Zotonic
Zotonic
zotonic@conference.zotonic.com
Monday, 10 December 2012< ^ >
arjan has set the subject to: Zotonic - the Erlang Content Management Framework
Room Configuration

GMT+1
[01:16:50] maas.maarten.zeeman leaves the room
[07:58:05] maas.maarten.zeeman joins the room
[07:59:57] maas.maarten.zeeman leaves the room
[08:02:58] maas.maarten.zeeman joins the room
[08:54:56] Arjan joins the room
[09:01:23] andreas.stenius joins the room
[09:37:15] <Arjan> morning
[09:51:27] <andreas.stenius> mornin' :)
[09:55:09] <Arjan> nice research on #308 :)
[09:59:58] <andreas.stenius> :)
[10:00:57] <andreas.stenius> It looked so strange, but now that I've grasped the concept, I appreciate the cleverness of it... :)
[10:01:48] <andreas.stenius> and I think that if we only use some other fallback during the rare circumstances that calls for repivoting the cat numbers, it's gonna be great
[10:02:18] <Arjan> yes indeed
[10:02:21] <andreas.stenius> going through old closed issues that has no assigned milestone...
[10:02:25] <Arjan> :)
[10:02:35] <Arjan> I just reopened one :p
[10:03:07] <andreas.stenius> found it strange when I looked at 0.9 and there were so few issues assigned to it (in total)...
[10:03:13] <andreas.stenius> aha :)
[10:04:10] <andreas.stenius> that was an unexpected bonus ;)
[10:04:42] <andreas.stenius> was so close to just dropping the label... *phew* :p
[10:06:16] <Arjan> heh
[10:06:30] <Arjan> its a weird issue though, hard to reproduce etc.
[10:12:02] <andreas.stenius> ok to rename mileston 0.8 to 0.8.x? or should we add extra milestones for each 0.8 release?
[10:12:21] <Arjan> renaming is ok
[10:12:31] <andreas.stenius> ok, good.
[10:12:38] <Arjan> however I'm not sure we're gonna do many 0.8 fixes once 0.9 is out?
[10:12:53] <andreas.stenius> no, it's for having a place to put old issues...
[10:12:58] <andreas.stenius> closed ones
[10:13:01] <Arjan> ok
[10:13:17] <andreas.stenius> and 0.8.x indicates more than one actual release...
[10:13:33] <andreas.stenius> more than just 0.8
[10:13:35] <Arjan> although this one is still relevant: https://github.com/zotonic/zotonic/issues/368
[10:14:33] <andreas.stenius> yeah, maybe create new milestones for not yet released versions, and only have the 0.8.x for older versions which didn't have milestones at the time of release...
[10:14:58] <andreas.stenius> so we can have a 0.8.3 milestone that has this isse
[10:15:11] <andreas.stenius> bah, I type like a duck
[10:15:27] andreas.stenius has cold fingers...
[10:15:32] <Arjan> :)
[10:15:59] <Arjan> I'll move 368 to 0.9.1
[10:16:02] <maas.maarten.zeeman> yup, working on that one.
[10:16:08] <Arjan> with the note that it need backporting
[10:16:11] <Arjan> maas.maarten.zeeman: awesome
[10:16:18] <andreas.stenius> ok, cool.
[10:16:24] <maas.maarten.zeeman> prob needs replacement of data bus though
[10:16:34] <Arjan> whoops :P
[10:16:38] <Arjan> no simple solution? :)
[10:16:48] <andreas.stenius> data bus?
[10:17:01] <maas.maarten.zeeman> well the page_session which is used as data bus
[10:17:08] <andreas.stenius> oh, ok
[10:17:18] <maas.maarten.zeeman> page-session dead... no more data
[10:17:23] <Arjan> indeed
[10:17:26] <andreas.stenius> sounds like quite a bit of work, then..
[10:18:21] <maas.maarten.zeeman> yes some internals have to change a bit.
[10:18:32] <maas.maarten.zeeman> working on it for channel
[10:18:51] <andreas.stenius> puh, /only/ 283 issues left to review :p
[10:19:04] <Arjan> gee
[10:19:16] <Arjan> 283??
[10:19:20] <andreas.stenius> I wont do them all...
[10:19:23] <Arjan> are you going over the closed ones?
[10:19:30] <andreas.stenius> yeah, looking at ... yep
[10:19:31] <maas.maarten.zeeman> on channel the page can now survive server reboots
[10:19:39] <andreas.stenius> cool
[10:19:48] <Arjan> nice
[10:20:24] <maas.maarten.zeeman> have to integrate it with the postback ws though. we now use two websockets.
