Pyogp/Chat Logs/Daily Meeting/18 Aug 2008
< Pyogp
Jump to navigation
Jump to search
- [9:29] Saijanai Kuhn: Hey Lock
- [9:29] Multi Gadget: v2.0.3b by Timeless Prototype, '/44 info'
- [9:31] Tao Takashi: Hi
- [9:31] Locklainn Linden: hi everyone
- [9:31] Tao Takashi: is lagged
- [9:31] Locklainn Linden: sorry, didn't have chat open :P
- [9:32] Enus Linden: hola
- [9:32] Locklainn Linden: hi enus
- [9:32] Tao Takashi: hola enus
- [9:32] Tao Takashi: so [1] is now defined as finished
- [9:33] Tao Takashi: as finished as it can be that is taking into account that I am building on a protocol which can change any day ;-)
- [9:33] Tao Takashi: Hi Infinity
- [9:33] Infinity Linden: lol
- [9:33] Infinity Linden: i came in right as tao said "can change any day"
- [9:33] Infinity Linden: which
- [9:33] Infinity Linden: sadly is the truth
- [9:33] Tao Takashi: is now?
- [9:33] Infinity Linden: though
- [9:34] Tao Takashi: that's ok, I put the proof of concept label on it ;-)
- [9:34] Infinity Linden: we're thinking that draft 3 will be fairly stable
- [9:34] Tao Takashi: that's even worse than change any day, it's more like "maybe does not work at all"
- [9:34] Saijanai Kuhn: yippie
- [9:34] Saijanai Kuhn: or not...
- [9:35] Enus Linden: so shall we run around the group?
- [9:35] Saijanai Kuhn: so this domain is as functional as the current aditi AD is?
- [9:35] Infinity Linden: suer
- [9:35] Enus Linden: i'll go. i sent an email to the list, i think i owe another
- [9:35] Infinity Linden: /dt 1 white meetin'
- [9:35] Enus Linden: and need to read and respond as needed to the list
- [9:35] Enus Linden: that's all for me
- [9:36] Tao Takashi: Sai: should be which means it logs you in and transfers the name to the region
- [9:36] Tao Takashi: but it does not do TP right now I think
- [9:36] Locklainn Linden: I've been working on the message refactor
- [9:36] Infinity Linden: right... that's what i've seen
- [9:36] Locklainn Linden: looking at Tao's code, and my own (which is in a branch)
- [9:36] Infinity Linden: login works fine, tp is still an issue
- [9:36] Saijanai Kuhn: ah, OK, so you get presence as in you can see the grid and avies on the grid can see you?
- [9:36] Locklainn Linden: and I'm writing up an official report/form of the problems teh refactor intends to solve
- [9:37] Tao Takashi: very cool, Locklainn
- [9:37] Enus Linden: ty locklainn
- [9:37] Saijanai Kuhn: Lock, should suupporting legacy login and TP part of that design goal or not?
- [9:37] Tao Takashi: Sai: I think others can see me but I never tested this ;-)
- [9:37] Locklainn Linden: hmm
- [9:37] Locklainn Linden: good question
- [9:37] Locklainn Linden: I think its a good goal
- [9:37] Enus Linden: legacy login is not part of the refactor at all is it?
- [9:37] Tao Takashi: isn't legacy more the thing which comes before any UDP happens?
- [9:37] Locklainn Linden: not currently enus
- [9:38] Locklainn Linden: yea, good point tao
- [9:38] Tao Takashi: basically sending the XML-RPC login request and getting sim etc from that?
- [9:38] Enus Linden: the refactor is a separate body of work imo
- [9:38] Locklainn Linden: the message refactor is really for udp communication
- [9:38] Enus Linden: and can be a separate efforty
- [9:38] Enus Linden: efoort*
- [9:38] Locklainn Linden: right
- [9:38] Tao Takashi: I think we can tackle legacy separatly then
- [9:38] Saijanai Kuhn: right though there is no AD in the structure.
