IRC Logs for #circuits Tuesday, 2013-03-26

*** Osso has joined #circuits03:52
*** ronny_ has quit IRC04:06
*** ronny_ has joined #circuits04:07
*** Osso has quit IRC04:15
*** Osso has joined #circuits04:20
Lamb3[Shadowlamb] - The newbie runner BusyR{14} got killed by the newbie NPC SecurityTroll[934313]05:45
prologicwho owns this bot?05:45
*** ronny_2 has joined #circuits05:50
*** ronny_ has quit IRC05:53
*** ronny_2 has quit IRC07:54
*** Osso has quit IRC09:29
*** ronny has joined #circuits10:31
*** ronny has quit IRC10:31
*** ronny has joined #circuits10:31
*** ronny has quit IRC10:38
*** ronny has joined #circuits10:38
*** ronny has quit IRC10:38
*** ronny has joined #circuits10:38
*** ronny_ has joined #circuits10:43
*** ronny_ has quit IRC10:43
*** ronny_ has joined #circuits10:43
*** ronny has quit IRC10:46
spaceoneprologic: i am sorry10:47
Lamb3spaceone: Ich verlasse nun und komme auch nicht von selbst wieder.10:47
spaceoneprologic: if we drop python2.6 support we can implement pythons  OrderedDict as base class for Headers10:48
ronny_wrong base class :)10:49
ronny_you want a ordered multidict, not just a dict :)10:49
spaceoneronny_: multidict?10:53
ronny_spaceone, well, http headers may happen more than once, either you keep correct order or you keep none of it10:54
spaceoneronny_: yes, but each header is seperated by an ','10:54
spaceoneso we don't need a multidict10:54
spaceoneronny_: HTTP allows:10:54
spaceoneX-Foo: 1\r\nX-Foo: 2\r\n10:54
spaceonebut this is the same as:10:55
spaceoneX-Foo: 1, 210:55
spaceoneand as this is simpler (for every client) we prefer this10:55
spaceoneronny_: you understood?12:03
ronny_spaceone, im recalling cases where it doesnt work that way (at least in http)12:07
ronny_but thats fuxxy12:07
ronny_eh fuzzy12:07
ronny_i'll have to check my knowledge, since i could be misstaken there12:08
spaceoneronny_: this would be client implementation errors12:09
spaceoneit is 100% valid HTTP12:09
spaceoneand i think every client should be able to understand this12:10
spaceone(which headers are arrays?) e.g. Authenticate is one of the headers12:10
spaceoneand of course all Accept-* header12:10
*** Osso has joined #circuits12:32
jgiorgiprologic: hows a fix for this generator issue coming?12:32
prologicjgiorgi, still coming - sorry dude13:14
prologichad to build a new filtering api to help solve it - which is done13:15
prologicheader order is not important and not required (recommended only)13:15
jgiorgiworks for me13:15
spaceoneprologic: yes13:15
spaceoneprologic: it is a 'nice to have'13:15
prologicyeah but not neccessary13:16
prologicwe can just subclass OrederedDict from collections13:16
spaceonei implemented it on my own that Date and Server is send first13:16
prologicso I think we'll drop 2.6 support13:16
spaceonedo we really want this?13:16
prologicwhy not13:16
spaceone(debian squeeze uses p2.6)13:17
prologic2.6 is obsolte13:17
spaceonebut squeeze is not obsolete13:17
prologicdebian squeeze also has 2.7 in ports13:17
spaceoneit is the current stable debian release13:17
spaceoneonly debian wheezy13:17
prologicyou're joking?13:17
prologicok well I have to get ready to go to work13:17
jgiorgidebian is always a version behind13:17
prologicmaybe we'll lave 2.6 support alone for a while longer13:18
jgiorgiredhats are always 10 version behinds13:18
spaceonejgiorgi: yes13:18
prologicdamn Debian13:18
jgiorgibetter than redsplat13:18
prologicdamn Redhat13:18
spaceonedebian is 'stable'13:18
prologicthey can all go to hell :)13:18
jgiorgithat's the nice way to put it13:18
prologicstable != old versions13:18
spaceonei am using wheezy now on my server13:18
prologicyou can't imply that either :)13:19
spaceonehmm, not at all13:19
prologicit's just their policy13:19
prologicit doesn't guarentee stabness and good security13:19
spaceonethere are much bugs on distributions which has the latest things13:19
prologicyou're actually better off with the latest versions of things ihmo13:19
spaceonenot always13:19
