process times out.
[13:08:26] Maas leaves the room
[13:10:42] Maas joins the room
[13:14:18] <Maas> smells fishy.
[13:15:29] Maas32428 joins the room
[13:15:29] Maas32428 leaves the room
[13:15:29] Maas32428 joins the room
[13:15:29] Maas32428 leaves the room
[13:15:29] Maas50792 joins the room
[13:15:29] Maas50792 leaves the room
[13:15:29] Maas53991 joins the room
[13:15:29] Maas53991 leaves the room
[13:15:29] Maas53840 joins the room
[13:15:38] Maas leaves the room
[13:18:45] <Maas53840> This is not integrated in a modular fashion... huge race condition.
[13:41:07] <Maas53840> hey @arjan.. do you know why mochiweb_http uses {active, once} instead of calling mochiweb_socket:recv ?
[13:41:43] <arjan> active once = asynchronous receive, right?
[13:41:52] <arjan> gets a packet as a erlang message
[13:42:09] <Maas53840> Otherwise it gets it as an answer from recv
[13:42:31] <Maas53840> it sets active once and then immediately goes to a receive
[13:42:32] <arjan> indeed so the process blocks on :recv
[13:42:36] <arjan> huh
[13:42:50] <Maas53840> now it blocks on receive, but it gets messages from our controllers.
[13:42:54] <arjan> yes ok
[13:43:11] <Maas53840> like flush, and down messages if the page session times out
[13:46:01] <Maas53840> The it returns error 400 for no good reason which messes up the request pipeline.
[13:50:22] <arjan> weird
[13:52:43] <Maas53840> I sometimes get these messages in my log which are from our controllers. They somehow are received in mochiweb_http:request but are totally unrelated to the socket and the current request.
[13:54:50] <Maas53840> Maybe there should not be a receive all pattern in request and headers, only tuples with length 2 or 3 with a port as second parameter.
[14:13:27] <Maas53840> nasty stuff...
[14:15:16] <Maas53840> probably messages from earlier requests.
[14:15:36] <Maas53840> on a keepalive socket...
[14:21:39] <z-bot> [kaos] ouch
[14:46:45] <Maas53840> eyeballing reset_process_dictionary in webmachine_mochiweb.. maybe that is a good place to add a flush of the mailbox of the process too.
[14:48:04] <Maas53840> warning the code there is quite ugly...
[15:11:03] Maas53840 leaves the room
[15:11:22] jeff.5nines leaves the room
[15:52:07] Maas joins the room
[15:57:37] jeff.5nines joins the room
[16:53:44] Maas leaves the room
[16:54:27] Maas joins the room
[17:22:24] jeff.5nines leaves the room
[17:40:00] Maas leaves the room
[17:57:30] Maas joins the room
[18:00:02] <Maas> Why didn't we stumble over this mochiweb/webmachine/zotonic integration problem earlier?
[18:00:26] arjan leaves the room
[19:32:52] Maas leaves the room
[19:51:20] arjan joins the room
[19:59:08] Maas joins the room
[21:08:13] arjan leaves the room
[21:17:07] <Maas> wtf, 204 responses closes the socket... that is not right too.
[21:42:58] Maas leaves the room
[21:55:48] Maas joins the room
[22:07:26] arjan joins the room
[22:07:55] arjan leaves the room
[23:09:05] Maas leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!