IRC Logs for #circuits Friday, 2013-12-20

*** eriknw has joined #circuits05:17
*** anunnaki has joined #circuits05:22
prologichi eriknw  :)05:31
prologicjust saw your email05:31
prologichi :)05:31
prologicI'll reply here quickly (wife should be picking me up soon!)05:31
prologicyes holidays are in full swing05:31
prologicI'm on leave as of (as soon as I leave here)05:31
prologicknow what that's like re working an extra work after finishing up somewhere - been there once or twice :/ not fun :)05:32
prologicI plan to work on circuits and get 3.0 out the door over the next 3 weeks (before I go back to work next year)05:32
prologicas well as more mio-lang work05:32
prologicand start a couple of new projects as wlel!05:32
eriknwnew projects, eh?05:33
prologicyeah maybe - at least one05:33
prologicI want to clone one of our projects here at work (written in Plone) in a different way (mostly just to prove a point!)05:33
eriknwheh, and working a few extra days was actually kind of interesting, because every day was taken to be my last day, so people would come by and say stuff like "well, in case I don't see you tomorrow, I just want to let you know..."05:33
eriknwproving a point is a perfectly valid reason!05:34
eriknwI worked on a new project for a couple hours today too05:34
eriknwwell, it's not completely new; I gave a lightning talk in July about it a day after I came up with the idea05:35
prologicahh nice05:35
prologicyeah I've been taking a page out of your book05:35
prologicand reserving pypi names :)05:36
eriknwheh, nice.  `eqpy` is mine!05:36
prologicso I have a few new ideas for projects up my sleeve as well that may come to fruition05:36
eriknwso are `sampling` and `metafunc` :)05:36
eriknwalthough I'm thinking of changing `metafunc` to `importz`05:36
prologicI've got bus, msgbus, protocols, object05:37
prologicand quite a few others :)05:37
*** anunnaki has quit IRC05:37
*** anunnaki has joined #circuits05:37
prologicand reuse!05:37
eriknw`eqpy` is a small project (I tend to like narrow scopes :-) ) that makes it easy to manage and solve systems of equations (linear or non-linear) symbolically  or numerically using sympy05:38
prologicanyway I look forward to working with you on mio soon :005:38
prologichopefully :)05:38
eriknwaha, excellent names you have reserved there05:38
prologicI could do with some help with some of the sematics05:38
prologicre eqpy cool sounds nice :)05:38
eriknwaye, sounds good to me05:38
prologicI think I've decided that we should just implement the python calling semantics of functions/methods05:39
prologicso that we get a nice *args, **kwargs, etc working nicely05:39
prologichowever you want to generalize the rules as such :)05:39
prologicand continue to have the unevaluated arguments in ``call message args``05:40
eriknwah, keeping things familiar to python users doesn't seem like such a bad thing, although I don't have a fully developed opinion on this yet05:40
prologicthere are many aspects of Python I do quite like05:41
prologicso yeah Python will probably be a heavy influencing factor I think05:41
prologicbut that's not set in stone obviously :)05:42
prologicyou find inspiration everywhere! and in any language :)05:42
eriknwsyntax-wise, I prefer Lua.  Library-wise, I much prefer Python05:42
prologicI haven't much practical experience with Lua I'm afraid05:42
prologicbut between you and I I think we can design a nice lang :)05:42
eriknwI bet so!05:42
prologicalso NB; the whole **kwargs and keyword arguments concept in mio05:43
prologicisn't actually part of the grammar at all05:43
prologicit's actually syntactic sugar on top of mio itself05:43
prologicas part of the bootstrap std. lib05:43
prologicso in reality there are only args05:43
prologicnice property of homioiconic languages like mio, io, smalltalk, (and others?)05:43
prologicI was actually thinking of re-implementing Block itself in mio anyway05:44
prologicor at least trying to see if it can be done05:44
eriknwoh boy, I wasn't prepared to jump into this tonight05:44
eriknwI just wanted to say hello then goodnight :)05:44
prologicand then re-implement the block/method blocks (functions) as well05:44
prologicand move them out of the Object proto05:45
prologicsorry :)05:45
prologicit's late in the afternooon05:45
prologicI'm on holidays in 5 4 3 205:45
prologicnearly :)05:45
prologicand I don't give two hoots about work right now :)05:45
prologichi and g'night then :)05:45
eriknwsweet!  and, yeah, our timezones are way off.  it's nearly 1am here05:45
prologicwe seem to be in differing timezones :)05:45
prologicyouch :)05:45
prologicgo to bed :)05:46
prologicnice chatting and meeting (sort of)05:46
prologiccya around :)05:46
eriknwwill do, and we'll continue this later!05:46
*** SX has joined #circuits06:56
prologichey SX06:59
*** SX has quit IRC07:32
*** qwebirc26798 has joined #circuits08:04
prologichi qwebirc2679808:05
prologicqwebirc26798, see the irc logs for my response to your query08:05
prologicI think you went offline and didn't see it :)08:05
prologicjust been playing with cocos2d08:06
prologicI like it :)08:06
prologicI'm going to build a event source driver for pyglet/cocos2d :)08:06
prologicfor circuits taht is08:06
qwebirc26798Yes, I have seen it08:07
