IRC Logs

2008 9
Mo Tu We Th Fr Sa So
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30          

03. 09 2008

[00:42:04] * Kaelten has quit IRC
[00:45:37] * lakin has joined #pocoo
[00:50:56] * lakin has quit IRC
[00:54:09] * lakin has joined #pocoo
[01:16:18] * highwaychile has quit IRC
[01:51:40] * lakin has quit IRC
[02:52:42] * izibi has quit IRC
[03:07:35] * michelp has joined #pocoo
[03:33:15] * davidcramer has joined #pocoo
[04:37:23] * aconbere_ has joined #pocoo
[04:40:34] * michelp has quit IRC
[04:59:25] * EnTeQuAk has joined #pocoo
[05:12:59] * EnTeQuAk has quit IRC
[05:23:04] * grumpy has joined #pocoo
[05:49:42] * grumpy has quit IRC
[06:24:40] * rcscomp has joined #pocoo
[06:25:26] <rcscomp> Is there a good way to validate/convert GET arguments? The routing system works great for "slash" arguments, but I have arguments from the GET string I need to work with too.
[06:27:43] <hads> The routing system is only for the path really. For the query string you will need to use request.args.
[06:30:02] <rcscomp> hads: thanks, I figured that. Any "best practice" recommendations for handing request.args?
[06:30:45] <rcscomp> I suppose one could build a light declarative wrapper around FormEncode validation.
[06:31:06] <hads> Not quite sure what you are after there sorry.
[06:34:22] <rcscomp> Just that it would be nice if you could code something like "page = IntValidator('page')" and then have IntValidator know it needs to go to request.args, get the 'page' value, check to make sure its an int, and assign the value converted to an integer to the local variable page. Exceptions could be handled in a consistent manner in a controller super class and this would save someone from having to write a lot of "if isint(page) then retu
[06:34:56] <rcscomp> obviously a few different validators could be written for that purpose
[07:09:09] * robart has joined #pocoo
[07:09:16] * robart has left #pocoo
[07:16:39] * boink__ has joined #pocoo
[07:28:51] * raz has quit IRC
[07:28:51] * boink__ is now known as raz
[07:56:42] * damjan has quit IRC
[07:56:56] * damjan has joined #pocoo
[08:14:53] * aconbere_ has quit IRC
[08:16:56] <davidcramer> TypeError: coercing to Unicode: need string or buffer, SilentUndefinedType found
[08:17:00] <davidcramer> anyone know what causes this typically?
[08:17:10] <davidcramer> in a typical django environment :P
[08:17:24] <davidcramer> http://www.pastethat.com/sa0zE
[08:17:27] <davidcramer> its happening with this chunk
[08:30:57] <thatch> davidcramer: my guess would be that GAME is not in scope
[08:31:43] <davidcramer> hrm k
[09:18:51] * boink__ has joined #pocoo
[09:31:07] * raz has quit IRC
[09:31:08] * boink__ is now known as raz
[10:01:51] <asmodai> birkenfeld: mmm, how far along are you in BH?
[10:04:02] <birkenfeld> asmodai: prologue ;)
[10:04:17] <asmodai> ahhh
[10:04:20] * asmodai hushes then
[10:18:05] <asmodai> must say
[10:18:10] <asmodai> this Google Chrome works pretty ok
[10:19:14] <xorAxAx> so whats your chrome GUID?
[10:19:27] <xorAxAx> every chrome user has a GUID, a persistent cookie
[10:21:12] <xorAxAx> of course this guid is only transmitted to google and nobody else
[10:23:38] <asmodai> oh, haven't checked
[10:42:19] <dennda> on updates only afaik
[10:45:28] * EnTeQuAk has joined #pocoo
[10:54:03] <asmodai> what I find the most intriguing is:
[10:54:17] <asmodai> Google has a positive image in general, whereas Microsoft doesn't, right?