- [9:38] Enus Linden: :P
- [9:38] Saijanai Kuhn: if its just for the messaging system, shouldn't make any difference
- [9:38] Locklainn Linden: right
- [9:38] Tao Takashi: at least for now
- [9:39] Tao Takashi: if messaging changes in OGP then we maybe need to think about how to support it
- [9:39] Locklainn Linden: of course
- [9:39] Locklainn Linden: I'm not sure how it would change that would break our system
- [9:39] Locklainn Linden: maybe flags and such?
- [9:39] Saijanai Kuhn: EQG vs rhttp, but EQG of some kind will be the fallback anyway
- [9:39] Locklainn Linden: ye
- [9:39] Locklainn Linden: yea*
- [9:39] Tao Takashi: no SuperIM messages ;-)
- [9:39] Tao Takashi: or what it was called
- [9:40] Tao Takashi: Improved..
- [9:40] Tao Takashi: although it does not sound like improved
- [9:40] Saijanai Kuhn: ImprovedInstantMessage
- [9:40] Tao Takashi: I wonder what the non-improved version looked like then ;-)
- [9:40] Saijanai Kuhn: scary thought
- [9:40] Tao Takashi: but anyway, maybe let's go on
- [9:40] Locklainn Linden: yep
- [9:40] Saijanai Kuhn: https://wiki.secondlife.com/wiki/ImprovedInstantMessage
- [9:40] Locklainn Linden: Shrek's turn?
- [9:41] Tao Takashi: huh?
- [9:41] Locklainn Linden: you are shrek
- [9:41] Locklainn Linden: :)
- [9:41] Locklainn Linden: never seen shrek?
- [9:41] Tao Takashi: I am not that fat ;-)
- [9:41] Locklainn Linden: haha
- [9:41] Tao Takashi: and wasn't he green?
- [9:41] Infinity Linden: nor that green
- [9:42] Tao Takashi: notes that Locklainn is color blind
- [9:42] Locklainn Linden: hmm
- [9:42] Tao Takashi: anyway, I was working on experimenting with networking stuff at the weekend
- [9:42] Locklainn Linden: seems green to me
- [9:42] Infinity Linden: it' okay... i can't see the color purple
- [9:42] Locklainn Linden: oh, yea, guess he is yellow
- [9:42] Locklainn Linden: looks green from over here, anyway
- [9:42] Infinity Linden: adjust teh tint on your monitor?
- [9:42] Tao Takashi: so I was first trying to get into networking at all as I hardly did that the last couple of years ;-)
- [9:43] Tao Takashi: and looked then into twisted and how you would do things with twisted compared to eventlet and a threaded approach
- [9:43] Tao Takashi: which resultet in long discussions with Locklainn in which we hopefully now are heading to a good solution
- [9:43] Enus Linden: aweosme
- [9:43] Infinity Linden: freaky... but definitely interesting
- [9:43] Enus Linden: awesome*
- [9:43] Tao Takashi: to let the user decide which system to use
- [9:43] Saijanai Kuhn: confusing to moi, but I made a transcript
- [9:43] Tao Takashi: I also sent some emails to the list regarding how to do logging and about the dev process
- [9:44] Tao Takashi: we hope to explain this in some docs in more detail
- [9:44] Tao Takashi: with examples
- [9:44] Tao Takashi: I guess we are doing that at least ;-)
- [9:44] Locklainn Linden: nice
- [9:44] Tao Takashi: as for logging I wanted to implement some example in e.g. the caps modules
- [9:44] Tao Takashi: module
- [9:44] Tao Takashi: as I also would need better logging for that agent domain
- [9:44] Locklainn Linden: I'm curious,
- [9:44] Locklainn Linden: why is there a logger per module?
