IRC Logs for #circuits Sunday, 2014-03-09

spaceonebut bitbucket doesn't allow me to assign myself00:00
spaceoneprologic: did you mean the commit message of the pull request?00:07
spaceonei didn't want to repeat everything what was already meantioned in the single commits00:08
spaceoneafaics the pull request message is also not shown anywhere00:09
prologicwell the commit message itself00:11
prologicI think Pull Requests are a feature og github/bitbucket00:11
prologicand nothing to do with the scm00:11
prologicalso re bitbucket isssues and assigning yourself00:11
prologiclemme see00:11
prologicshould be able to now00:12
prologicadded you as a developer of the circuits team00:12
spaceonebut what was wrong with the commit messages?00:15
spaceonebtw: the % string interpolation feature is not deprecated. str.format is just an supplement for lack of functionality in %. But % syntax is easier to read and write and should therefore be maintained because there is no need of str.format00:51
prologicI'm not sure if % string interpolation is deprecated or not02:45
prologicbut I honestly prefer .format()02:46
prologicbecause it's more OO/functional02:46
prologicre commit message02:46
prologicjust remember to put in the commit message02:46
prologicFixes Issue #n02:46
prologicor See Issue #n02:46
prologicin case it relates to a documented issue02:46
prologicwhich most commits ideally should :)02:46
prologici.e: bug fixes, new features, etc02:46
spaceonehm, ok. but i will not create issues for trivial things02:52
spaceone(i hate the bitbucket web interface)02:52
prologicbut yeah obviously not for trivial things04:10
prologicI don't :)04:10
*** Myke_ has joined #circuits06:28
prologichi Myke_06:44
Myke_about circuits06:55
Myke_would a DAC + counter be easier to generate a stepped ramp than integrating known area pulses?06:56
Myke_for a crude curve tracer06:56
*** koobs has quit IRC08:36
prologicsorry Myke_08:52
prologicyou want #electronics08:52
prologicthis is about the python application framework "circuits"08:53
prologicperhaps we should move this channel to #pythoncircuits one day :)08:53
spaceoneprologic: core/manager.py08:56
spaceone824:            print("Unhandled ERROR: {0:s}".format(e))08:56
spaceone825:            print(format_exc())08:56
spaceone→ shouldn't this printed to stderr?08:56
spaceoneif so, i will change it08:56
spaceoneand what about core/Debugger? i think this would be good too, if it is printed to stderr08:57
prologicyes please08:59
prologicI complletely agree08:59
prologicthings like this should go to stderr08:59
spaceonei will fix every occurrence08:59
prologicyes same with Debugger08:59
prologicalso did you see my comment son your PR #2?08:59
prologiclet's get rid of that useless "Using circuits-dev serial"08:59
spaceonei already commited that09:00
prologicahh k09:00
spaceonedo i need to make another pull request then?09:00
prologicI think if you git push -f09:00
prologicforce push09:00
prologicthe existing PR should be updated09:00
spaceoneah ok09:00
prologicbut I didn't get a notification about it09:00
spaceonei didn't pushed yet09:00
spaceone1 sec09:00
prologicprobably because it's just the underlying commits that changed09:00
prologicoh :)09:00
prologicyes okay force push then :)09:00
prologicwe'll see if I get a notify about the PR being updated09:01
spaceonecheck ;)09:01
prologicand btw thank you for these fixes/changes/improvements09:01
prologicgreatly appreciated09:01
*** koobs has joined #circuits09:05
prologicokay so I didn't get a notification about the updated PR09:06
prologicbut I merged it anyway09:06
*** Myke_ has quit IRC09:22
spaceoneprologic: i creat PR#309:27
*** FSX has joined #circuits09:50
spaceonei didn't change the print from the tree printer10:26
prologicI've added a Travis-CI config to the project10:32
prologicso hopefully that'll integrate nicely with github's ui10:32
prologicand show up there10:32
prologicI guess we have Travis and Drone as our CI's right now10:33