[10:20:48] <andreas.stenius> that's how I found #309
[10:58:17] <andreas.stenius> added ~30 closed issues to the 0.9.0 milestone... and I think that was most of them. the remaining 220 or so closed issues not assigned to any milestone is either 0.8.x or earlier...
[10:59:09] <andreas.stenius> we still have 57 open issues not assigned to any milestone.
[10:59:53] <andreas.stenius> Arjan: the 0.9.1 issue, is that a reminder for self about making the release.. ?
[11:09:03] <Arjan> which one?
[11:09:06] <Arjan> you mean the milestone?
[11:16:51] <andreas.stenius> no, the issue #469
[11:17:11] <andreas.stenius> the milestone is a given, to assign other issues to, of course
[11:17:35] <Arjan> oh lol no
[11:17:41] <andreas.stenius> also, how come the 0.7.2 - 0.7.5 were never tagged?
[11:17:45] <Arjan> I think I tried to create a milestone but got confused
[11:17:55] <andreas.stenius> ok, so we can delete that one?
[11:18:15] <Arjan> done
[11:19:19] <andreas.stenius> :)
[11:20:27] <Arjan> implemented category thing
[11:20:38] <Arjan> did not expect it to be so inefficient as it apparently is..!
[11:20:46] <andreas.stenius> nice :)
[11:20:49] <andreas.stenius> is it?
[11:20:50] <Arjan> oh no wait
[11:20:53] <Arjan> mistake
[11:34:40] <andreas.stenius> wow, I had quite a few pull requests a year ago... (more than I remembered them to be.. ;)
[11:40:31] <Arjan> hehe nice
[11:40:47] <Arjan> good that the tickets are being cleaned up! :)
[11:40:57] <Arjan> that way we can have a nice overview with all the bugs fixed for the release
[11:41:15] <andreas.stenius> yeah, doing some house work once in a while...
[11:41:34] <Arjan> I'm almost finished with https://github.com/zotonic/zotonic/issues/308 :)
[11:41:46] <andreas.stenius> it'll get better now that the procedures are settling a bit, I think
[11:41:54] <Arjan> yep
[11:48:07] <Arjan> category renumbering is also quite slow because it only updates 500 resources per 10 seconds
[11:49:00] <Arjan> I'll increase that to 1000 resources per 10 seconds
[11:49:03] <andreas.stenius> ah, right... that's 3000 / minute, 180'000 / hour... yeah with many resources...
[11:49:09] <Arjan> indeed
[11:49:31] <Arjan> but now that 308 is fixed it doesnt matter so much anymore ;)
[11:49:38] <andreas.stenius> no
[11:49:39] <Arjan> oh wait I have to commit it :p
[11:49:53] <andreas.stenius> but how long does it take to update 500 (or 1000) resources?
[11:50:01] <andreas.stenius> lol
[11:50:24] <andreas.stenius> I mean, wouldn't it be nice to tune it so it's more or less evenly spread out... ?
[11:50:51] <Arjan> true but we want not to overload the db while its doing it
[11:50:55] <Arjan> e.g. give other queries also some time
[11:51:08] <Arjan> I tried with 5000 per 10 seconds
[11:51:19] <Arjan> then the db was pretty much busy all the time
[11:51:32] <andreas.stenius> yeah, that's not too good
[11:51:36] <Arjan> so 1000 seems right
[11:51:39] <andreas.stenius> ok
[11:51:42] <Arjan> :)
[11:52:08] <andreas.stenius> felt like an arbitrary raise, but I see you did your homework ;)
[11:54:48] <Arjan> :)
[11:55:02] <Arjan> yep I'm also not too fond of magic numbers
[11:56:02] <andreas.stenius> there should be a NICE argument you could pass in with the SQL query...
[11:58:38] <andreas.stenius> ah, cool, didn't think of putting the dirty flag in the config. that way one can force it to run with a dirty tree to test responsiveness etc... nice :)
[11:59:03] <andreas.stenius> (without going through too much trouble)
[12:00:48] <andreas.stenius> ack --erlang pivot_catgory_nr shows a few more spots that might benefit from looking at the dirty flag...
