Monday, 3 December 2012< ^ >
arjan has set the subject to: Zotonic - the Erlang Content Management Framework
Room Configuration

[09:01:37] Arjan joins the room
[09:05:18] andreas.stenius joins the room
[09:20:54] maas.maarten.zeeman joins the room
[09:56:55] Arjan leaves the room
[09:58:37] Arjan joins the room
[11:43:53] <maas.maarten.zeeman> :q
[11:44:06] <maas.maarten.zeeman> oops. escaped vi command
[11:44:35] <maas.maarten.zeeman> got this: {error_logger,{{2012,12,3},{10,41,46}},"Can't set short node name!\nPlease check your configuration\n",[]}
[11:44:48] <Arjan> euh
[11:44:54] <Arjan> where?
[11:45:01] <maas.maarten.zeeman> when calling zotonic stop
[11:45:31] <maas.maarten.zeeman> because of the shortname change.
[11:45:41] <maas.maarten.zeeman> strange thing is that it started just fine.
[11:45:47] <Arjan> thats weird
[11:46:04] <Arjan> can you put an echo in front of the $ERL of the zotonic-stop script
[11:46:08] <Arjan> see what it tries to do
[11:47:03] <maas.maarten.zeeman> have a separate script to set custom variables. it sets sname and hostname, max ports and max ets tables.
[11:47:59] <maas.maarten.zeeman> I'll play with it locally. This instance has live sites running.
[11:49:12] <Arjan> maybe your custom script needs changing
[11:49:35] <Arjan> although I kept it BC by using the $SNAME variable if set
[11:50:06] <maas.maarten.zeeman> think so. it sets export SNAME="beta"
export HOSTNAME="channel.im"
export MAX_PORTS=5000
export MAX_ETS_TABLES=10000
[11:50:17] <maas.maarten.zeeman> and then calls the zotonic script
[11:51:23] <maas.maarten.zeeman> the hostname should probably be scrapped. will check the other scripts too. Also the heart setup. That can be tricky.
[11:51:43] <Arjan> HOSTNAME is supposed to be set by the OS
[11:52:03] <maas.maarten.zeeman> this is vairtual
[11:52:32] <maas.maarten.zeeman> this is a vps which has a wrong hostname.
[11:52:37] <maas.maarten.zeeman> :-(
[11:54:24] <Arjan> that's fixable right?
[11:54:54] <maas.maarten.zeeman> yes, the hostname is ok now. :-)
[11:55:09] <Arjan> :)
[11:59:43] <maas.maarten.zeeman> aha. i understand. script do this: $ERL $NAME_ARG ${NODENAME}_shell_$$@$HOSTNAME -remsh $NODENAME@$HOSTNAME
[11:59:54] <maas.maarten.zeeman> -remsh is a long name
[12:00:04] <maas.maarten.zeeman> this one was for zotonic shell.
[12:00:47] <Arjan> I dont get it
[12:00:52] <andreas.stenius> me neither...
[12:01:08] <andreas.stenius> :p
[12:03:20] <maas.maarten.zeeman> that is a longname
[12:04:51] <Arjan> you mean the -sname variant should not add @$HOSTNAME?
[12:05:28] <maas.maarten.zeeman> I think so. i'll look at it later.
[12:07:49] <maas.maarten.zeeman> kids taking over my workspace... at home for lunch.
[12:08:11] <andreas.stenius> lol-- kids.
[12:08:39] <andreas.stenius> not much sense for workspace...
[12:35:26] maas.maarten.zeeman leaves the room
[12:40:04] maas.maarten.zeeman joins the room
[12:57:46] <maas.maarten.zeeman> localname setup doesn't work
[12:58:11] <maas.maarten.zeeman> the name of the node is zotonic001@channel
[12:58:31] <maas.maarten.zeeman> scripts try to access zotonic001@channel.im <mailto:zotonic001@channel.im>
[12:58:56] <maas.maarten.zeeman> don't have time right now to look into it though
[12:59:34] <maas.maarten.zeeman> I think this is also wrong: https://github.com/zotonic/zotonic/blob/master/src/scripts/helpers/zotonic_setup#L94
[13:00:06] <maas.maarten.zeeman> if looking at https://github.com/zotonic/zotonic/blob/master/src/scripts/helpers/zotonic_setup#L83
[13:01:27] <andreas.stenius> indeed, looks like it ought to be using the $NODE value
[13:03:07] <maas.maarten.zeeman> yes
[13:04:32] <maas.maarten.zeeman> ill look into it later. no time right now.