prologicI'm going to move away from ShiningPanda10:33
prologicas it costs money10:33
prologicwe'll see if my account gets into an "open free" state soon10:33
spaceonegood decicion10:33
spaceoneprologic: where should feature requests be done? in bitbucket?10:34
prologicin terms of documenting them?10:35
prologicas issues?10:35
spaceoneuhm, missing features/functionality10:36
prologicyeah write them up as new issues on bitbucket10:36
prologicso we can discuss/resolve them10:36
prologicI mean we could use github's issue tracker too10:40
prologicbut then we have two places to document issues10:40
prologicand that's a bit of a pain :)10:40
spaceoneor host a real issue tracker → bugzilla10:41
spaceonebut this is currently overhead10:41
prologicbugzilla is overkill10:41
prologicway too complex an issue tracker :)10:42
prologicI like github's and bitbucket's10:42
prologicthey're both good10:42
spaceonebugzilla is very good for large projects10:42
spaceonemaybe one day i will rewrite bugzilla in python having neat features of trac and github10:44
spaceoneprologic: i woul write a new component for it, this is probably easier than adapting the current11:13
spaceoneand probably cleaner11:13
prologicsounds good11:16
prologicthat's precisely what I was thinking too11:16
prologicthe name Socket() for the component is okay11:16
prologicbut it's a little misleading I thikn11:16
prologicbecause has Client and Server base sockets11:16
prologicthen TCP, UDP and UNIX varients11:17
prologicjust voicing my opinion :)11:17
prologicbut you raise a good point11:17
prologicit should be possible to create a generic socket component11:17
spaceoneimho there is no difference between client and server11:17
prologicthat which can then be bound to multiple listening interfaces11:17
prologicand respective events forwarded on11:17
prologicinto a single channel11:17
prologicokay so here's the thing11:18
prologicI realize there isn't much differenc eeither11:18
spaceoneimplementation: should all those single interfaces have an own channel?11:18
prologicbut there are some nice conveniences in the API(s) for Client/Server11:18
prologicso ...11:18
spaceonebe simple is guess11:18
prologicif you implemented a more generic Socket component11:18
prologicwhere respective Client/Server parts were just as simple to use11:18
prologicwe could throw this in to
prologicas the new API(s)/Interface(s)11:19
prologicand deprecate the old ones over time11:19
prologicone thing current socket components in circuits do not have either11:20
prologicin terms of cilent connectivity11:20
prologicis multiple connections11:20
prologicit would be nice to multiplex many connections onto a single component instnace11:20
prologice.g: web crawler11:20
spaceonethe bind format is probably important11:24
prologicmaybe use the same bind format as is there now11:26
prologicwhich supports both ipv4 and v611:26
prologicand seems fairly common11:26
spaceoneplus unix11:27
prologictravis-ci isn't working out so well11:28
spaceoneso def bind(foo, secure=Fase)11:28
spaceoneso def bind(intrerface, secure=Fase)11:28
prologicbind(interface, secure=False)11:29
prologicI think that's okay11:29
prologiccovers most use-cases?11:29
prologicwhere interface is the form11:29
spaceoneinterface is either tuple in ipv4/6 or string with : as ip4/6 or unix file?11:29
prologicfor ipv411:29
prologicand whatever the ipv6 form is11:29
spaceoneyes as i said before11:30
spaceoneprologic: does it make sense to have 2 circuits irc channels?11:30
prologichave a look at the def parse_* methods11:30
spaceone(i would merge #dev into here)11:30
prologicyou should just be able to reuse these11:31
prologicas is11:31
prologicyes I think I'll merge #circuits-dev not here too11:31
prologicgood idea11:31
prologicI'll make a note of it11:31
spaceonethe channel there is commonly unused11:32
spaceonebut contains the important messages11:32
prologicyeah I've gotta bring my ircnotifier bot back online11:33
spaceonei'll look11:47
