Pyogp/Chat Logs/Daily Meeting/16 jul 2008

From Second Life Wiki

Jump to: navigation, search

[9:32] Saijanai Kuhn: Ah ha!
[9:32] Infinity Linden: hola folks
[9:33] Opensource Obscure: : )
[9:33] Locklainn Linden: hi
[9:33] Infinity Linden: an accident on hwy 17 almost made me seriously late
[9:33] Enus Linden: glad you made it
[9:33] Saijanai Kuhn: ouch, but you're ok?
[9:33] JB Kraft: howdy folks :)
[9:33] Infinity Linden: fortunately it wasn't too bad
[9:34] Infinity Linden: so.. i'm pressed for time
[9:34] Infinity Linden: and
[9:34] Infinity Linden: a bit embarassed Ididn't do as much as i wanted to do yesterday
[9:34] Infinity Linden: so
[9:34] Infinity Linden: i think i'm going to skip myself
[9:34] Enus Linden: i go i go!
[9:34] Enus Linden: as i don't have much
[9:34] Infinity Linden: and... anyone on #pyogp? is Tao coming?
[9:35] Saijanai Kuhn: he online
[9:35] Locklainn Linden: on his way
[9:35] Infinity Linden: Enus.. we should probably wait for Tao
[9:35] Locklainn Linden: he should go first
[9:35] Enus Linden: i will
[9:35] Locklainn Linden: when he comes on
[9:35] Saijanai Kuhn: and I certainly hven't done anything
[9:35] Locklainn Linden: says hes gotta be quick
[9:35] Infinity Linden: he's a heavy lifer with respect to this projcet
[9:35] Infinity Linden: yay! Tao's here!
[9:35] Infinity Linden: i just mentioned i did next to nothing yesterday
[9:36] Tao Takashi: I posted all I did to the list
[9:36] Infinity Linden: and Enus called dibs on going next
[9:36] Enus Linden: we;re communicating rather well on that list...
[9:36] Enus Linden: so me: i created an internal jira, DEV-17910
[9:36] Infinity Linden: krikey... that's what i get for going on a cereal jag with girlfriends last night instead of reading email
[9:36] Tao Takashi: I try to as it's better than reading chatlogs
[9:36] Enus Linden: it is literally composed of one wiord currently
[9:36] Infinity Linden: yeah... probably a good idear
[9:37] Locklainn Linden: nice
[9:37] Enus Linden: i;'ll fix that today
[9:37] Tao Takashi: if it's the right word...
[9:37] Locklainn Linden: or phrase 'git r done'
[9:37] Enus Linden: i also am speaking with sue and robla about how best to support pyogp on pjira
[9:37] Enus Linden: git r dun is a fav of mine lol
[9:37] Locklainn Linden: haha
[9:37] Enus Linden: wrt pjira, we are debating having pyogp as a metaproject
[9:38] Enus Linden: or as a component to an "open source" meta project
[9:38] Enus Linden: i'll let you know how that turns out
[9:38] Tao Takashi: we finish it in the meanwhile ;-)
[9:38] Locklainn Linden: yep
[9:38] Locklainn Linden: haha
[9:38] Enus Linden: I also have most of a wiki structure proposal sketched out
[9:38] Infinity Linden: i don't know that I care one way or another, so i'll keep my nose out of that debate
[9:38] Enus Linden: and will circulate somehow later today
[9:38] Infinity Linden: yay
[9:38] Tao Takashi: cool
[9:39] Locklainn Linden: ok
[9:39] Infinity Linden: wikis are like totally cool
[9:39] Tao Takashi: actually they suck ;-)
[9:39] Infinity Linden: hmm... saying things in val-speak in SL doesn't really work too well
[9:39] Enus Linden: and friday at 9:30 will be my 'office hours', a forum for a weekly pyogp status and other cool qa-y stuff as needed
[9:39] Infinity Linden: hmm...
[9:39] Tao Takashi: so do I do that ZCA stuff during that OH?
[9:39] Infinity Linden: Tao was goinna do ZCA
[9:39] Enus Linden: sounds appropriate
[9:39] Infinity Linden: okay... cool
[9:40] Tao Takashi: it depends on who is around. It would be good if it's mostly people who really want to work on it
[9:40] Tao Takashi: and have looked at it
[9:40] Infinity Linden: i gotta make sure it's on my Goog calendar
[9:40] Locklainn Linden: yea
[9:40] Tao Takashi: otherwise I expect chaos
[9:40] Enus Linden: i'll setup an agenda page somewhere smometime
[9:40] Locklainn Linden: exactly
[9:40] Enus Linden: done
[9:40] Infinity Linden: well... hopefully we won't pull as many folks as Zero's recent office horus
[9:40] Enus Linden: (early on, it will be just us)
[9:40] Tao Takashi: JB: Did you do Zope3 work, too?