[12:01:14] <andreas.stenius> most notable last_modified in m_category.erl
[12:03:07] <Arjan> hmmm
[12:05:05] <Arjan> ack is nice, isn't it? :)
[12:05:35] <andreas.stenius> yeah, it's great :)
[12:05:50] <andreas.stenius> nice colorful output, too
[12:05:50] <Arjan> so much faster than grep
[12:06:13] <Arjan> I also have these in my ackrc:
[12:06:15] <Arjan> --type-set=tpl=.tpl
--type-set=rst=.rst
[12:06:28] <andreas.stenius> no more find . -name \*.erl | xargs grep foo :p
[12:06:50] <andreas.stenius> ah, nice
[12:06:52] <Arjan> hehe
[12:06:59] <Arjan> ack --tpl blah
[12:10:47] <andreas.stenius> that's really nice... grepping templates has been a pain
[12:21:41] <maas.maarten.zeeman> find . -name ".tpl" | xargs grep blah
[12:23:22] <Arjan> maas.maarten.zeeman: we're discussing http://betterthangrep.com/
[12:23:49] <Arjan> you'll like it
[12:23:51] <Arjan> :p
[12:24:07] <maas.maarten.zeeman> grep and find combined
[12:25:02] <Arjan> saves keystrokes :)
[12:29:32] <maas.maarten.zeeman> alias?
[12:30:24] <Arjan> it is a perl script
[12:33:07] <maas.maarten.zeeman> alias grep-erl='find . -name "*.erl" | xargs grep $1'
[12:33:11] <maas.maarten.zeeman> ?
[12:33:21] <maas.maarten.zeeman> :-p
[12:34:09] <Arjan> http://betterthangrep.com/documentation/ :p
[12:37:53] <maas.maarten.zeeman> o, just change $1 with $* then ;-) ...
[12:38:09] <maas.maarten.zeeman> no, looks like a nice tool though
[12:38:12] <maas.maarten.zeeman> :-)
[12:38:34] <Arjan> surely
[12:38:46] <maas.maarten.zeeman> have "just" been grepping for quite some time now..
[12:40:06] <Arjan> me too but this tool got me to switch my habits
[12:44:23] <Marc Worrell> Yeah we are getting rich!
[12:44:23] <Marc Worrell> Revenue last month: €0.20
Total available revenue: €30.77
Full report: https://flattr.com/transaction/view/id/1272756 <https://flattr.com/transaction/view/id/1272756?utm_source=email&amp;utm_campaign=revenue-report>
[12:45:32] <Arjan> wow
[12:45:35] <Arjan> :p
[12:53:15] <maas.maarten.zeeman> Niet vergeten op te geven bij de belastingdientst :-p
[12:54:01] <Marc Worrell> Jaja, dat wordt een groot probleem :p
[12:55:39] <Arjan> andreas.stenius: can you remove http://kaos.github.com/zotonic-docs/ <http://kaos.github.com/zotonic-docs/#>
[12:55:44] <Arjan> it is googleable
[12:55:49] <Arjan> :)
[12:55:58] <Marc Worrell> or set a noindex :)
[12:56:16] <Arjan> yep
[14:54:49] <maas.maarten.zeeman> oops... zotonic stuck in a restart loop. zotonic_sup keeps trying to call z_dispather$sitename, dispatch info, and that process is not started.
[14:58:14] <maas.maarten.zeeman> all started due to a timeout on the module indexer which caused a restart. The shutdown failed for some reason.
[14:58:24] <maas.maarten.zeeman> Slow machine.
[15:10:41] <maas.maarten.zeeman> @arjan. Ack is sure nice, but it seems to cache results and sometimes returns old answers
[15:15:29] <Arjan> oh? weird
[15:15:32] <Arjan> never got issues with that
[15:16:23] <maas.maarten.zeeman> probably a time sync thing. have share mounted via ssh
[15:17:34] <maas.maarten.zeeman> maybe the dates wrong so it doesn't see the file has changed.
[15:27:31] Arjan leaves the room
[15:36:28] Arjan joins the room
[15:37:00] <andreas.stenius> Arjan: removed zotonic-docs. didn't find any option to set seonoindex... and I don't care for the repo any longer.
[15:37:08] <Arjan> k
[15:38:04] <andreas.stenius> oh, right. it was for the gh-pages branch... a bit slow on the up-take.. been out a few hours...
[15:38:12] <andreas.stenius> but it doesn't matter ;p
[16:25:04] <maas.maarten.zeeman> ow cool. just discovered a new zotonic trick
[16:25:43] <maas.maarten.zeeman> ext_...
[16:26:59] <Arjan> :)
[16:27:26] <Arjan> new rule proposal: the discoverer must document the features he finds :p
[16:28:18] <Arjan> but I'll do it
[16:28:19] <maas.maarten.zeeman> haha
[16:28:25] <Arjan> I added the feature about a year ago
[16:28:35] <Arjan> its quite a small thing
[16:28:36] <Arjan> wondering where to put it
[16:28:43] <Arjan> where in the doc
[16:29:02] <maas.maarten.zeeman> was checking zotonic_sup. it was looping like crazy
[16:29:10] <Arjan> mm
[16:30:30] <maas.maarten.zeeman> A shutdown went wrong in a dispatcher somewhere. It somehow went into a state from which it couldn't recover on its own
[16:30:52] <andreas.stenius> ext_?