[10:55:20] <asmodai> rofl: http://i454.photobucket.com/albums/qq262/freehostingbucket/gordonfreeman.jpg
[11:11:49] <asmodai> so given that image, people are likely to pick up on Google Chrome easily
[11:11:57] <asmodai> I wonder if that means it'll cut into IE's market share
[11:14:58] <aat> how embarrassment: http://www.readwriteweb.com/archives/security_flaw_in_google_chrome.php
[11:15:11] * davidcramer has quit IRC
[11:16:09] <asmodai> aat: ouch :(
[11:20:09] <asmodai> aat: admit it, your fault
[11:20:18] <aat> damn you
[11:22:01] <asmodai> aat: why? :)
[11:35:49] <asmodai> aat: http://tapthehive.com/discuss/This_Post_Not_Made_In_Chrome_Google_s_EULA_Sucks
[11:37:35] <aat> yeah, that eula is shite
[11:37:50] <asmodai> I'm sure it's just an overzealous lawyer that put that in
[11:37:57] <asmodai> and will soon be rectified after the web explodes over it :D
[11:41:32] <asmodai> aat: but all in all I found it a refreshing experience
[11:41:48] <asmodai> aat: just funnyto see the vista gui on XP
[11:43:27] * izibi has joined #pocoo
[12:27:51] * grumpy has joined #pocoo
[12:28:02] * grumpy has left #pocoo
[12:41:53] <apollo13> who is going to join us for a newforms fork?
[12:42:47] * highwaychile has joined #pocoo
[12:46:27] <asmodai> a newforms fork?
[12:46:45] <apollo13> jep
[12:46:55] <ronny> hmm what about wtforms?
[12:47:06] <apollo13> we don't like it that much
[12:47:12] <apollo13> (doesn't mean it's bad though)
[12:47:29] <ronny> why dont you like it ?
[12:47:39] <ronny> (it seems fine for the most cases)
[12:47:45] <ronny> so what does it lack
[12:48:13] <apollo13> no docs, no formset_factories
[12:48:31] <apollo13> I don't now if there is a way to return different data in something like a clean method
[12:48:45] <apollo13> eg user enters ronny clean returns ynnor
[12:49:35] <apollo13> and converting to another formlib might take us long ;)
[12:49:51] <ronny> formset factory?
[12:50:13] <ronny> custom data processing is possible
[12:50:15] <apollo13> essentialy the same form displayed more than once, like django inlines
[12:50:36] <ronny> (i make it turn texts into db objects using query.get_by)
[12:51:15] <apollo13> and we would like to see babel support
[12:51:19] <ronny> apollo13: you mean like show a FooForm for all items in the list thats at Barobject.foo ?
[12:51:37] <apollo13> ronny: yeah
[12:51:40] <apollo13> basically
[12:52:30] <ronny> yeah, currently that doesnt work, but something like listhandler should be simple
[12:53:00] <apollo13> and I somehow can't get my head around the validator approach
[12:53:16] <prencher> that's a first
[12:53:19] <prencher> apollo13 - custom data conversion is a matter of defining custom fields (this will be made easier)
[12:53:54] * hads likes WTForms :)
[12:53:59] <apollo13> well I think we could live mostly without data conversion, cause we merly use it
[12:54:04] <apollo13> but it's nice to have
[12:54:34] <ronny> its rather simple
[12:54:40] <apollo13> and I dislike the _validate_x sytax
[12:54:45] <apollo13> validate_x would be cleaner
[12:54:49] <prencher> and you still blabber on about i18n support, when everything you need is already exposed, there is no need for the framework to know anything about it
[12:55:15] <prencher> it's unicode clean, and anything that outputs anything has a message= attribute to override, i fail to see how i18n is even a factor
[12:55:25] <apollo13> phone brb
[12:55:38] <prencher> it's made specifically that way exactly because we can't possibly give you good error messages anyway, only generic mostly-debug ones
[12:57:50] <apollo13> re
[12:59:43] <ronny> apollo13: so far the only valid issue you told is formsets, and those arent that hard to add to wtforms
[13:00:50] <apollo13> hmmm
[13:02:59] <apollo13> we'll se, we tend to use newforms, cause we are used to it I think
[13:03:28] <apollo13> and it might take us quite long to move from newforms to wtforms
[13:04:27] <hads> That's a more understandable reason.
[13:05:03] <apollo13> only to some extend, forking newforms is work too ;)
[13:06:12] <hads> I didn't say I agreed with it, just that it's more understandable :)
[13:06:33] <apollo13> btw what about initial data?
[13:07:07] <hads> With WTForms? Can be populated from an object or arguments.
[13:07:31] <apollo13> how?
[13:08:53] <hads> It's in the crash course form memory.
[13:09:10] <prencher> apollo13 - i'm not being dismissive of your concerns just as an aside, but if you would post on the group maybe we can get some discussion about it, so i can get a clear picture of what people really want from the next version (other than documentation)
[13:10:06] <prencher> there's also a rearchitecture of fields, converters and validators coming so that'll be cleaner.. there will also be a very simple widget system so you can render e.g. html4 without custom templates, but not in the django newforms widget sense
[13:10:37] <prencher> but beyond that, i do want some sort of builtin mechanism for repeated fields and forms (and subforms)
[13:10:43] <hads> FWIW without any input from me on what I'd like you're doing exactly that from the commits I've read :D
[13:11:47] <prencher> apollo13 - with the rearch, fields will be able to easily define default validators, so django newforms style fields will be added
[13:12:07] <prencher> e.g. an EmailField or similar that has a validators.not_empty and validators.email set
[13:12:10] <apollo13> how long will this take?
