Zotonic
Zotonic
zotonic@conference.zotonic.com
Wednesday, 25 September 2013< ^ >
Arjan has set the subject to: Zotonic - The Erlang Web Framework & CMS
Room Configuration

GMT+2
[00:07:23] Mawuli Adzaku joins the room
[00:09:35] Mawuli Adzaku leaves the room
[00:10:28] Mawuli Adzaku joins the room
[00:12:16] Mawuli Adzaku leaves the room
[00:12:39] Mawuli Adzaku joins the room
[00:15:56] Mawuli Adzaku leaves the room
[00:16:56] Mawuli Adzaku joins the room
[00:21:00] Mawuli Adzaku leaves the room
[00:21:40] Mawuli Adzaku joins the room
[00:26:02] Mawuli Adzaku leaves the room
[00:26:56] Mawuli Adzaku joins the room
[00:32:07] Mawuli Adzaku leaves the room
[00:33:09] Mawuli Adzaku joins the room
[00:36:22] cillian.deroiste leaves the room
[00:36:55] Mawuli Adzaku leaves the room
[00:37:22] Mawuli Adzaku joins the room
[00:37:25] Mawuli Adzaku leaves the room
[04:08:25] Andreas Stenius leaves the room
[08:18:23] arjan joins the room
[08:22:55] arjan leaves the room
[08:50:24] cillian.deroiste joins the room
[08:59:51] arjan joins the room
[09:00:37] arjan leaves the room
[09:01:14] arjan joins the room
[09:25:43] Kaos joins the room
[09:44:52] <arjan> my screen is overflowing with messages like this
[09:45:02] <arjan> 09:44:58.197 [error] webmachine error: path="/comet": {exit,{timeout,{gen_server,call,['z_session_manager$milledoni',{start_session,ensure,"qAYgqQnTazBc7w60fUNU","dvyNDf84scMGWRB54WD4"}]}},[{gen_server,call,2},{z_session_manager,start_session,3},{z_context,ensure_session,1},{z_context,ensure_all,1},{controller_comet,forbidden,2},{webmachine_controller,controller_call,3},{webmachine_decision_core,decision,3},{webmachine_decision_core,handle_request,2}]}
[09:45:11] <arjan> anybody have a clue?
[09:45:17] <arjan> the process is alive
[09:46:45] <arjan> all DB connections say: postgres: zotonic milledoni_live 127.0.0.1(38616) idle in transaction
[09:48:04] <z-bot> [kaos] you can try to pin down the troubling spot with tracing, perhaps..
[09:49:25] <z-bot> [kaos] dbg:tracer(), dbg:p(all,call), then add tracing to the funcs you want to trace: dbg:tp(MOD, []) for all funs in MOD or, dbg:tp(MOD, FUN, []) for a single fun, use tpl instead of tp for local funs
[09:49:34] <arjan> I did a rebuild search queries so the pivot queue was full
[09:49:46] <arjan> I emptied it and restarted, now it is running again
[09:49:55] <arjan> but apparently reindexing all blocks up all database connections
[09:49:57] <z-bot> [kaos] Ah..
[09:50:19] <arjan> might have to do with a non-responsive solr or something
[09:53:33] maas.maarten.zeeman joins the room
[10:06:03] <Marc Worrell> @arjan - reindexing should only use a single db connection - there is only one process doing it
[10:06:13] <arjan> yes but apparently it did not
[10:06:27] <arjan> it blocked all db connections
[10:06:50] <arjan> only when I cleared the pivot table and restarted it was OK again
[10:07:02] <Marc Worrell> that is very strange, we can't have parallel indexers - that is the whole idea of the one indexer process
[10:07:16] <arjan> im afraid to hit the reindex button again
[10:07:19] <arjan> but need to
[10:07:21] <arjan> :P
[10:08:41] <Marc Worrell> ff kijken naar het pivot process
[10:08:43] <z-bot> [kaos] isn't there a "slow" reindex option ? :p
[10:10:11] <Marc Worrell> are you doing anything on the #pivot_done{} notification?
[10:10:20] <arjan> lemme see
[10:10:50] <arjan> eh, yes.
[10:10:51] <Marc Worrell> that one goes async
[10:10:51] <arjan> :P
[10:12:02] <arjan> so better on #custom_pivot{}?
[10:12:04] <z-bot> [kaos] how often is the piviot_done notification sent?
[10:12:04] <arjan> that is a map
[10:12:12] <arjan> once per resource
[10:12:19] <z-bot> [kaos] oh...
[10:13:05] <Marc Worrell> maybe make the #pivot_done{} als synchronous with a map?
[10:13:11] <Marc Worrell> also
[10:13:33] <Marc Worrell> but with a cleaned up context - you might not want to work in the same transaction?
