IRC Logs for #circuits Friday, 2015-06-12

spaceoneokay ^^00:06
riotspaceone: https://github.com/hackerfleet/hfos00:13
riotbut yeah, we should definitely collect these00:13
riotreferences \o/00:13
rioti think i'm going to present circuits to the c-base community during next hack'n'tell00:13
prologicnice! :)00:22
prologiclook forward to that00:22
prologicLet's collect these in an open issue00:23
prologicfor when the site gets rebuilt00:23
prologicmy plan is to rebuild a new modern static site00:23
prologichttps://github.com/circuits/circuits/issues/1300:23
prologichttps://github.com/circuits/circuits/issues/9200:29
spaceoneriot: thanks!00:35
spaceonehmm... i am getting a recursion error :(00:36
spaceoneprologic: components.py contains this:00:38
spaceone120             if v is not self and isinstance(v, BaseComponent) \00:38
spaceone121                     and v not in ('parent', 'root'):00:38
spaceone122                 v.register(self)00:38
spaceonei don't see a way to disable this00:38
spaceonei am storing the parent at my component before registering it (the application does not run yet)00:39
spaceoneas workaround i will set it to None, call __init__() and restore it00:39
prologiceh? say what?00:42
prologicthat's pretty normal to do00:42
prologicand works just fine00:43
prologicwhat are you doing differently?00:43
prologicare you starting a threaded component graph00:43
prologicthen attaching/registering that to another running graph?00:43
prologicthat's invalid and will not work :)00:43
prologicwell it *could* work00:44
prologicbut is gets vastly more complex building a system like that00:44
prologicthreading, thread safety et all00:44
prologicavoid threading :)00:44
spaceonei don't use threads00:48
spaceonewould be nice if methods starting with _ and @property would be ignored for this00:51
prologic_ are ignored defiantely00:52
prologicsorry wait what00:52
prologicwhat are you saying?00:52
prologicmethods perfixed with _ are not decoratoed with @handler00:52
prologicnot sure about properties00:52
prologicwe should fix that if not00:53
spaceonei am not speaking about handlers01:03
spaceonei am speaking about automatic register() calls in __init__01:04
prologicoooh01:07
prologiclet me have a looksie01:07
prologichttps://github.com/circuits/circuits/blob/master/circuits/core/components.py#L11701:09
prologicthis functionality?01:09
prologicand v not in ('parent', 'root'):01:09
prologic^^^ this makes no sense either01:09
prologicv is going to be an instance of BaseComponent01:09
prologicI think what this allows us to do is this:01:10
prologichttps://github.com/circuits/circuits/blob/master/examples/cat.py#L2001:10
*** Osso has quit IRC01:13
spaceoneprologic: ja, but if this is unwanted one cannot disable it01:13
prologicthat is true01:14
prologicwhat's the use-case for not registering the component though?01:14
prologicif you've added it as a class attribute01:14
prologicsurely you'll eventually register it?01:15
spaceonei am registering it by hand01:17
spaceone/manually01:17
spaceonein the other way round01:17
spaceonenot v.register(self) → self.register(v)01:17
spaceoneand when doing this i am ending in some recursion error01:17
prologicahh01:20
prologicthen shoudln't that check be fixed01:20
prologicoh wait01:20
prologicwhen are you registering it?01:20
prologicin the __init__?01:20
spaceonewell i have to do it before __init__01:24
spaceoneuhm, no afterwards01:25
prologichmm01:26
prologicok I'm confused01:26
prologicI don't see the issue anymore01:26
prologicif you registered in the __init__01:27
prologicit *should* be fine01:27
spaceonebecause it would be ignored?01:29
prologichmm01:31
prologiccan you pastebin a clear/concise example of the issue you're running in to?01:32