prologicnot bleeding edge13:19
spaceonei also think that debian is a little bit too slow13:20
prologicI agree13:20
prologicanyway gotta get ready13:20
spaceone( i am working on a debian based distribution at work )13:20
prologicyeah ok we will leave Python 2.6 alone for now13:43
jgiorgiprologic: "make circuitscontrib part of the circuits namespace" how, why?13:51
prologicwhy not?13:52
prologicit makes importing nice13:52
prologicfrom circuits.contrib import blah13:52
prologicits still separate though13:53
jgiorgiok fair enough13:53
prologicif its not installed it just wont be there13:53
jgiorgijust warn me next time, that breaks 18 apps13:53
prologicoh nevause i changed the import paths?13:54
prologicoh ooops ;)13:54
jgiorgii'll fix em :)13:55
prologicsimple sed should fix?13:55
prologici just wanted to add my dmsll 2c worth snd make it nice ;)13:55
prologici want to use this stuff too13:55
jgiorgiit's nice, tbh i honestly didnt know how to do that13:55
jgiorgii just attempted to patch something and everything broke lol, i was confused13:56
prologicim happy eith the layout and structure now,13:56
jgiorgiyou're welcome to use any of it, i tried to document it fairly well, i have half the wiki done, i'll push it when i finish it13:59
prologicwhy not sphinx focs?14:04
jgiorgifigured i'd just use the builtin bitbucket wiki14:07
prologici think mist projects use sphinx and publish to these days14:09
jgiorgii'll have to learn the markup then14:09
prologicits pretty easy14:12
prologicpip install sphonx14:12
jgiorgiok i'll look at it14:13
jgiorgii've got an import error i cant figure out14:13
prologictheres also sphinx-apidoc to auto generate API docs14:14
jgiorgioh im trying to fix the init drop in script14:36
prologicis it suppos to be importable?14:37
jgiorgithe dropin imports circuits.contrib.core.main14:37
jgiorgiand that import is where it's broken14:37
jgiorgibut it seems to be a problem with the python path, i think i fixed the actual imports14:38
prologicyou font use virtualenv /   irtualenwrapper ?14:39
prologiclearn it ;)14:40
prologicvery nice tools ;)14:40
jgiorgii think some older project use it, back when the circuits api was slightly less stable14:40
jgiorgiis there an easy way for mercurial to dynamically write the revision number into a file?14:40
jgiorgii think you used to do that somehow14:41
prologicyeah used to use hgtools14:41
prologicbut dropped it14:42
jgiorgiah ok14:42
jgiorgiwell i have no g-d idea why but i can import circuits.contrib as root but not as anybody else14:48
jgiorgithink development install is broken14:52
jgiorgistandard install fixed it14:52
prologicill confitm when i get to work14:53
prologicit was working for me just fine14:53
ronny_prologic, $comment_you_know14:55
ronny_told you so carma14:55
prologicno dont14:55
ronny_prologic, i still strongly suggest getting rid of that namespaces thing, it will bite some more else14:58
jgiorgii like it but it's causing me hell14:58
prologicif its do horribly broken14:58
prologicwhy habe it?14:59
prologiczope and others use namespaces14:59
ronny_prologic, they also use buildout, and not develop installs, that evades the problem14:59
prologicjgiorgi:  let me retest when i get into work...14:59
jgiorgidamn it now i have to change all my imports back14:59
ronny_prologic, as faras i can tell you did not yet make a release with the namespace thing?15:00
jgiorgii have no idea why15:01
ronny_prologic, then you dont have to support it that way15:01
jgiorgieverything seems to work but my site is broken15:01
ronny_prologic, al i know is that it breaks horrible in some cases15:01
ronny_and if you just use it, you will most likely to inflict that break on users15:02
ronny_(this is precisely the reason why flask dropped the flaskext namespace and advices extensions to go for flask_something)15:03
prologicsuggest an alternative thats still lets one do: from circuits.contrib import blah15:03
ronny_there is none :)15:04
ronny_except for15:04
prologicthen well figure this out15:05