- [9:44] Tao Takashi: hard to find out now what happened when you only see a 500 in the access log
- [9:44] Saijanai Kuhn: a reminder to all: put the category:pyogp_kitchen_sink in any page that way we can find it no matter what part of the hierarchy its in
- [9:44] Locklainn Linden: you can't set it then to a level
- [9:44] Locklainn Linden: for all logging
- [9:45] Tao Takashi: it's per module and level
- [9:45] Tao Takashi: if you listen to "pyogp" and error you only get all errors and critical messages for the whole lib
- [9:45] Tao Takashi: you might set up a different logger for pyogp.lib.base.message to only get log statements from the message system
- [9:45] Locklainn Linden: so it works for all the loggers in every module?
- [9:46] Tao Takashi: yes, if you remove parts from the path you get more stuff
- [9:46] Tao Takashi: it just collects it
- [9:46] Tao Takashi: a path is splitted by .
- [9:46] Locklainn Linden: I see
- [9:46] Locklainn Linden: ok
- [9:46] Tao Takashi: I think that's quite nice
- [9:46] Tao Takashi: so I hope to do an example soon. I would suggest though not to do such smaller changes in a branch
- [9:47] Tao Takashi: it's still easy to remove those statements again anyway and I won't do the whole lib in one go
- [9:47] Tao Takashi: this removes the burden of merging then which is not really fun with svn
- [9:47] Locklainn Linden: hmm
- [9:47] Locklainn Linden: I don't know about that
- [9:47] Locklainn Linden: it may be small
- [9:47] Tao Takashi: and I will take care that all the tests still run of course
- [9:47] Locklainn Linden: but a single line could break stuff
- [9:48] Tao Takashi: that's why I take care that the tests run
- [9:48] Locklainn Linden: just saying for future reasons
- [9:48] Locklainn Linden: I know logging won't
- [9:48] Locklainn Linden: I guess this is a question of the development process
- [9:48] Locklainn Linden: for me
- [9:48] Tao Takashi: yep, we maybe should discuss this further outside this meeting
- [9:49] Tao Takashi: I also would like some feedback on exceptions btw so I can put them in place :)
- [9:49] Locklainn Linden: yea,
- [9:49] Tao Takashi: maybe this can be done with logging in one go with caps as an example
- [9:49] Locklainn Linden: enus, inf, don't we have some docs being made up for exceptions?
- [9:49] Infinity Linden: yup
- [9:49] Enus Linden: i've paid no attention to that yet myseld
- [9:49] Locklainn Linden: any idea when they will be ready?
- [9:49] Infinity Linden: what we're doing is documenting the "exception" and "error" responses from each message
- [9:50] Locklainn Linden: right
- [9:50] Tao Takashi: cool
- [9:50] Infinity Linden: internally... i can get you something that's vaguely stable
- [9:50] Infinity Linden: but it needs another round of review before publishing
- [9:50] Infinity Linden: publicly
- [9:50] Locklainn Linden: so, Tao, probably have to wait for this
- [9:50] Tao Takashi: well, I at least should be able to catch connection errors and such
- [9:50] Locklainn Linden: yea
- [9:50] Infinity Linden: but
- [9:50] Tao Takashi: you can always add more exceptions for more detailed errors
- [9:51] Infinity Linden: things that are stable:
- [9:51] Infinity Linden: * we're definitely moving from 302 to 200
- [9:51] Locklainn Linden: well, they may change too
- [9:51] Infinity Linden: for both auth failures and auth successes
- [9:51] Locklainn Linden: like that :)
- [9:51] Infinity Linden: and
- [9:51] Infinity Linden: i think we're removing legacy login from the specification
- [9:51] Tao Takashi: that doesn't reflect the exception though, it might always be UserNothAuthorized or so
- [9:51] Infinity Linden: and
- [9:52] Tao Takashi: just when it's raised is different
- [9:52] Infinity Linden: well
- [9:52] Infinity Linden: you acn handle it how you like
- [9:52] Infinity Linden: but
- [9:52] Infinity Linden: if you do not provide the correct credentials
- [9:52] Infinity Linden: you'll get a distinct message in return
- [9:52] Tao Takashi: this can be put into the exception then
- [9:52] Infinity Linden: for things like maintenance
- [9:53] Infinity Linden: and whether or not the authentication failure was due to improper login options, etc
- [9:53] Tao Takashi: we need to discuss how much detail we want or need to model in exceptions anyway
- [9:53] Infinity Linden: the place_avatar "family" of calls are also mostly stable
- [9:53] Tao Takashi: but I would be in favor of replacing those string exceptions or prints we have already with some class based exception
- [9:53] Locklainn Linden: well, Tao, I don't think that is up to us, no?