prologicsorry but I'm at work and getting lost in this :)01:32
prologicseeing code might help clear my head :)01:32
spaceoneprologic: i will create a clear example in a new issue in the next days :)01:34
prologickk :)01:36
prologicsorry!01:36
prologicI think I need more caffeine :)01:36
spaceonecaffeine is in the end not good for the health01:41
spaceoneit generates stress hormones01:42
spaceoneit is better to drink real(!) green-tea01:42
spaceonegreen-tea has a lot of vitamines and antioxidiants, which have the powerfull effect that they bring the body in a positive mood state01:43
spaceonewith a long lasting effect01:43
spaceonewhile coffeine pushs you immediately01:44
prologichaha01:57
prologicI'll keep that in mind :)01:57
spaceonei hope so :D01:57
spaceoneit's better for your health01:57
spaceoneotherwise you will recognize it if you are old01:58
spaceone;)01:58
spaceoneprologic: https://bpaste.net/show/f02955b9fffc → I don't understand this, can you explain what is wrong?02:12
prologicwtf02:14
prologicwhy are you doing02:14
prologicA("A")02:14
prologicand B("B")02:14
prologic?02:14
prologicthis makes no sense02:15
prologicproviding an arbitrary argument should hav eno effect02:15
spaceoneisn't this the channel?02:16
spaceonewell, that was only a fast example.. in my class construction i pass channel='A'02:17
spaceonesame error02:17
prologichttps://gist.github.com/prologic/66cf53e922f4412431ef02:21
prologicwell yeah no channel *has* to be a keyword argument02:21
prologic*args here are ignored02:21
prologicunless your subclass does something with them02:21
prologicbut yeah02:21
prologicI think you've found a bug02:21
prologicalbeit rare :)02:21
prologicnote that .root is defined02:22
prologicbut I wonder when we define .manager02:22
spaceone^^02:22
spaceonehow can i unregister then?02:22
spaceoneroot.unregister(parent)02:22
spaceoneor parent.unregister(child)02:23
spaceonei can't remember those02:23
spaceonedoes it mater where i call unregister?02:23
prologicyes02:23
prologicyou normally call it on the child02:23
prologicas it unregistered from it's "manager"02:23
prologichence I think this is a bug :)02:23
prologicyou register to a manager02:23
prologicand unrgister()02:23
prologic.register(manager)02:24
prologic.unregister()02:24
spaceonedoesn't it have arguments? → nice then02:24
spaceonechild.unregister() removes itself from the parent02:24
spaceone→ that would be the that what i would expect02:24
prologichmm02:25
prologicdo we define .manager anywhere anymore02:25
prologicwe might have changed this02:25
prologicand forgot to update __isub__02:26
prologicand __sub__02:26
prologic?02:26
spaceoneprobably02:26
spaceonei think i remember that change02:26
prologicyeah can you help me find it?02:28
prologicI think I recall such a change too02:28
prologica - b02:28
prologicand02:28
prologica -= b02:28
prologicare quite useful shortcuts02:28
prologicmuch like + and +=02:28
spaceoneyou want to find the git-revision?02:29
prologicyse!02:32
prologicthen file a bug report/issue02:32
prologicand fix this blasted thing :)02:32
prologicbe an easy fix though :)02:32
prologicgood spotting!02:33
spaceonei don't find the commit :/02:33
spaceonebut i will create a issue02:33
prologicfound it :)02:41
prologicfeel free to take this on and take the credit :)02:43
prologicassign it to yourself02:43
prologicnice one :)02:59
spaceoneprobably not the only one :D03:02
prologicperhaps03:03
prologicalso we should think about releasing a new version soon03:03
prologicand then think about splitting up circuits into separate namespaces I think03:03
prologicwe actually have quite high coverage on circuits.core03:03
prologicand circuits.web03:03
prologicbut very little on anything else03:03
spaceonewell, let's wait 3 month until new release03:04
prologickk03:04
