Zotonic
Zotonic
zotonic@conference.zotonic.com
Thursday, 1 November 2012< ^ >
arjan has set the subject to: Zotonic - the Erlang Content Management Framework
Room Configuration

GMT+1
[07:28:26] Ilya Rezvov joins the room
[07:38:54] maas.maarten.zeeman joins the room
[07:40:48] Ilya Rezvov leaves the room
[08:43:49] andreas.stenius joins the room
[08:56:06] Ilya Rezvov joins the room
[08:59:43] arjan joins the room
[09:00:27] <arjan> that as well
[09:21:25] <andreas.stenius> yay, I'm getting my new 240G SSD disc today :)
[09:24:25] <arjan> yeah
[09:24:53] <arjan> I have a 200G
[09:25:06] <arjan> its really a upgrade well worth it
[09:25:29] <andreas.stenius> My 80G is has been pretty filled up for well over a year now...
[09:25:40] <andreas.stenius> yeah
[09:26:28] <andreas.stenius> so it's spilled over to my old 400+ G regular disk, slowing it down every time I need to access something from it...
[09:27:08] <andreas.stenius> even more so, as it almost always has time to fall asleep inbetween too... :p
[10:27:48] <arjan> wow, travis-ci is weird
[10:27:50] <arjan> Executing your script (make all edocs) took longer than 30 minutes and was terminated.
[10:27:53] <arjan> https://travis-ci.org/#!/zotonic/zotonic/jobs/3015524
[10:29:19] <andreas.stenius> but that seems like somehting weird was going on, as the other erl versions built in under 2 minutes...
[10:30:11] <andreas.stenius> ah, yes. looking at where it stopped, it was during submodule update... i.e. a git clone....
[10:30:30] <andreas.stenius> uh, or maybe not...
[10:32:32] <andreas.stenius> nah, I don't know...
[10:32:33] <arjan> network failure?
[10:32:42] <arjan> I dont think it's my latest commit anyway :P
[10:32:59] <arjan> that's what I dont like about CI systems
[10:33:03] <andreas.stenius> well, there has been issues with the git clone previously....
[10:33:23] <arjan> if something in the CI or test system breaks, you think your code is broken
[10:33:30] <andreas.stenius> yeah, I'm sure it's no wrong with the source per se, but the CI system is having build issues...
[10:33:51] <andreas.stenius> it's been flunky for a while...
[10:34:03] <arjan> it should be posible to make a more intelligent CI system
[10:34:12] <andreas.stenius> indeed
[10:34:39] <arjan> or really use the unix exit codes
[10:34:40] <andreas.stenius> at least mark it as a transient error, that may be caused by something else than a real error of the SUT
[10:34:58] <arjan> now, $? != 0 is a build failure
[10:35:04] <arjan> indeed
[11:38:53] andreas.stenius leaves the room
[12:05:18] arjan leaves the room
[12:39:14] <maas.maarten.zeeman> Looking at (custom) event binding with wire. It now works with $(selector).bind(evt_type, handler). We could make it more generic by using the new on method in jquery 1.7. That would allow us to wire elements which are not yet on the page. This is handy for tables which are loaded via ajax and things like that.
[12:55:25] arjan joins the room
[13:04:58] <maas.maarten.zeeman> Hmm, not so easy because you then miss a pickled trigger-id of the element which was *actually* triggered. Leaving room for security glitches.
[13:18:30] <maas.maarten.zeeman> Would be nice for admin pages though as you don't have to register a ton of postbacks.
[13:47:45] <maas.maarten.zeeman> or like the language menu. It would be more compact to have just one wire with a set_language postback on #languages. But then the language code (or other values) somehow must be security passed with the postback. $("#languages").on("a", handler) The handler could retrieve those from the this pointer of course.
[14:24:47] Andreas Stenius joins the room
[14:26:34] <Andreas Stenius> @maas.maarten.zeeman: interesting stuff... :)
[14:27:04] <Andreas Stenius> btw, I'm on Fedora now! wohoo :P
[14:27:31] <Andreas Stenius> (running on my new SSD disc)
[14:27:51] Andreas Stenius leaves the room
[14:28:13] <maas.maarten.zeeman> Cool, haven't used red-hat in a long long time.