- [9:53] Locklainn Linden: isn't that up to OGP spec?
- [9:54] Infinity Linden: but Tess and Leyla have been working mostly with them
- [9:54] Infinity Linden: no no... i think he's thinking of how the exceptions are reported in PyOGP
- [9:54] Locklainn Linden: ohh ok
- [9:54] Tao Takashi: Locklainn: how we implement it is up to us I think. The difference is if we have one special exception for every detail or if we put e.g. some reason inside the exception
- [9:54] Tao Takashi: yep
- [9:54] Locklainn Linden: oh ok
- [9:54] Locklainn Linden: I gotcha
- [9:54] Infinity Linden: so yer talking about having a generic exception
- [9:54] Tao Takashi: and if we have 10 detailed exceptions they probably derive from one more generic one
- [9:54] Locklainn Linden: so same exception, but presented differntly
- [9:54] Locklainn Linden: gotcha
- [9:55] Infinity Linden: yup
- [9:55] Tao Takashi: some exception tree,
- [9:55] Infinity Linden: what i was thinking too
- [9:55] Locklainn Linden: cool
- [9:55] Tao Takashi: like already started in exc.py
- [9:55] Tao Takashi: so each subcomponent would get it's own super-exception from which it then can derive further exception classes
- [9:55] Tao Takashi: e.g. MessageError to NotAValidBlockError
- [9:56] Tao Takashi: or maybe InvalidBlockError ;-)
- [9:56] Tao Takashi: BlockTooFunkyException and such
- [9:56] Tao Takashi: so we also have internal ones actually
- [9:56] Infinity Linden: right... and service consumers can choose to catch the superclass or the more detailed subclasses
- [9:56] Locklainn Linden: so, just pointing out that I'll be leaving in a week, and I won't be here tomorrow and Wed
- [9:56] Tao Takashi: yep
- [9:56] Locklainn Linden: so I am hoping to get finished this messaging refactoring
- [9:56] Infinity Linden: krikey
- [9:56] Locklainn Linden: and hopefully documenting it
- [9:57] Locklainn Linden: so, I hope we can put a lot of work into this
- [9:57] Locklainn Linden: and bust it out so it can be done before I leave
- [9:57] Saijanai Kuhn: quietly wipes a tear
- [9:57] Tao Takashi: it would be very good indeed to document how this whole thing works, also from a protocol level, not only from implementation
- [9:57] Tao Takashi: yeah, that will be a loss
- [9:57] Enus Linden: 10am, next meeting time
- [9:57] Locklainn Linden: maybe I'm missing it again, isn't that what the OGP spec is?
- [9:58] Infinity Linden: eek.
- [9:58] Infinity Linden: it's getting
- [9:58] Tao Takashi: but the details about acks and such...
- [9:58] Infinity Linden: late
- [9:58] Enus Linden: see yas on irc
- [9:58] Locklainn Linden: oh ok
- [9:58] Locklainn Linden: cya enus
- [9:58] Infinity Linden: lemme just say... i'm working on OGP spec
- [9:58] Infinity Linden: documenting
- [9:58] Tao Takashi: basically what a circuit does
- [9:58] Saijanai Kuhn: making transcript...
- [9:58] Tao Takashi: maybe this can then be part of the OGP spec :)
- [9:58] Infinity Linden: cool
- [9:58] Locklainn Linden: is there a wiki for ur work infinity?