[9:40] Locklainn Linden: [1]
[9:41] Enus Linden: unless sai pollutes the waters with mass spammage
[9:41] JB Kraft: yes, been zoping for quite a while
[9:41] Tao Takashi: right, Sai, please do not announce this too much ;-)
[9:41] Tao Takashi: JB: Cool, so you can help me with explaining interfaces, utilities and adapters then? :)
[9:41] Enus Linden: ty lock
[9:41] Locklainn Linden: np
[9:41] BigMike Bukowski: The entire crowd at zero's end's up making people trying to learn and catch up to be comepletly confused.
[9:41] Infinity Linden: at least until we have something really cool to show off and the beta program is up and running
[9:42] Infinity Linden: that is actually a good point
[9:42] JB Kraft: sure, im just ctaching up at the moment, just got the code out of svn today
[9:42] Tao Takashi: I still need the URL for OpenSim where I can call rez_avatar ;-)
[9:42] Enus Linden: one request: is it possible to keep this status mtg limited to 30 minutes?
[9:42] Tao Takashi: JB: well, it's more about the general principles as these are new to many
[9:42] Tao Takashi: Enus: +1 :)
[9:42] Enus Linden: what i've done/what i pan to do/blockers/issues, done.
[9:42] JB Kraft: gotcha
[9:42] Infinity Linden: i think the reason a lot of folks were bringing virtual pitchforks and torches was the idea that assets were going to freely flow across region domain boundaries
[9:43] Infinity Linden: it would be nice to have code that demonstrates that we _don't_ do that
[9:43] BigMike Bukowski: Yeah. That's what it seemed like.
[9:43] Tao Takashi: ok, so we do the round?
[9:43] Tao Takashi: or are we in the middle?
[9:43] Infinity Linden: okay... yeah... lets' finihs the round
[9:43] Enus Linden: tao, you're up
[9:43] Tao Takashi: ok, so I am a bit blocked by that grok decision
[9:43] Tao Takashi: I would like to get rid of my branch
[9:44] Tao Takashi: besides that I refactored a bit on that branch
[9:44] JB Kraft: what was the grok decision?
[9:44] Enus Linden: lcok, did you review? scope is focused, seems to make sense to merge it
[9:44] Tao Takashi: adding an Avatar class and that example high level API
[9:44] Tao Takashi: JB: yes or no basically
[9:44] Tao Takashi: actually not grok, but grokcore.component
[9:44] Locklainn Linden: I looked at it a bit
[9:44] Locklainn Linden: yea, that is the problem
[9:44] Locklainn Linden: how much of grok is taken in
[9:44] Locklainn Linden: but I guess with our Pyogp guidelines, thats not a problem haha
[9:45] Tao Takashi: it's just those grok.Adapter superclasses and some directives like grok.name() and grok.context()
[9:45] Tao Takashi: it basically replaces implements() and adapts()
[9:45] Tao Takashi: and does all the registration for you
[9:45] Infinity Linden: wait... does grok replace ZCA?
[9:45] Tao Takashi: and grokcore.component is just the registration bit taken out of grok
[9:45] Locklainn Linden: essentially, the unfamilair, (correct me if I'm wrong Tao), is allow you to register which utilities and adapters you suse
[9:45] Tao Takashi: no, grok makes using ZCA easier
[9:45] Infinity Linden: or layer beside / on top of it
[9:45] Infinity Linden: okay
[9:45] Infinity Linden: let's discuss that in email / irc
[9:45] Tao Takashi: normally you register components in XML
[9:46] Infinity Linden: oh oh
[9:46] Infinity Linden: cool
[9:46] JB Kraft: what does grok buy in this context, beyond base transport
[9:46] Tao Takashi: with grok you don't need to
[9:46] Infinity Linden: yeah... gets back to the whole "functionality vs. dependency" argument
[9:46] Locklainn Linden: not using xml JB
[9:46] Tao Takashi: it does enable easier registration of components without XML needed
[9:46] Tao Takashi: so you don't need to switch context and files to register
[9:46] Locklainn Linden: we need one, either xml, grok, or something, otherwise the modules themselves register their components
[9:46] Infinity Linden: okay... i need to schedule a little time playing with it
[9:46] Locklainn Linden: which is order dependent
[9:46] Tao Takashi: I explained this a bit in part 4 of my ZCA tutorial
[9:46] Infinity Linden: but let's move on
[9:47] Infinity Linden: Lock?