[13:08:36] <maas.maarten.zeeman> quickly switched to pre sname version
[13:08:54] <Arjan> maybe make a ticket
[13:09:18] <maas.maarten.zeeman> there are so many way's to configure this.
[13:13:11] <maas.maarten.zeeman> maybe this is nice too https://github.com/ShoreTel-Inc/erld
[14:07:18] maas.maarten.zeeman leaves the room
[14:33:15] maas.maarten.zeeman joins the room
[17:10:45] maas.maarten.zeeman leaves the room
[17:19:43] <Arjan> we're talking why every resource has "address" edit fields
[17:21:52] <Arjan> maybe we need something like traits for categories
[17:21:59] <Arjan> so you can say, this category has this feature
[17:22:09] <Arjan> and the feature is "address edit fields"
[17:22:49] <Arjan> so that behaviuor is not directly coupled to a category like it is now
[17:23:26] <andreas.stenius> well, that sounds like working around the limitation of having a single category per resource...
[17:23:34] <andreas.stenius> but perhaps a good way to do it :)
[17:23:49] <Arjan> traits indeed are a poor man's multiple inheritance
[17:24:08] <Arjan> but multiple inherintance is much harder
[17:24:18] <andreas.stenius> I can imagine
[17:25:45] <andreas.stenius> hmmm.... what if we keep the singleton relationship rsc-category, but let categories have multiple parents... ? is that equally hard...?
[17:27:04] <andreas.stenius> should be possible to work out a single category trail to hide the multi parent scenario from the rest... that only cares if a certain category is in the ancestry...
[17:29:33] <Arjan> but then how do you do catincludes?
[17:29:40] <Arjan> if you get conflicting templates
[17:29:45] <Arjan> which one do you choose
[17:30:03] <Arjan> or access control
[17:30:34] <andreas.stenius> those parts would still work with the "flat" category list
[17:31:10] <Arjan> but then functionality from one can hide it from the other
[17:31:12] <andreas.stenius> so, the trick would be to translate a category hierarchy to a flat version
[17:31:40] <andreas.stenius> isn't it so already?
[17:32:09] <andreas.stenius> I think I fail to see your point
[17:32:32] <Arjan> ok so,
[17:32:53] <Arjan> say you have categories A and B who both define a foo.a.tpl and foo.b.tpl
[17:33:02] <Arjan> category C inherits from both A and B
[17:33:11] <Arjan> and a catinclude "foo.tpl" is requested for a rsc from cat C
[17:33:17] <Arjan> does it show foo.a.tpl or foo.b.tpl?
[17:33:40] <Arjan> not both -- that is "all catinclude"
[17:34:05] <andreas.stenius> ok, I see your conflict.
[17:34:23] <andreas.stenius> but how likely is that? I have to get a example working in my head...
[17:34:28] andreas.stenius is thinking...
[17:34:39] <Arjan> multplie inheritance also does not work with the way we represent the category tree now in SQL
[17:35:01] <Arjan> I think
[17:35:05] <andreas.stenius> yeah, I guess the db would have to adapt...
[17:35:17] <Arjan> you know using right/left nested sets
[17:35:27] <andreas.stenius> yep
[17:35:44] <andreas.stenius> briefly, anyhow
[17:38:19] <andreas.stenius> I'll ponder this some more over dinner... ;)
[17:38:49] <Arjan> :)
[17:39:15] <Arjan> I think it's nice to separate rsc features from categories more
[17:39:37] <Arjan> to be able to have the admin interface contain less parts that resources do not need
[17:39:46] <andreas.stenius> yeah, that could be a good move
[17:40:03] <Arjan> it confuses the hell out of users
[17:40:16] <Arjan> and as it is right now it is quite hard to customize the admin interface
[17:40:23] <Arjan> or actually, not hard, but cumbersome
[18:05:34] Arjan leaves the room
[19:23:29] Arjan joins the room
[20:51:05] Arjan leaves the room
[20:59:14] andreas.stenius leaves the room
[21:23:30] Maas joins the room
[22:10:59] Arjan joins the room
[22:32:09] Arjan leaves the room
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!