[16:31:30] <maas.maarten.zeeman> That allows you to add non-standard children to zotonic_sup
[16:33:08] <maas.maarten.zeeman> The person who is going to need that probably knows what to do ;-)
[16:33:09] <andreas.stenius> aha :)
[16:36:48] <maas.maarten.zeeman> weird loop.
[16:37:23] <maas.maarten.zeeman> Looks like a problem with the state of z_sites_dispatcher.
[16:37:42] <maas.maarten.zeeman> it is never restarted by zotonic_sup
[16:38:56] <Arjan> does it crash?
[16:39:01] <Arjan> zotonic_sup is a regular supervisor right
[16:39:45] <maas.maarten.zeeman> yes indeed. can't see why z_sites_dispacher is never restarted when it is obviously having a problem.
[16:40:25] <maas.maarten.zeeman> bla looking at the wrong pid
[16:40:47] <maas.maarten.zeeman> it is restarted.
[16:43:49] <Arjan> ah
[16:44:18] <maas.maarten.zeeman> why couldn't it get the dispachinfo of a site then?
[16:44:22] <maas.maarten.zeeman> on every restart
[16:44:30] <maas.maarten.zeeman> more code digging
[16:46:08] <maas.maarten.zeeman> it failed in fetch_dispatchinfo because the dispatcher of the site wasn't running.
[17:00:24] <maas.maarten.zeeman> hmm, looks like z_sites_manager somehow thought the site was running while it was actually not.
[17:07:05] <andreas.stenius> pushed tags for release-0.7.2 -- .5
[17:14:03] <Arjan> andreas.stenius: nice :)
[17:18:21] <maas.maarten.zeeman> zotonic_sup just keeps running. one_for_one, 1000, 10
[17:19:57] <maas.maarten.zeeman> Not that is not something my netbook is going to achieve. :-)
[17:20:41] <maas.maarten.zeeman> once the error get's that high, maybe everything should be restarted.
[17:20:46] <maas.maarten.zeeman> one_for_all?
[17:39:27] Arjan leaves the room
[17:41:59] Arjan joins the room
[17:55:07] <maas.maarten.zeeman> Try running a couple of times. With this I can now get zotonic in restart loop. erlang:exit(whereis(z_sites_manager), kill), erlang:exit(whereis(z_sites_dispatcher), kill).
[17:58:28] <maas.maarten.zeeman> (a different one)
[18:05:05] <maas.maarten.zeeman> changing zotonic_sup into a one_for_all fixes that
[18:06:15] <maas.maarten.zeeman> maybe the site dispatcher and manager, which depend on each other, should be started under a one_for_all supervisor together.
[18:06:49] <maas.maarten.zeeman> leaving the email and media out.
[18:08:47] andreas.stenius leaves the room
[18:10:49] andreas.stenius joins the room
[18:40:11] Arjan leaves the room
[18:45:45] Arjan joins the room
[19:02:33] Arjan leaves the room
[19:17:13] Arjan joins the room
[19:27:20] Arjan leaves the room
[20:11:30] Arjan joins the room
[21:04:52] <Arjan> andreas.stenius: what do you mean?
[21:05:15] <andreas.stenius> hang on...
[21:05:38] <andreas.stenius> go to http://zotonic.com/edoc/modules (make sure there's no trailing slash)
[21:05:47] <andreas.stenius> try a link to a modules
[21:05:51] <Arjan> aah, a slash
[21:06:07] <Arjan> hmm
[21:06:12] <andreas.stenius> ah, did you get the mail notification, were my trailing slash was missing?
[21:06:19] <andreas.stenius> I added it the second after posting...
[21:06:22] <Arjan> :)
[21:06:25] <Arjan> yes via email
[21:06:41] <Arjan> i would actually consider this a bug in controller_file_readonly
[21:06:46] <andreas.stenius> github should hold on to the notification a minute.. :p
[21:06:55] <andreas.stenius> I considered that too...
[21:06:59] <Arjan> where it serves a directory index, it should ensure a trailing /
[21:07:09] <andreas.stenius> yep
[21:20:25] <Arjan> fixed
[21:27:34] Arjan leaves the room
[21:42:09] <andreas.stenius> nice :)
[22:30:35] andreas.stenius leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!