[9:47] Tao Takashi: can post the link later
[9:47] Tao Takashi: and I checked in my agent domain stuff
[9:47] Infinity Linden: (I think Open said he was just going to observe)
[9:47] Infinity Linden: w00t!
[9:47] Locklainn Linden: been writing a message template parser for the bodies of packets
[9:47] Tao Takashi: but we can discuss this at some other point as it's far from finished
[9:47] Tao Takashi: so Lock :)
[9:47] Opensource Obscure: ( yes. i'm a learning lurker here )
[9:47] Locklainn Linden: I've been writing a message template parser for the bodies of packets
[9:47] Locklainn Linden: nearly finished
[9:48] Locklainn Linden: Tao broke it today :)
[9:48] Tao Takashi: that's what you get for helping out ;-)
[9:48] Enus Linden: file a jira.
[9:48] Locklainn Linden: for the better
[9:48] Enus Linden: oh, wait...
[9:48] Infinity Linden: lol
[9:48] Tao Takashi: what does it return?
[9:48] Locklainn Linden: haha
[9:48] Locklainn Linden: dict of packets, which you can get their bodies out of
[9:48] Locklainn Linden: and should be able to get the types of each of the data fields in the body
[9:48] Infinity Linden: so... gonna keep on keepin' on with the parser?
[9:48] Tao Takashi: btw, here I would like to see some doctest which explains interactively how to use it :)
[9:49] Locklainn Linden: yes, I have been using unittest
[9:49] Tao Takashi: maybe I write one
[9:49] Locklainn Linden: not for this stuff so far
[9:49] Locklainn Linden: but for the packet decoder/encoder
[9:49] Locklainn Linden: i will write some tests
[9:49] Tao Takashi: cool
[9:49] Tao Takashi: how do you get the types?
[9:49] Infinity Linden: hmm... just dawned on me that the message template is a great place to find which "resource bundles" should be defined first
[9:49] Tao Takashi: Did you create your own classes or are they just names?
[9:49] Locklainn Linden: strings, but now I need to convert to the actual data type
[9:49] Tao Takashi: ok, ic
[9:49] Infinity Linden: yeah... i'm a fan of having real types to work with too
[9:50] Tao Takashi: and are you parsing actual packets or the message templates to python templates?
[9:50] Infinity Linden: not just LLSD blobs floating around
[9:50] Tao Takashi: maybe I just look at your tests :)
[9:50] Locklainn Linden: haha
[9:50] Locklainn Linden: I'll email out what its doing
[9:50] Tao Takashi: cool
[9:50] Infinity Linden: okay.. let's let Sai have his say
[9:50] Locklainn Linden: and mayabe you can give me feedback on what you'd like to see it do
[9:51] Tao Takashi: sure, will have a look at it
[9:51] Infinity Linden: then talk more about it on #pyogp
[9:51] Locklainn Linden: yep
[9:51] Tao Takashi: or the list ;-)
[9:51] Locklainn Linden: haha he chat-o-phobic
[9:51] Infinity Linden: and... of course.. BigMike and JB if you want to join in
[9:51] Tao Takashi: not really, it just takes more time and is harder to keep track of ;-)
[9:51] Tao Takashi: anyway, next one :)
[9:51] Locklainn Linden: haha sai, ur up
[9:51] Infinity Linden: Sai?
[9:52] Infinity Linden: anyone... anyone. .Beuler?
[9:52] Locklainn Linden: afk
[9:52] Infinity Linden: okay
[9:52] Infinity Linden: JB / BigMike
[9:52] Saijanai Kuhn: noithing going on with me. Trying to unblock on RL job, currently
[9:52] Infinity Linden: wanna jump in and volunteer to do anything>
[9:52] Infinity Linden: ?
[9:52] Infinity Linden: okay
[9:52] Infinity Linden: Sai has the RL thign going on
[9:52] Tao Takashi: what's that?
[9:52] Locklainn Linden: reminds me, we should start talking about what components need to be worked on
[9:52] Locklainn Linden: so we can start to distribute work
[9:53] Infinity Linden: oh... and Lor...??
[9:53] JB Kraft: im still absorbing thus far :)
[9:53] Lor Gynoid: I'm just listening in at the moment. :)
[9:53] Tao Takashi: I think once the UDP handler is in place we can distribute handlers to people ;-)
[9:53] Enus Linden: w00t!