ronny_a module thats not15:05
prologicunless i cannot make it work15:05
ronny_in flask we have the flask.ext object, which is a module subclass15:05
ronny_it dispatches to flask_foo15:05
jgiorgifixed it, want the namespace issue15:07
jgiorgiit wasnt the namespace issue, i was stupid and lost a couple files15:14
prologicahh ;)15:16
prologicok bbs15:17
*** Osso has quit IRC15:28
*** Osso has joined #circuits15:38
prologicyeah ok15:54
prologicso the way we're using namespaces here15:54
prologictotally works fine15:54
prologicin installed and develop15:55
jgiorgiok cool15:55
jgiorgii must have something wrong with my path somewhere that made develop fail15:56
jgiorgii have all 18 apps reconfigured and working15:56
prologicwell I was able to see what ronny was talking about15:56
prologicit's easily fixed with a clean environment15:56
prologicif you python install15:56
prologicand later python develop15:56
prologicthings get confused15:56 is now running based off circuits.contrib15:57
prologicoh nice :)15:57
prologicI'll check it out15:57
prologicyeah this is confirmed15:57
prologiccircuits.contrib defined the circuits namespace package15:57
prologic(not circuits)15:57
prologicif you install both in develop, it's fine15:57
prologicif you install both it's fine15:57
jgiorgimy environment is far from clean, it's been abused15:59
prologicthis is confirmed16:00
prologicit breaks if you start mixing install/develop16:00
prologicshould you shouldn't be doing anyway16:00
prologici.e: cd circuits && python install && python develop16:00
prologiccd circuits.contrib && python install && python develop16:00
prologicif you do that16:00
prologicthings breaks16:00
jgiorgii have no idea how many times things have been installed and/or developed anymore16:02
prologicI highly recommend virtualenvwrapper :)16:03
prologicit's awesome16:03
prologicand helps find import bugs more quickly16:03
prologicthat and fox + pytest obviously16:03
prologicif you're not already using them16:03
jgiorgihrmm ok16:16
jgiorgii'll add it to my todo list16:16
prologicyou deploy your apps using wsgi?16:16
*** Osso has quit IRC16:17
jgiorgiprologic: no, i deploy via proxy16:21
jgiorgiwsgi is also on the list but it seemed like too much trouble and i have a lot of work to do16:21
*** ronny_ has quit IRC16:21
*** Osso has joined #circuits16:22
prologicI noticed with though16:22
prologicX-Powered-By: circuits-2.1.016:22
prologicIf you were doing reverse proxy16:22
prologicit would be Server: circuits-2.1.016:22
prologicI thought16:22
prologicahh no16:23
prologicnevermind :)16:23
prologicI'm wrong16:23
prologicyou just have your web server setup to not identify what it is16:23
prologicX-Powered-By is correct in this case16:23
jgiorgiyeah i have apache setup to be as invisible as possible16:35
spaceoneServerTokens Prod16:36
spaceoneServerSignature Off16:36
spaceoneOptions -ExecCGI16:36
spaceoneyou can also remove the Server header completely i think16:36
prologicfair enough16:36
spaceonewith mod_headers16:36
prologicI don't really care about stuff like that :)16:37
jgiorgifor security purpose it's honestly best to hide most of what and what version is in use16:37
jgiorgiif an exploit is ever discovered that information is just helping the would-be attacker16:37
prologicthat may be so16:37
prologicbut I still don't really care that much :)16:37
prologicit's not like my sites are heavily trafficked :)16:38
jgiorgii'll be deleting the X-Powered-By header soon16:38 is going to attract a bit of traffic16:38
prologicoh well16:38
prologicthis is why web server stats are skewed :/16:39
spaceoneprologic: jeah, i found the security whole in your app without looking at the server header :D16:39
prologicdoesn't matter16:39
jgiorgii like it in principle but it's just too risky, remember what happened with bind a few years ago?16:39
prologicsecurity should not be done via obscurity ihmo16:39
prologicif you want to be scure16:39
prologicuse a jail and chroot16:39