[13:12:37] <prencher> approximately X
[13:12:55] <prencher> haven't had a lot of time to work on wtforms specifically lately, but hopefully soon
[13:13:40] <apollo13> hmm you would have time until we kick django
[13:13:45] <apollo13> then we decide I think
[13:13:54] <apollo13> and that decission might be final
[13:14:29] <prencher> anyway, to avoid spamming the channel further and hopefully get some discussion going, post to http://groups.google.com/group/wtforms if you would with your concerns (and/or other stuff)
[13:15:08] <apollo13> what's wrong with discussiong it here?
[13:15:22] <asmodai> it's been quiet anyway :D
[13:15:23] <prencher> well it's not a pocoo project and i don't want to impose my projects too much here
[13:16:13] <prencher> apollo13 - but again, not to be rude or dismissive, but i'm really not looking to win people over; I've used wtforms for a long long time (years), it was opensourced mainly for some people in here liking what i described of it and them being fed up with newforms
[13:16:30] <hads> Thanks :)
[13:16:34] <prencher> if you use it, great.. if not then that's great too, just not for you
[13:16:50] <apollo13> well I would like to use wtforms as it would mean one lib less for us to maintain
[13:16:52] <prencher> was mainly hads, ronny and aa_ (or was it aat?) i believe
[13:17:15] <prencher> well mainly ronny, loudmouthed tart ;)
[13:17:30] <asmodai> *laughs8
[13:17:37] <ronny> GAH ME IZ NOT LOUD
[13:17:41] <hads> I think I came in after the fact, but I'm certainly greatful.
[13:18:03] <ronny> LOLCAT DID!
[13:32:52] * EnTeQuAk has quit IRC
[13:37:13] <ronny> bbl
[13:41:14] <apollo13> prencher: I still think you could mark all your strings translateable
[13:41:38] <apollo13> looking at DateTimeField for instance u'Date is invalid.' is pretty much all you need to say
[13:42:08] <apollo13> or am I missing something?
[13:43:29] <apollo13> some for the validator messages, why should I always come up with my one? what's more to say than: 'Invalid email address.'
[13:44:01] * EnTeQuAk has joined #pocoo
[13:48:43] * izibi has quit IRC
[13:49:33] <prencher> apollo13 - i can see it from both sides, but can i add this without adding a dependency on babel or whatever?
[13:49:44] <apollo13> prencher: of course
[13:50:05] <apollo13> at least if I understand translation systems right
[13:50:27] <apollo13> just check if _ is in globals if not, put _ = lambda x: x in there
[13:50:42] <prencher> i've not needed to i18n anything generically like that before, so im unsure of the process.. but i refuse to add any non-stdlib depedencies
[13:51:01] <prencher> well, if that's all it takes, ill be happy to add it to all standard messages
[13:51:19] <apollo13> ask someone with more knowledge about this, but I think it should be enough
[13:51:30] <apollo13> I'll ask cmlenz
[14:01:10] * EnTeQuAk has quit IRC
[14:07:01] <apollo13> prencher: you might wanna look how formencode is doing this
[14:07:57] <apollo13> it relies on __builtin__._ and if it's not existent setting up it's own _ using gettext
[14:34:41] * michelp has joined #pocoo
[14:37:35] <prencher> hey asmodai
[14:37:49] <prencher> seen imdb lately? they've changed the layout to the kind you love: http://www.imdb.com/title/tt0468569/
[14:42:03] <asmodai> oh yes
[14:42:05] <asmodai> narrow it down
[14:42:09] <asmodai> imagine using your viewport
[14:42:41] <prencher> dude
[14:42:44] <prencher> your screne needs more gray
[14:43:26] <prencher> asmodai - the worst part is..
[14:43:34] <prencher> that design is so obviously designed for fullscreen
[14:43:39] <prencher> this just looks so force
[14:43:40] <prencher> d
[14:44:14] <asmodai> makes you wonder: WHY?!
[14:44:43] <prencher> upcoming redesign?