[9:53] Tao Takashi: JB at least was checking out the code :)
[9:53] Locklainn Linden: then we should discuss the UDP handler
[9:53] JB Kraft: as we speak :)
[9:53] Lor Gynoid: I may need to learn a new programming language or two. :)
[9:54] Infinity Linden: okay...
[9:54] Infinity Linden: ditto for web services and cap messages
[9:54] Tao Takashi: I have to look at the old example I think
[9:54] Tao Takashi: one thing I wanted to clean up was the serialization component
[9:54] Locklainn Linden: handling cap messages you mean infinitY?
[9:54] Infinity Linden: i think we're going to announce ou plans for prioritizing things in the beta "any day now"
[9:54] Infinity Linden: yup
[9:55] Locklainn Linden: right
[9:55] Infinity Linden: basically just mean "don't forget the HTTP messages"
[9:55] Infinity Linden: whatever we want to call them
[9:55] Locklainn Linden: so we need UDP messages and event queue handling
[9:55] Tao Takashi: as the beta is about TP, the question is: What is missing for TP and how is TP working at all?
[9:55] Locklainn Linden: good question
[9:55] Tao Takashi: so maybe some Linden can find out as you should know most about it ;-)
[9:55] Locklainn Linden: Ill try to uncover that today
[9:55] Infinity Linden: i think teh functionality of TP is based on Zha's patch and the viewer updates
[9:55] Tao Takashi: I can maybe think about how we best do the UDP loop
[9:56] Infinity Linden: that are supposed to go in "any day now"
[9:56] Locklainn Linden: Tao, then post some pseudo to the list
[9:56] Tao Takashi: like how can we do it and still have user input etc.
[9:56] Locklainn Linden: so we can look at it and talk
[9:56] Infinity Linden: hmm... add a ping UDP message?
[9:56] Infinity Linden: I could claim to be fixing a bug or two and sneak it in
[9:56] Tao Takashi: and what I would like to know is: What is the minimum we need to handle to keep the avatar alive?
[9:56] Locklainn Linden: haha
[9:56] Locklainn Linden: Sai already did that
[9:57] Tao Takashi: ok
[9:57] Locklainn Linden: its in Enus' old code
[9:57] Locklainn Linden: I can put that somewhere viewable
[9:57] Locklainn Linden: to work off of
[9:57] Infinity Linden: okay... let's make sure we have that documented
[9:57] Tao Takashi: Well, I was more thinking about a more textual explanation ;-)
[9:57] Locklainn Linden: you want pictures :P
[9:57] Tao Takashi: no ;-)
[9:57] Infinity Linden: caues with all the different code, some of it is bound to get excised
[9:57] Infinity Linden: okay
[9:57] Locklainn Linden: haha jk
[9:57] Infinity Linden: coming close to the end of our half hour
[9:57] Tao Takashi: just which packets we need to answer and how
[9:57] Locklainn Linden: exactly
[9:58] Locklainn Linden: I'm supposed to be talking withthe interop people about that
[9:58] Infinity Linden: and i'm gonna die if i don't get coffee before my next meeting
[9:58] Infinity Linden: so
[9:58] Infinity Linden: i'm going to sneak off
[9:58] Tao Takashi: and it would be great if we can add this to a general repository of packets and their meaning
[9:58] Infinity Linden: and rejoin you all in #pyogp
[9:58] Locklainn Linden: ok infinity
[9:58] Enus Linden: ditto. will post transcript shortly...
[9:58] Infinity Linden: cheers all!
[9:58] Tao Takashi: not in the pyogp namespace but the general one
[9:58] JB Kraft: adios!
[9:58] Saijanai Kuhn: a word on presence... To keep presence, all you need to do as ack incoming packets. wheee
[9:59] Locklainn Linden: right
[9:59] Locklainn Linden: Tao,
[9:59] Tao Takashi: ok, seems not to be so hard then
[9:59] Locklainn Linden: what do you mean?
[9:59] Locklainn Linden: add what to general repo?
[9:59] Tao Takashi: but still the question is how we setup the udp loop so that it actually answers things in time but you can still hook some thing into it
[10:00] Tao Takashi: isn't there some list of packets somewhere on the wiki?
[10:00] Locklainn Linden: yea
[10:00] Saijanai Kuhn: I though we were using eventlet for that?
[10:00] Locklainn Linden: but not very helpful
[10:00] Tao Takashi: If they are fully defined then ok but I guess they aren't
[10:00] Locklainn Linden: Sai, yes eventlet should be able to do that
[10:00] Saijanai Kuhn: messages
[10:00] Tao Takashi: sure but I wouldn't like to bind it to eventlet
[10:00] Locklainn Linden: what sort of things do you imagine we are going to hook into it?