spaceoneprologic: you're often writing some english words capitalized11:47
spaceonei think this is incorrect11:47
spaceoneotherwise it improoves readability for e.g. germans11:48
prologichabit :)11:49
prologicfeel free to correct11:49
prologicI guess I like emphasizing certain things11:49
spaceonei don't have the time, and it is OK for me :D11:49
spaceoneand you are writing too fast without rereading your texts11:50
prologicyes yes :)11:55
spaceoneprologic: hmm my comments have somehow been removed11:56
prologicI blame it on my lack of sight11:56
prologicI am blind after all :/11:56
prologicyeah kinda weird11:56
prologicbitbucket bug!11:56
prologicI had to review each one via email11:56
spaceonei did some o5ther ones11:56
prologicwhen I updated th PR with new commits11:56
prologicthe comments disappears11:56
prologicI got all 311:56
spaceonei hope ypu will receive them11:56
spaceonei did one more11:56
spaceonedid you gret the "wording sounds odd"?11:57
prologiconly got 311:57
prologicremove the11:57
spaceonehmm :(11:57
prologicused the11:57
prologicwording sounds odd? no11:58
prologicwhich part?11:58
spaceonei will comment again11:58
spaceonedone btw12:00
prologicgot it12:01
prologicI'll merge this in then12:01
spaceoneprologic: can you please reread issue #6512:02
prologicsure one sec12:03
spaceoneissue #60 seems to be ok12:03
prologicwhat about #65?12:04
prologicdid I miss something in the tutorial?12:04
spaceoneuhm sorry s/65/60/g12:04
spaceoneprologic: does circuits support verified ssl connections?12:07
prologicclient-side verification?12:07
prologicor client verification rather12:07
spaceoneclient and server side12:07
spaceoneboth ways12:07
prologicre #60 I think I should split some of the comments out into separate issues there12:07
prologicand just keep the first comment and close it with PE #1112:08
prologicyes afaik circuits supports both12:08
prologicclient and server ssl12:08
prologicand certificate verification12:08
prologicat least there is code for it :)12:08
prologicand a test I think12:08
spaceonedon't trust too much tests.. i will proove that there are many untested things or broken things in circuits.web which can be seen by a simple code review12:09
spaceonebut this is a different project, i need some time to make a writeup12:10
prologicnot all things are covered12:13
prologicthe majority are12:13
spaceonebtw. do you have a public view of the coverage report?12:14
prologicthe filter-attribute of an event handler should be documented. (does it still exists? I remember some event.stop() ideas).12:16
prologicit's .stop()12:16
prologicand is documented in Event.stop()'s docstrings12:16
prologicI will delete this comment?12:16
prologicspaceone, re coverage view - no CI systems I've used do this very well12:17
spaceonehm, wait. this belongs into the tutorial?! this is a mayor question12:17
prologicI will try to regularly publish it somewhere12:17
prologicihmo I don't thinkd we should introduce everything in the tutorial12:18
prologicit'll overwhelm the new user12:18
spaceoneuhm :(12:18
prologicperhaps as a separate issue12:18
prologicbut in the "Manual"?12:18
spaceonemaybe creating a FAQ?12:18
prologican extension of the tutorial?12:18
prologicbut not in a tutorial style12:18
spaceonethe thing is that when you want to start fast you donÄt12:19
prologicThere is the circuits Users Manual12:19
spaceonewant to read the docstrings12:19
prologicI think this belongs there no?12:19
prologicI just think event filtering is something that not many will do12:19
prologicnot new users striaght away anyway12:20
spaceonei did it12:20
prologicyou're a special case? :)12:20
prologicthing is12:20
spaceonein the very first day i wondered how to do this12:20
prologicif we were to add event.stop() in the tutorial12:20
spaceonealso just to know how this can be doned12:20
prologicwe'd have to talk about a whole new section12:20
prologicand add a whole new example12:20
prologicof fireing an event12:20
spaceonethere is a chapter about filtering12:20