[14:29:55] Andreas Stenius joins the room
[14:30:22] Andreas Stenius leaves the room
[14:30:27] <maas.maarten.zeeman> It is a new scomp I think. I'm trying to get something usefule and more generic for custom js events.
[14:31:03] <maas.maarten.zeeman> Some pseudo code what I have so far:
[14:31:06] <maas.maarten.zeeman> <ul id="#languages">
{% on id="languages" trigger="a" postback={set_language} delegate="mod_translation" %}
{% for code, lang in list %}
  <li><a href="#" data-postback="{% pickle code=code %}">{{ lang }}</a></li>
{% endfor %}
</ul>
[14:32:28] andreas.stenius joins the room
[14:33:14] <maas.maarten.zeeman> Maybe easier to use too as you don't have to generate ids and add that to a new wire.
[14:33:44] <maas.maarten.zeeman> Now for custom events:
[14:34:05] <maas.maarten.zeeman> {% on id=" document" type="my.event" postback={my_custom_postback foo=bar} %}
[14:34:16] <maas.maarten.zeeman> $(document).trigger("my.event", {x: 100, y: 100});
[14:34:37] <maas.maarten.zeeman> Then ideally the js values are transferred as json.
[14:35:28] andreas.stenius leaves the room
[14:37:35] andreas.stenius joins the room
[14:38:30] <andreas.stenius> ah, cool! :)
[14:40:39] andreas.stenius leaves the room
[14:40:48] Andreas Stenius joins the room
[14:45:35] Andreas Stenius leaves the room
[14:47:39] andreas.stenius joins the room
[14:48:58] andreas.stenius leaves the room
[14:49:01] Andreas Stenius joins the room
[14:49:12] <Andreas Stenius> hmm... connection seems a bit shaky...
[14:49:22] arjan leaves the room
[14:52:49] arjan joins the room
[14:55:39] Andreas Stenius leaves the room
[14:58:33] <maas.maarten.zeeman> maybe not even a new scomp as there is just one extra attribute... the extra selector.
[14:59:44] andreas.stenius joins the room
[15:05:20] andreas.stenius leaves the room
[15:07:49] andreas.stenius joins the room
[15:15:19] andreas.stenius leaves the room
[15:15:54] <maas.maarten.zeeman> @andreas I see a pattern. Connection is up 5 minutes then drops... ahhh, extended with 2 minutes in the last cycle....
[15:16:24] andreas.stenius joins the room
[15:17:29] <andreas.stenius> could be due to me fiddling about in a lot of dialogs and what not.... but still...
[15:18:00] andreas.stenius is exploring his new environment... :p
[15:18:11] <maas.maarten.zeeman> :-)
[15:25:23] andreas.stenius leaves the room
[15:26:29] andreas.stenius joins the room
[15:30:21] andreas.stenius leaves the room
[15:32:36] andreas.stenius joins the room
[15:35:16] andreas.stenius leaves the room
[15:40:11] <arjan> From a developer's perspective, I would like to see that the interaction between the browser and server becomes more javascript centric
[15:40:33] <arjan> as opposed to the template-tags it needs now
[15:41:42] <arjan> when the logic on the client becomes more complex, I think it is nicer to be able to do all zotonic interactions from javascript
[15:42:03] <arjan> especially as some HTML frameworks are fully JS based
[15:42:46] <arjan> for instance, dojo toolkit, or sencha / ext.js
[15:43:35] <arjan> we can still keep the template tag stuff but let it call the javascript library instead
[15:43:48] <arjan> so that you can also call that library itself directly without needing to write template tags
[16:02:22] <Ilya Rezvov> arjan: +100500 :)
[16:02:31] <arjan> :)
[16:02:51] <Ilya Rezvov> i often think about it
[16:03:27] <arjan> me too
[16:04:06] <arjan> the current JS codebase feels a bit adhoc
[16:04:45] <arjan> not really designed in a proper, JS-native way
[16:04:57] <arjan> for startes, al these global functions...