[10:13:38] <arjan> Im doing m_rsc:update calls inside pivot_done
[10:13:46] <arjan> that is maybe not the smartest thing there is to do
[10:14:11] <Marc Worrell> nope :p
[10:14:39] <Marc Worrell> that might trigger another pivot later again?
[10:15:04] <Marc Worrell> (except if you don't "touch" the resource)
[10:15:27] <arjan> yes its not updating in all cases
[10:15:46] <arjan> but still it did m_rsc:p
[10:15:53] <arjan> so hitting the db
[10:17:15] <Marc Worrell> Shall we make the #pivot_done{} a synchronous map? Might be safer....
[10:19:02] <Marc Worrell> Killing IE7 support in mod_admin_frontend…. http://caniuse.com/hashchange
[10:19:26] <Marc Worrell> Use of IE7 is down the drain anyway
[10:41:51] <arjan> the zotonic status site is a bit weird lately
[10:42:12] <arjan> when the site list refreshes it shows like this, http://stick.im/s/3L5
[10:42:35] <arjan> and also the top nav bar is gray
[10:42:51] <arjan> is that on purpose? because the logo is now invisible (white on gray)
[10:43:24] <Kaos> the navbar color changed with the upgrade of bootstrap from black to gray..
[10:44:40] <arjan> ah
[10:44:47] <arjan> that means the logo needs to be inverted
[10:44:53] <arjan> hold on :P
[10:55:10] <arjan> better
[10:55:11] <arjan> http://stick.im/s/IFW
[11:02:17] <arjan> Kaos: bootstrap changes were not backported to 0.9 I hope?
[11:02:21] <arjan> :)
[11:02:37] <z-bot> [kaos] no, I don't think so.. :)
[11:03:42] <arjan> k
[11:04:03] <z-bot> [kaos] this issue, milestone set to 0.10.. https://github.com/zotonic/zotonic/issues/641
[11:05:18] <Kaos> nope, it's only on master :)
[11:31:29] Steffen Hanikel joins the room
[11:34:12] cillian.deroiste leaves the room
[11:34:56] cillian.deroiste joins the room
[14:08:16] Kaos leaves the room
[14:12:29] Kaos joins the room
[14:46:50] <maas.maarten.zeeman> > iolist_to_binary(z_utils:js_object([{is_wtf, false}])).
<<"{is_wtf:true}">>
[14:47:01] <maas.maarten.zeeman> :/
[14:49:45] <arjan> huuuuuuuhh
[14:50:20] <maas.maarten.zeeman> js_prop_value(_, false, _OptContext) -> <<"true">>;
[14:50:25] <maas.maarten.zeeman> :-)
[14:50:30] <maas.maarten.zeeman> oops
[14:50:45] <arjan> iolist_to_binary(mochijson:encode(z_convert:to_json([{is_wtf, false}]))).
[14:50:58] <arjan> I never used that js_object function
[14:51:13] <arjan> (zotonic001@x1)18> iolist_to_binary(mochijson:encode(z_convert:to_json([{is_wtf, false}]))).
<<"{\"is_wtf\":false}">>
[14:51:14] <arjan> :P
[14:51:25] <maas.maarten.zeeman> I need to output js :p
[14:51:36] <arjan> :P
[14:54:34] <maas.maarten.zeeman> One of those moments...
[14:58:58] <maas.maarten.zeeman> Must have been there for ages.
[15:41:56] <arjan> why don't we have the geo coordinate of a rsc in the database as a point column.. grr
[17:35:16] maas.maarten.zeeman leaves the room
[17:54:46] arjan leaves the room
[19:52:22] arjan joins the room
[21:00:08] <Marc Worrell> @arjan just decode the quad-coordinate :p
[21:00:27] <arjan> how efficient is that?
[21:00:31] <arjan> in SQL?
[21:00:40] <arjan> I want to do distance calculations in the db
[21:01:18] <Marc Worrell> oh :p
[21:01:29] <Marc Worrell> then better pivot the two values as well
[21:01:48] <arjan> postgres has a nice native datatype...
[21:01:50] <Marc Worrell> you want to use a psql geo index?
[21:01:58] <arjan> turns out epgsql cannot binary encode a point
[21:02:22] <Marc Worrell> yeah - this quadtile is because we also want to work with sqlite et al
[21:05:54] Steffen Hanikel leaves the room
[21:08:45] Maas joins the room
[21:09:37] <Marc Worrell> maybe you can use a function to encode the binary point? And then inject the textual presentation of the two coordinates?
[22:08:37] Maas leaves the room
[22:36:12] Maas joins the room
[22:44:51] cillian.deroiste leaves the room
[22:48:51] cillian.deroiste joins the room
[23:11:31] Steffen Hanikel joins the room
[23:35:18] Maas leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!