prologicand having multiple event handlers lisetn to it12:20
prologicand then somehow stopping the other handlers from executing12:21
prologicwhich in the Dog/Pound style tutorial12:21
prologicwould be hard to write :)12:21
spaceonebut filter=True isn't meationed12:21
spaceonethis was what i inteded with the issue comment12:21
prologicwell filter=True is gone too12:21
prologicit's just event.stop()12:21
prologicso shall we write about this in the manual?12:21
prologicor you still want it in the tutorial (more work)12:21
spaceoneyesterday you told that filter=True still exists12:22
prologicpretty sure I didn't :)12:22
spaceoneand is the preferred way12:22
spaceoneif it doesn't exists then some examples are really broken12:22
prologicsearch for "filter=True"12:23
prologicspaceoneprologic: does filter=True still exists?01:1012:23
prologicspaceoneis it recommended to replace the arg by event.stop?01:1812:23
prologicprologicit was replaced by event.stop()12:23
prologicnot many uses of this left to get rid of12:24
spaceoneprologic: ahh i misunderstood because i thought the answer was to my second question12:26
spaceoneis it recommended?12:27
spaceonefilter replaced event.stop12:27
spaceonethis was my interpretation12:27
spaceonegood to know12:27
prologic*nods* :)12:28
prologicthe new ay is much better and cleaner ihmo12:29
prologicthe old way with sorting via filter=True attirbutes12:29
prologicand returning a True-ish value12:29
prologicwas kinda stupid and sometimes problematic12:29
prologicalso every event in have the same inherited docstring. This should definately be fixed.12:32
prologicThis is untrue?12:32
prologicI see12:33
prologicDoes this occur for all auto generated event subclasses in the API Docs?12:33
prologicseems to only be
spaceonethis is because of imports?12:33
spaceonedunno anymore12:33
spaceoneand there are often newlines missings on various places so that the output is in one line12:34
spaceoneyou need to make 2 newlines to fix it12:35
prologicI've created separate issues for each point in #6012:35
prologicthat wasn't about the changes I made for P# #1112:35
prologicso I'll fix all those separately12:35
prologicI'm using sphinx's apidoc tool to generate thse12:36
prologicbut I'm not 100% happy with it12:36
spaceonei know, it uses RST syntax12:36
spaceonei like it12:36
prologicno I like it too12:37
prologicbut I don't quite like the output it generatoes12:37
prologicit's not perfect :)12:37
prologicbed time12:38
prologicI'm tired, sick with a cold12:39
prologicand still have to drag my sorry ass to work12:39
prologicthis was fun thank you :)12:39
prologiclook forward to more issues resolved and future PR(s)12:39
spaceoneprologic: i created PR#412:45
spaceonedef bind(interface, secure=False) is maybe ok, but the parsing of interface is not ok12:57
spaceone* one cannot differentiate between TCP or UDP12:58
spaceone* one cannot tell if this should be a client or server12:58
spaceonei would btw prefer uri's12:58
spaceone'unix://', 'tcp+ip4://' …12:58
spaceonei have no idea how to conform the client/server thing with the current structure13:14
spaceoneplus we can make shortcuts for the uri ('https://' would be 'tcp+ip4+ssl://$host:443/'13:22
*** koobs has quit IRC13:35
*** Osso has joined #circuits13:55
*** koobs has joined #circuits14:03
spaceonehmm, in the first step it is easier to def __init__(self, client=True)14:12
spaceoneso this cannot conflict14:12
spaceoneit's probably also better in the philosophy of circuits components14:13
spaceoneprologic: can you please set the circuits description in the github title?14:15
*** bleep has joined #circuits16:05
*** FSX has quit IRC16:08
*** zleap has joined #circuits17:30
*** ikarusm has joined #circuits18:27
*** ikarusm has quit IRC18:45
*** flikker has joined #circuits20:01
prologicspaceone, yes ofc20:08
prologicI'll do it when I get into work :)20:08
prologiceating breakfast and getting ready to go :)20:08
*** zleap has quit IRC23:04

Generated by 2.11.0 by Marius Gedminas - find it at!