qwebirc26798right now im studying the io.serial code. It's not 100% windows compatible... yet08:08
prologicI suspected it not to be :)08:09
prologicand as I've mentioned before here08:09
prologicto the other guys08:09
prologicit's really really hard to unit test08:09
prologicso there's no unit test for it08:09
prologicif you feel up to it, please write one :)08:09
prologicI would mock out the pyserial library part08:09
prologicand just test the circuits api side08:10
prologicI would also use the pytest monkeypatch fixture08:10
qwebirc26798Well till now I just work with pyserial, worked quite well, but im note sure how it performance with higher bandwith08:13
qwebirc26798I suspect you are an electrical engineer as well or why did you choose circuits as the name for this lib? It's sometime hard to google it because of the name conflict :)08:16
prologicahh yes08:16
prologicjust google python + circuits08:16
prologicor better yet08:16
prologicpython circuits framework08:16
prologicit's not impossible to google08:16
prologicbut then again it's not like I've done any marketting08:16
prologicor link spamming :)08:17
prologicin any case08:17
prologichopefully we'll see more traction and exposure from next year onwards08:17
prologicwith contintious releases08:17
prologicand I'll try to be more active on SO with an occainsally answer like:08:17
prologic"Here's how you do it in circuits"08:17
prologicre name choice08:17
prologicthe reason for circuits08:17
prologicis because that's the conotation08:17
prologicbuilding apps in circuits is like plugging together pieces of electrnoic components in a circuit board08:18
prologicapp = (A() + B() + C())08:18
prologicthe idea is that everything in your app is a small components08:18
prologicand larger functionality is horizontally made up of smaller components08:18
qwebirc26798yes this app = (A() + B() + C()) ... is quite clear to understand and clean to implement08:20
prologicwell it is when you understand why08:22
prologicand how08:22
prologica) it's a shortcut and convenience08:22
prologicb) that's the design and architecture of circuits08:22
prologicapp = (A() + B() + C()).run()08:22
prologic(A() + B() + C()).run()08:22
prologicis equivilent to:08:22
prologica = A()08:23
prologicb = B()08:23
prologicc = C()08:23
qwebirc26798i just started to programm an app which acquires data form the serial port, logging them, do some numpy/scipy magic, plot them08:23
prologicand so on08:23
prologicshould be able to do that all nicely with circuits08:23
qwebirc26798with your framework it should extandable easily08:25
qwebirc26798do you have some examples with threading and/or multiprocessing?08:30
qwebirc26798i like to distribute the numpy number crunching to different processes08:31
qwebirc26798I'm back soon08:33
prologicqwebirc26798, yes08:35
prologicsee the examples/ example08:36
prologicI wrote that for lit*08:36
prologicuses the Worker component08:36
prologicwhich itself wraps multiprocessing08:36
*** Ossoleil has joined #circuits08:47
prologicheya Ossoleil08:56
qwebirc26798prologic, thanks i will look into the examples/factorial.py09:22
*** SX has joined #circuits09:36
qwebirc26798prologic, do you have an example how to use/integrate tick() in a GUI event loop (GTK,...)?13:06
*** SX has quit IRC13:22
*** qwebirc26798 has quit IRC17:38
*** Ossoleil has quit IRC18:52
prologicqwebirc26798> prologic, do you have an example how to use/integrate tick() in a GUI event loop (GTK,...)? <-- not tick()20:40
prologicthat's the old 2.1.0 way20:40
prologicwrite a generate_events event handler20:41
prologicsee circuits/core/timers.py20:41
*** Osso has joined #circuits22:48
prologicheya Osso22:58
prologicI assume you have a bit of time :)23:01
prologicwant to quickly discuss namespace packages again23:01
prologicI think we really should do this for circuits23:01
prologicnamely because some things like e.g:23:02
prologicreally should exist in it's own package23:02
prologicand I have an idea for how we can maintain good versions overall23:02
Ossook tell me more23:23
prologicso basically23:27
prologiccircuits itself becomes a namespace package23:28
prologicwe move the __date__, __author__ and __version__ to circuits.core23:28
prologica sub-package23:28
prologicand everything in the circuits-dev repo becomes it's own package23:28
prologicand published separately on pypi23:28
prologic-but- the top-level pulls all of what is currently circuits in23:28
prologicso there is no brekage with circuits 3.0.0.dev23:28
prologicnow for the versioning23:28
prologicwe build a version management such that the version of the top-level circuits23:29
prologicis comprised of all the sub-package we (the core devs) maintain23:29
prologicso basically it would currently be this:23:29
prologiccircuits.core.__version_info__ = (3, 0, 0, "dev")23:29
prologicand all sub-packages would be23:29
prologic__version_info__ = (0, 0, 0)23:30
prologicso that the sum of all sub-packages would equla 3.0.0.dev23:30
prologicalso I would move out of the core circutis23:30
prologicsince it has an external dep23:30
prologicand as well23:30
prologicbut everything else remains in the same repo23:30
prologicthis would also give us and others the ability to add to the circuits namespace reasily23:31

Generated by 2.11.0 by Marius Gedminas - find it at!