[14:44:54] <prencher> hahahahaha...imdb redesigning is about as likely as google.com changing
[14:46:35] * michelp_ has joined #pocoo
[14:50:08] * Kaelten has joined #pocoo
[14:59:40] * michelp has quit IRC
[15:01:05] * highwaychile has quit IRC
[15:06:01] * michelp_ is now known as michelp
[15:29:47] * lakin has joined #pocoo
[16:04:55] * michelp_ has joined #pocoo
[16:09:37] * michelp has quit IRC
[16:18:07] * Kaelten has quit IRC
[16:18:36] * prencher has quit IRC
[16:20:04] * prencher has joined #pocoo
[16:25:17] * Kaelten has joined #pocoo
[16:45:16] <rcscomp> prencher: you still around and up for some more wtforms off-topic discussion?
[17:04:47] * M3ntor5 has joined #pocoo
[17:24:19] * michelp_ has quit IRC
[17:39:34] * [diecast] has joined #pocoo
[17:54:27] * highwaychile has joined #pocoo
[18:10:35] * EnTeQuAk has joined #pocoo
[18:35:24] <asmodai> Kaelten: are you involved with wowaceupdater?
[18:35:37] <Kaelten> yes
[18:35:45] <Kaelten> I cofounded WowAce.com :)
[18:35:54] <asmodai> Kaelten: What could cause:
[18:36:05] <asmodai> rightclick addon: open forum thread
[18:36:13] <asmodai> Shell Error
[18:36:17] <asmodai> Application not found?
[18:37:13] <Kaelten> not sure, but on a guess its either trying to use something other than http, or it can't launch your browser properly
[18:37:31] <asmodai> worked before
[18:37:34] <asmodai> I wonder what got changed
[18:44:43] * Tik-Tok has joined #pocoo
[18:46:00] <prencher> god asmodai
[18:46:02] <prencher> you broke it
[18:46:09] <asmodai> did not!
[18:46:16] <prencher> besides Kaelten is being a jerk and shutting down the ace updater soon anyway
[18:46:37] <prencher> just cause it uses some 40tb of bw, pfft
[18:47:09] <asmodai> :|
[18:47:12] <asmodai> seriously?
[18:47:33] <Kaelten> 40-60
[18:47:52] <asmodai> holy moly
[18:48:08] <prencher> he had some fun stories to tell about patch days where they max a 1gig pipe
[18:48:27] <Kaelten> hehe, ya seeing 800+ mbit/s of addons get downloaded is.... fun
[18:48:29] <prencher> so much at one point they had to use two, was it Kaelten?
[18:48:33] <asmodai> sweet bejesus
[18:48:47] <asmodai> that's insane
[18:49:10] <prencher> all off of one box running nginx though, quite impressive
[18:49:18] <Kaelten> ya, nginx is awesome
[18:49:27] <Kaelten> maxed the gige pipe, I was very impressed
[18:49:43] <asmodai> i'm getting a bit sick of the lighttpd developer mentality
[18:49:44] <Kaelten> server runs at a constant 7 + load
[18:49:47] <asmodai> been pondering nginx
[18:49:53] <Kaelten> asmodai: I like it.
[18:49:58] <Kaelten> er, correction, love it
[18:50:09] <Kaelten> its the best thing out for static file serving
[18:50:13] <Kaelten> and not bad for cgi aps
[18:50:14] <asmodai> well, the last time I looked at it most documentation was in Russian, heh, things have much improved by now
[18:50:19] <prencher> asmodai - it's faster, it scales better (it has N workers that run the event model, rather than one like lightty)
[18:50:29] <Kaelten> someone was even working on a mod_wsgi port to it
[18:50:29] <prencher> also.. the nginx config format is not retarded
[18:50:38] <prencher> Kaelten - that port is a piece of shit
[18:50:43] <prencher> run apache+mod_wsgi behind it
[18:50:43] <Kaelten> really?