prologicrun separate jailed vas (openvz is good)16:40
prologicthis is my plan for my server upgrades16:40
prologicopenvz + chrooted/jailed apps16:40
spaceoneprologic: btw are you going to fix it soon in the stable branch?16:40
spaceoneand on your wikisite16:40
prologicI alredy have?16:40
spaceonehm, ok16:40
prologicdidn't I pull in your changesets already?16:40
spaceonedid not look16:40
prologicjgiorgi:  think about what I said though16:41
spaceonein circuits-dev, but also in circuits?16:41
prologicobscurity != security16:41
prologicuse chroots with bad jails or linux vz (openvz)16:41
jgiorgiprologic: obscurity is a layer of security imo, anything to make it a bit harder, i do have some jails setup but not as tight as i should16:42
spaceonejeah but many people don't use chroot16:42
prologicspaceone:  we merge into circuits (stable) when we're ready to release16:42
spaceoneprologic: there were fixes which have been released before16:42
spaceone(the windows socket thing)16:42
prologicI will merge it up soon16:42
prologicI normally don't though16:42
prologicI expect users to follow the dev branch16:43
prologicif they want the latest changes16:43
prologicit's what we develop against16:43
prologicand regularly test16:43
prologicso it's just as stable :)16:43
jgiorgididnt used to be lol16:43
prologicno this is true16:43
prologicmany years ago16:43
prologicbefore we seriously did unit and system testing16:44
prologicand before we have CI (We now use ShiningPanda CI)16:44
jgiorgii truly appreciate the stability improvements, changing code all the time blows16:44
prologicbe aware we're deprecating filter=True btw16:44
jgiorgii heard16:44
prologicwe'll probably drop it in 2.216:44
prologicwhen we build a bunch of other new features16:45
jgiorgii'm ok with that, it will break some code but with a definite purpose16:45
prologicit's a small change16:45
prologic   ...16:45
spaceoneprologic: btw is there a changelog?16:45
prologicchanges to:16:45
prologicdef on_foo(self, event, ...):16:45
prologic   event.stop()16:45
prologicspaceone:  of course16:45
prologicthe reason we only merge with the stable branch btw16:49
prologicuntil end of sprint and release16:49
prologicis that's where the issues and stories get closed16:49
spaceoneit is the correct way16:49
spaceonebut sometimes errata updates are recommendet16:49
prologicjgiorgi:  I setup commit messages to go to #circuits-dev (here) and for circuits.contrib16:56
prologicspaceone:  whilst I agree16:57
prologiccircuits-dev always becomes circuits (stable)16:57
prologicso there is no reason whatsoever why someone could not use dev16:57
prologicwe try very hard not to break anything :)16:57
jgiorgiprologic: that's fine by me16:57
jgiorgii'll keep it clean16:58
jgiorgiwell im off, back in a few hours17:01
prologicwhere is our bot?17:02
prologicoh it's on #circuits-dev17:03
spaceoneshall i get Lamb back?17:06
prologicwhat for?17:07
prologicwhat's it do?17:07
prologicthe circuits bot runs on my home del conection17:07
prologicwhich I've noticed is extremely unstable lately17:07
spaceonejust listening to command you give him17:07
prologicneed to contact my provider and get them to fix the line17:08
spaceonee.g. translate things17:08
spaceoneor wiki, or google something17:08
prologicyou speak english quite well :)17:08
spaceoneor WTF something17:08
spaceoneor slap someone17:08
prologicI have a bot that does all that17:08
prologicbut I don't really run it much17:08
spaceoneor make a badjoke17:08
spaceoneor display a random quote17:09
spaceoneleave messages for someone17:10
spaceonewho is offline17:10
spaceonetelling the date17:11
spaceoneand time17:11
spaceoneannounce svn commits17:11
*** Osso has quit IRC17:26
spaceonegn8 ;)17:46
*** ronny_ has joined #circuits22:09
jgiorgiprologic: just letting you know i created an unstable fork of circuits.contrib so i can break things at will22:16
prologicjgiorgi, yes I know :)23:10
prologicI get all the emails :)23:10

Generated by 2.11.0 by Marius Gedminas - find it at!