prologicsounds good03:04
spaceonei will do a lot of things in the next month03:04
prologicwe've often told our users to use master anyway03:04
prologicand with good reason03:04
*** Osso has joined #circuits08:24
*** Coldblackice has quit IRC09:31
*** Coldblackice has joined #circuits09:41
*** Coldblackice_ has joined #circuits09:43
*** Coldblackice has quit IRC09:47
*** Coldblackice_ has quit IRC09:50
*** Coldblackice has joined #circuits09:52
*** Coldblackice_ has joined #circuits09:53
*** Coldblackice has quit IRC09:57
*** treemo has joined #circuits12:18
riotprologic: i have the testing working. mysteriously. i think i did indeed just have to leave/reenter the virtual environment, after i made hfos a namespace package.13:05
riotoh, wow! no tests but i have 11% coverage! \o/13:07
prologicriot, ahh good o :)13:07
riotyaaay13:07
prologicyeah I thought it was a bit strange things weren't working for you13:07
prologicbut I guess I take my own experience for granted sometimes :)13:07
rioti have coverage reports13:09
prologicwhoohoo :)13:10
riotoh, boy.. testing this will be teh hell13:13
riotsome things at least appear easy.13:13
riotre: py.test problems: "Namespaces are one honking great idea -- let's do more of those!"13:56
riotgrrr ;)13:56
spaceoneriot: hm?13:57
riotspaceone: all good ;) does your py.test stuff work?14:03
spaceoneriot: no :D14:03
riotprologic: can i copy some of your documentation for the developer docs i need?14:03
riotspaceone: your tests didn't find your package, right?14:03
spaceonethey did14:04
spaceonei had other strange errors14:04
spaceonesome attributes do not exists14:04
riot*rereads*14:04
riothmm. was that up ni14:08
riotin the log somewhere.. (ni?!)14:08
riotyaay, i just built the first sphinx doc package.14:29
spaceone^^14:31
spaceoneeasy14:31
riotprologic: do you maintain the changelog manually?15:04
rioti think i'll handle mine with git's precommit hook :>15:09
spaceoneand how do you want to do this?15:09
spaceone/what should the hook do?15:09
riothttps://github.com/MartinSeeler/auto-changelog-hook15:09
spaceonewell, that is very easy afterwards15:10
riotnot sure if this is a good idea...15:10
spaceoneit's just a git log15:11
spaceoneyou still need to filter out unneeded things manually15:11
riotyeah, i just checked the output.. nah15:11
riothmmm.. there must be a better way.15:11
spaceoneit adds noice to commits15:12
riotanother bad thing, exactly15:12
riota commit for every commit.. wah15:12
spaceone^^15:12
riotwell, how to get a busy looking project15:12
riot(look, a tiny robot crawls after each commit and comments it..)15:12
spaceonebest would be to have the discipline to create the entries every week15:14
riota client just added a "publish" flag to his issues ;)15:14
spaceoneso one doesn't forget the details15:14
riotthing is, i have rather tight integration of github issues in my workflow, so i don't want to write nor sort this more than once ;)15:15
riotthis is really helpful: https://waffle.io/hackerfleet/hfos15:16
spaceonenow*15:18
*** treemo has quit IRC16:10
*** green1 has joined #circuits17:19
*** Ueqo has joined #circuits17:57
*** green1 has quit IRC18:00
*** Coldblackice has joined #circuits19:04
*** Coldblackice_ has quit IRC19:06
*** Coldblackice has quit IRC19:08
*** green1 has joined #circuits19:40
*** Ueqo has quit IRC19:40
*** FSX has quit IRC20:07
*** FSX has joined #circuits20:07
*** Coldblackice has joined #circuits20:44
prologicriot> prologic: can i copy some of your documentation for the developer docs i need? yes of course :)21:28
prologicriot> prologic: do you maintain the changelog manually? <-- not really. I use releases21:28
*** green1 has quit IRC22:39
*** Osso has quit IRC22:53

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!