[18:50:46] <Kaelten> heh
[18:50:56] <asmodai> prencher: I'll look at it soon then
[18:51:01] <Kaelten> I'd like to be able to ditch apache one day
[18:51:07] <Kaelten> I was hoping it'd turn out good
[18:51:41] <prencher> Kaelten - it has no daemon mode i believe, and no isolation
[18:51:56] <prencher> also nginx can't scale the workers, and as a result, neither can it's mod_wsgi (based on grumpys work)
[18:52:08] <Kaelten> hrm
[18:52:14] <Kaelten> thats sad
[18:52:15] <prencher> also, iirc, you lock up an entire process when serving a wsgi request
[18:52:20] <Kaelten> ewe
[18:52:23] <prencher> which means you have to basically proxy nginx with nginx
[18:52:32] <prencher> or just do the sane thing and run nginx+apache/mod_wsgi
[18:52:59] <prencher> which affords you grumpys awesome work, as well as apaches ability to scale processes
[18:53:15] <Kaelten> heh
[18:53:16] <prencher> or daemon mode - on our server we had nginx in front serving static and proxying to a single apache
[18:53:27] <prencher> that apache in turn talked to some 5-6 different daemon based wsgi apps
[18:53:32] <prencher> which all ran under seperate users
[18:53:44] <Kaelten> i normally just have nginx serving static files on a separate ip
[18:54:14] <prencher> and of course, mod_wsgi can recycle procs after a timeout, so the procs not often accessed, like trac, will recycle down to the minimum shared memory footprint
[18:55:00] <prencher> Kaelten - sure, that works too - apache is quite good as a frontend server if you can give it some 400-500mb on the box
[18:55:16] <prencher> personally, i prefer to just save the memory, and put nginx in front, use that extra couple hundred megs for extra wsgi instances
[18:55:39] <Kaelten> heh
[18:55:43] <Kaelten> sadly I'm still running prefork
[18:55:44] <prencher> varies from setup to setup which works better, but the nginx<->apache proxy overhead is undetectable on the same box
[18:55:47] <Kaelten> I don't trust django
[18:55:56] <prencher> nothing wrong with using prefork
[18:56:16] <prencher> python doesn't scale very well with threading anyway, so you tend to not really get any extra perf out of threads
[18:56:33] <Kaelten> GIL ?
[18:56:36] <prencher> indeed
[18:56:41] <Kaelten> shame
[18:56:56] <prencher> you can get better concurrency, but usually it comes at extra latency
[18:57:14] <Kaelten> *nods*
[18:57:20] <Kaelten> hrm, lunch time
[18:57:20] <Kaelten> bbiab
[18:57:25] <prencher> however - it does mean you can serve more requests, though with slightly higher latency on each, at a much lower footprint than prefork
[18:58:07] <prencher> this is purely based on my own experience testing and deploying though, YMMV
[19:05:52] * dennda just "upgraded" to zine (lazily)
[19:06:10] <prencher> did the world explode?
[19:06:21] <dennda> almost
[19:06:32] <dennda> but being a hero I saved it
[19:06:39] <prencher> that's good
[19:53:51] * _paneb has joined #pocoo
[19:54:17] <_paneb> does anyone have an example of using the Routing system for a controller setup similar to the one in pylons?
[20:02:11] * izibi has joined #pocoo
[20:31:42] * jinks has quit IRC
[20:32:17] * jinks has joined #pocoo
[20:42:29] * boink__ has joined #pocoo
[20:47:54] * EnTeQuAk has quit IRC
[20:54:32] * raz has quit IRC
[20:54:32] * boink__ is now known as raz
[21:57:37] <birkenfeld> hmm
[21:57:41] <birkenfeld> zine?
[22:12:08] <prencher> mits new name for textpress
[22:12:20] <prencher> gawd birkenfeld, keep up ;)
[22:13:18] <birkenfeld> oh
[22:13:19] <birkenfeld> hmm
[22:13:24] <birkenfeld> not easily googlable
[22:13:32] <prencher> well, yet anyway
[22:13:38] <prencher> he only just changed the name a couple days ago
[22:13:57] <birkenfeld> uh, am I blind or is there really no way to nonrecursively add directories in hg?
[22:14:36] <prencher> hm, good question
[22:14:42] <birkenfeld> ah
[22:14:47] <prencher> you could try hg add --exclude foo/*/ foo
[22:14:49] <birkenfeld> you can just go into the dir and add what you need
[22:15:09] <prencher> or that, but an exclude pattern ought to work
[22:15:29] <prencher> (you can use -n to do a dry run and see if it does what you need)
[22:21:42] * Kaelten has quit IRC
[22:22:23] * Kaelten has joined #pocoo
[22:25:14] * Kaelten has quit IRC
[22:25:22] * Kaelten has joined #pocoo
[22:45:58] * Manni is now known as Wheerd
[22:46:37] * M3ntor5 has left #pocoo
[22:49:26] * Wheerd is now known as Wheerd_
[22:49:38] * Wheerd_ is now known as Manni
[22:51:21] * Manni is now known as Wheerd
[23:04:04] * lakin has quit IRC
[23:29:48] * kaelten_ has joined #pocoo
[23:30:20] * izibi has quit IRC
[23:33:11] * luks has quit IRC
[23:33:15] * luks has joined #pocoo
[23:42:05] * davidcramer has joined #pocoo
[23:47:11] * Kaelten has quit IRC