[16:09:15] <maas.maarten.zeeman> That is easily fixed by wrapping it in an anonymous function closure. Some of the global variables are also private
[16:10:40] <Ilya Rezvov> arjan: i started refactoring js-part already
[16:11:02] <arjan> Ilya Rezvov: how?
[16:11:10] <Ilya Rezvov> it has low modularity
[16:11:19] <arjan> maybe its good we discuss stuff like that first ;-)
[16:11:20] <Ilya Rezvov> arjan: localy :)
[16:11:28] <arjan> if you want it to be accepted :p
[16:11:34] <Ilya Rezvov> i think make proof of concept
[16:11:43] <Ilya Rezvov> and present it for discussion
[16:11:56] <arjan> makes sense
[16:12:10] <Ilya Rezvov> i think it low interesting for all :)
[16:13:07] <Ilya Rezvov> i thinked*
[16:14:21] <arjan> I thnk its very interesting
[16:14:30] <arjan> if you have something, just push it to your github
[16:14:34] <arjan> in a branch
[16:15:21] <Ilya Rezvov> i'll try to formulate my concept the evening
[16:16:18] <arjan> nice
[16:16:33] <Ilya Rezvov> i said some my thoughts to Mark
[16:16:41] <Ilya Rezvov> sorry, Marc*
[16:19:14] andreas.stenius joins the room
[16:20:30] <maas.maarten.zeeman> The template tags are very nice for less technical front-enders. Also security of values already coming from zotonic.
[16:24:41] <Ilya Rezvov> i think two ways will be more flexible
[16:25:27] andreas.stenius leaves the room
[16:25:32] <arjan> I agree with that the tpl tags are nice
[16:25:42] <arjan> but underneath they should just call some Zotonic JS API
[16:25:47] <arjan> which can also be called directly
[16:25:51] andreas.stenius joins the room
[16:27:43] andreas.stenius leaves the room
[16:28:05] <maas.maarten.zeeman> Yep, I agree. the generated wire scripts now also assume $ == jQuery. Should probably also be wrapped in an anonymous function. Something like this. Also allows extending the zotonic api. (function(zotonic) { var $ = zotonic.jQuery; .... })(zotonic);
[16:29:15] andreas.stenius joins the room
[16:30:06] <arjan> anybody of you on mac?
[16:30:14] andreas.stenius leaves the room
[16:30:21] <arjan> can you do "echo $HOSTNAME" in the terminal?
[16:30:39] <maas.maarten.zeeman> yep, mac here
[16:31:07] <maas.maarten.zeeman> $ echo $HOSTNAME
edo-2.fritz.box
[16:31:19] <arjan> ok
[16:31:21] andreas.stenius joins the room
[16:31:23] <arjan> and if you do now:
[16:31:28] <Ilya Rezvov> i on mac
[16:31:28] <arjan> erl -sname foo@$HOSTNAME
[16:31:34] <arjan> it probably errors
[16:32:07] <Ilya Rezvov> $ echo $HOSTNAME
dhcp173-23-red.yandex.net
[16:32:12] <andreas.stenius> arjan, why don't you leave out the @hostname part when using -sname
[16:32:23] <arjan> yes that's what I'm considering
[16:32:51] <maas.maarten.zeeman> yes, that fails
[16:32:55] <arjan> ok
[16:32:56] <maas.maarten.zeeman> can't set short name
[16:33:01] <arjan> yup
[16:33:08] <arjan> I'll fix it like andreas suggested
[16:33:30] <maas.maarten.zeeman> also watch out with the long names... they are case sensitive :-)
[16:33:37] <maas.maarten.zeeman> unlike domain names.
[16:34:21] <arjan> with long names zotonic now always uses the output of hostname -f
[16:35:08] <maas.maarten.zeeman> for erlang foo@test.com <mailto:foo@test.com> and foo@Test.com <mailto:foo@Test.com> are different names. The handshake will fail then.