[10:00] Locklainn Linden: input and the like?
[10:00] Tao Takashi: so we maybe need some way to allow anything to hook into it
[10:00] Saijanai Kuhn: aside from wwhat various people have added, the wiki just lists the message_tmplate list, one per page
[10:01] Tao Takashi: so maybe we can add to those packet descriptions as we go and explore them
[10:01] Locklainn Linden: essentially, in order for the UDP and HTTP stuff to work together, we have to have non-blocking input
[10:01] Tao Takashi: so as a sideeffect we have more documentation created
[10:01] Locklainn Linden: so, unfortunately, we have only use libs that can handle that
[10:01] Locklainn Linden: YES Tao
[10:01] Locklainn Linden: we DEF need to do that
[10:01] Locklainn Linden: as far as LL agrees
[10:01] JB Kraft: isnt asyncore under the zope stuff?
[10:01] Tao Takashi: well, I would first then start with functions which do the handling. And some loop needs to call them then
[10:02] Saijanai Kuhn: https://wiki.secondlife.com/wiki/Category:Messages
[10:02] Locklainn Linden: right
[10:02] Tao Takashi: there are many things to do async stuff, twisted, eventlet, asyncore etc.
[10:02] JB Kraft: right, ok
[10:02] Locklainn Linden: hmm
[10:02] Saijanai Kuhn: I was thinking we needed to handle: UDP, EQG, r-ttp for OGP and rhht for external comands
[10:02] Locklainn Linden: actually, threading would solve this, no?
[10:02] Tao Takashi: and it would be cool if the application can choose which one to usd
[10:02] Tao Takashi: use
[10:02] Saijanai Kuhn: r-http*
[10:03] Enus Linden: transcript needs to end in a sec unless sai can stay
[10:03] Tao Takashi: threading might also be an option, just run it in a different thread and send it messages
[10:03] Saijanai Kuhn: I can stay
[10:03] Enus Linden: ty
[10:03] Locklainn Linden: right
[10:03] JB Kraft: if i rem, asyncore can handle UDP and threads
[10:03] Enus Linden: lates y'all
[10:03] Locklainn Linden: bye Enus
[10:03] Saijanai Kuhn: Donovan and which were suggesting threading on Python was really icky
[10:03] Tao Takashi: but as said, we can start with those functions and the easiest thing then would to explain how you handle this in some loop
[10:03] Tao Takashi: we can also ship an example
[10:03] Locklainn Linden: right
[10:03] Locklainn Linden: for information,
[10:03] Tao Takashi: yes
[10:03] Locklainn Linden: there are only 2 things to handle
[10:04] Locklainn Linden: UDP and TCP
[10:04] Locklainn Linden: TCP is the reverse http,
[10:04] Locklainn Linden: aka, eventqueue
[10:04] Tao Takashi: this might also get a bit tricky ;-)
[10:04] Saijanai Kuhn: also an external command script could use rhttp
[10:04] Locklainn Linden: right
[10:05] Locklainn Linden: and, both AD and sim has eventqueue
[10:05] Tao Takashi: yep
[10:05] Locklainn Linden: but only sim has UDP messages, yes?
[10:05] Tao Takashi: does LL have some example rhttp implementation one can test against?
[10:05] Locklainn Linden: have u seen the wiki?
[10:05] Locklainn Linden: i think donovan has a script using eventlet
[10:05] Tao Takashi: well, except the Donovan script ;-)
[10:05] Locklainn Linden: haha
[10:05] Locklainn Linden: then probably not
[10:06] Tao Takashi: this actually does not run for me with that hub stuff
[10:06] Tao Takashi: I also dissected it into a server and client sciprt
[10:06] Tao Takashi: so it's better understandable
[10:06] Locklainn Linden: but still doesnt run?
[10:06] Saijanai Kuhn: hmmm. DOnt' sims talk to each other with UDP, or is it all TCP now?
[10:06] Tao Takashi: it does when I commend this hub stuff out
[10:06] Locklainn Linden: oh
[10:06] Tao Takashi: no idea what it actually is good for ;-)
[10:06] Locklainn Linden: not sure Sai
[10:06] Tao Takashi: anyway, I gotta go
[10:06] Tao Takashi: more on the list
[10:06] Locklainn Linden: alright
[10:07] Tao Takashi: cya later!
[10:07] Locklainn Linden: cya Tao
[10:07] Locklainn Linden: Sai, you said you were going to post the transcript?
[10:07] Saijanai Kuhn: yar