[16:35:20] andreas.stenius leaves the room
[16:35:56] andreas.stenius joins the room
[16:41:46] <maas.maarten.zeeman> had to start wireshark to actually find out what was wrong.
[16:45:25] andreas.stenius leaves the room
[16:49:31] andreas.stenius joins the room
[16:54:44] <arjan> fix pushed for -sname issue
[16:55:30] andreas.stenius leaves the room
[16:56:05] andreas.stenius joins the room
[16:58:24] <maas.maarten.zeeman> @arjan maybe we should also warn that R15 isn't working
[16:58:34] <maas.maarten.zeeman> pfff
[16:58:34] <arjan> eh
[16:58:44] <arjan> what do you mean
[16:59:10] <maas.maarten.zeeman> I see on zotonic list that he tries to start on os x 10.7.5 with R15B02
[17:00:15] <maas.maarten.zeeman> or was iconv already fixed?
[17:01:08] <maas.maarten.zeeman> aw crap. i will fix that later this evening. first cook food for sick wife and kids.
[17:05:29] andreas.stenius leaves the room
[17:06:38] andreas.stenius joins the room
[17:15:07] <arjan> :)
[17:15:24] andreas.stenius leaves the room
[17:15:34] <arjan> maas.maarten.zeeman: as long as you don't do receiving mail, zotonic will work fine on R15
[17:15:44] <arjan> thats the only iconv use, right?
[17:16:29] <maas.maarten.zeeman> k
[17:16:39] <maas.maarten.zeeman> no issues with compiling then
[17:19:09] <maas.maarten.zeeman> we use iconv a lot as we retrieve quite some data from other servers.
[17:19:41] andreas.stenius joins the room
[17:25:18] andreas.stenius leaves the room
[17:25:30] <maas.maarten.zeeman> eiconv i mean iconv, has a 64 kb limit
[17:26:45] andreas.stenius joins the room
[17:30:15] andreas.stenius leaves the room
[17:31:21] andreas.stenius joins the room
[17:33:56] Ilya Rezvov leaves the room
[17:35:20] andreas.stenius leaves the room
[17:37:03] andreas.stenius joins the room
[17:45:32] andreas.stenius leaves the room
[17:49:56] andreas.stenius joins the room
[17:55:25] andreas.stenius leaves the room
[17:57:42] andreas.stenius joins the room
[18:09:52] <andreas.stenius> ok, I'm now running win 8 on fedora using virtualbox, seems to go rather smooth ;)
[18:15:33] andreas.stenius leaves the room
[18:16:47] andreas.stenius joins the room
[18:30:10] andreas.stenius leaves the room
[18:30:51] andreas.stenius joins the room
[18:35:20] andreas.stenius leaves the room
[18:36:26] andreas.stenius joins the room
[18:36:34] andreas.stenius leaves the room
[18:49:31] andreas.stenius joins the room
[18:50:24] andreas.stenius leaves the room
[18:55:05] andreas.stenius joins the room
[18:55:34] andreas.stenius leaves the room
[18:56:00] andreas.stenius joins the room
[19:05:18] andreas.stenius leaves the room
[19:07:34] andreas.stenius joins the room
[19:15:08] andreas.stenius leaves the room
[19:16:38] andreas.stenius joins the room
[19:24:56] andreas.stenius leaves the room
[19:26:12] andreas.stenius joins the room
[19:26:34] andreas.stenius leaves the room
[19:28:23] Andreas Stenius joins the room
[19:29:00] <Andreas Stenius> ok, see if swift is more stable than the default empathy client...
[19:30:27] Andreas Stenius leaves the room
[19:30:54] Andreas Stenius joins the room
[19:33:56] <Andreas Stenius> nah... maybe it's not on this end... ?
[19:35:21] Andreas Stenius leaves the room
[19:35:55] Andreas Stenius joins the room
[19:45:47] Andreas Stenius leaves the room
[19:45:57] Andreas Stenius joins the room
[19:55:24] Andreas Stenius leaves the room
[19:55:58] Andreas Stenius joins the room
[20:15:56] Andreas Stenius leaves the room
[20:17:00] Andreas Stenius joins the room
[20:23:26] maas.maarten.zeeman leaves the room
[20:24:35] maas.maarten.zeeman joins the room
[20:27:30] Andreas Stenius leaves the room
[20:38:49] Andreas Stenius joins the room
[20:40:48] <Andreas Stenius> hmmm... ok, swift was easy to setup, but was reconnecting every 10 minutes... see how pidgin does it (starting to think there's some other service interrupting my connection...)
[20:45:31] Andreas Stenius leaves the room
[20:47:47] Andreas Stenius joins the room
[20:48:57] Andreas Stenius leaves the room
[20:52:06] Andreas Stenius joins the room
[21:15:36] Andreas Stenius leaves the room
[21:16:55] Andreas Stenius joins the room
[21:25:44] Andreas Stenius leaves the room
[21:26:38] Andreas Stenius joins the room
[21:30:08] Andreas Stenius leaves the room
[21:32:37] Andreas Stenius joins the room
[21:35:36] Andreas Stenius leaves the room
[21:37:35] Andreas Stenius joins the room
[21:45:32] Andreas Stenius leaves the room
[21:47:32] Andreas Stenius joins the room
[21:55:34] Andreas Stenius leaves the room
[21:56:54] Andreas Stenius joins the room
[22:05:44] Andreas Stenius leaves the room
[22:08:43] Andreas Stenius joins the room
[22:15:28] Andreas Stenius leaves the room
[22:16:28] Andreas Stenius joins the room
[22:29:26] Andreas Stenius leaves the room
[22:29:55] Andreas Stenius joins the room
[22:30:20] Andreas Stenius leaves the room
[22:32:26] Andreas Stenius joins the room
[22:33:35] <Andreas Stenius> I'm guessing it's the google talk server that's not really compatible with the jabber clients...
[22:33:37] maas.maarten.zeeman leaves the room
[23:13:27] <Andreas Stenius> weird. the "about this blog" menu of a newly created blog site always points to the current page...
[23:15:16] <Andreas Stenius> hah, even so in the admin...
[23:15:34] <Andreas Stenius> pressing "save & view" simply reloads the edit rsc page...
[23:17:29] <Andreas Stenius> aaahh...
[23:17:38] <Andreas Stenius> 23:17:14.942 [info] [zotonic_dev] warning @ m_rsc:634 Failed to get page url path. Is the `page' dispatch rule missing?
[23:22:52] <Andreas Stenius> what's preferred, adding mod_base_site, or adding the page rule to the blog skeletons dispatch file?
[23:24:41] <Andreas Stenius> ah, I'll add the mod_base_site to the list of enabled modules to the blog skeleton to avoid duplicating the page dispatch rules...
[23:30:25] Andreas Stenius leaves the room
[23:31:35] Andreas Stenius joins the room
[23:31:54] maas.maarten.zeeman joins the room
[23:32:46] <maas.maarten.zeeman> got that one too. didn't investigate fully, but it seemed to have something to do with thumbnails
[23:35:17] Andreas Stenius leaves the room
[23:37:19] andreas.stenius joins the room
[23:37:52] <andreas.stenius> thumbnails?
[23:38:46] <maas.maarten.zeeman> I got a ton of page dispatch rule missing when I opened the admin page. It wants to show little thumbnails.
[23:38:48] <andreas.stenius> no.. in my case, it was the default page dispatch rule that was missing. the other pages on the menu had a page path specified, so they worked. The articles work because they don't rely on the default page rule, but have rules per category
[23:39:01] <maas.maarten.zeeman> hmm.
[23:39:03] <andreas.stenius> oh, ok
[23:39:59] <andreas.stenius> anyhow, I'm off to bed now... g'night ;)
[23:40:09] andreas.stenius leaves the room
[23:40:21] <maas.maarten.zeeman> me too.good night
[23:57:11] maas.maarten.zeeman leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!