- [9:58] Infinity Linden: not really
- [9:58] Infinity Linden: but
- [9:59] Locklainn Linden: so documenting in doc file or something?
- [9:59] Infinity Linden: wait
- [9:59] Tao Takashi: it's more in XML ;-)
- [9:59] Locklainn Linden: haha
- [9:59] Locklainn Linden: she thinks in XML
- [9:59] Infinity Linden: we're asking people to put comments on the documentation page for teh existing spec
- [9:59] Infinity Linden: well
- [9:59] Infinity Linden: i think in objects
- [9:59] Infinity Linden: XML is just the "common tongue" of middle-grid
- [10:00] Tao Takashi: I guess for now it would be good to have it in any form
- [10:00] Locklainn Linden: :)
- [10:00] Locklainn Linden: ok,
- [10:00] Locklainn Linden: so would it be a bad idea for us to put this stuff on the Pyogp page?
- [10:00] Infinity Linden: and if you don't like it, i'll write it in Quenya or Esperanto
- [10:00] Saijanai Kuhn: Quenya?
- [10:00] Tao Takashi: probably they want to mangle it through the speclanguage machine anyway
- [10:00] Infinity Linden: or Sindarin
- [10:00] Infinity Linden: yup
- [10:00] Saijanai Kuhn: has a 50-100-year-old-book on Esperanto
- [10:00] Saijanai Kuhn: Ah, evlish
- [10:01] Saijanai Kuhn: elvish
- [10:01] Infinity Linden: a lot of us are all SGML types
- [10:01] Infinity Linden: yup
- [10:01] Tao Takashi: Locklainn, there should be some pages about messages and such already on the wiki
- [10:01] Tao Takashi: maybe put it there somewhere
- [10:01] Locklainn Linden: yea
- [10:01] Locklainn Linden: that's what I'm asking
- [10:01] Infinity Linden: elvish SGML + DSSSL freaks
- [10:01] Locklainn Linden: seems the messaging stuff on wiki isn't very doced
- [10:01] Locklainn Linden: so shoudl we put this al there?
- [10:01] Tao Takashi: so it's also useful for the other folks out there
- [10:01] Saijanai Kuhn: there's a page lock wrote on notes, or maybe it was infinity? I think I put it under kitchensink
- [10:01] Locklainn Linden: yea,
- [10:01] Locklainn Linden: notes has message format
- [10:01] Locklainn Linden: currently
- [10:01] Locklainn Linden: and I also have started a protocol page
- [10:02] Locklainn Linden: that lists exceptions and the like
- [10:02] Tao Takashi: maybe it would even be good for LL employees to look at ;-)
- [10:02] Locklainn Linden: maybe should go on a non-pyogp page
- [10:02] Saijanai Kuhn: is it under kitchensink? ;-)
- [10:02] Locklainn Linden: lemme find
- [10:02] Saijanai Kuhn: nothing wrong with pyogp kitchensink referring to non pyogp pages
- [10:02] Locklainn Linden: [2]
- [10:02] Tao Takashi: you can also put links onto some messagesystem project page under pyogp
- [10:03] Locklainn Linden: haha
- [10:03] Locklainn Linden: this is what I'm asking
- [10:03] Locklainn Linden: where should this stuff be?
- [10:03] Locklainn Linden: pyogp page
- [10:03] Tao Takashi: which point outwards
- [10:03] Locklainn Linden: general messaging page
- [10:03] Locklainn Linden: I know they CAN go in tehse places
- [10:03] Locklainn Linden: but where SHOULD they go
- [10:03] Infinity Linden: yup
- [10:03] Tao Takashi: if it's in general about the protocol put it somewhere close to other related pages outside pyogp
- [10:03] Tao Takashi: and link from pyogp to it
- [10:03] Tao Takashi: I think that way people can find it best
- [10:03] Infinity Linden: that page is good... we needed more of an overview for teh spec
- [10:04] Infinity Linden: we have it on our to-do list, but its pretty far down
- [10:04] Infinity Linden: also
- [10:04] Locklainn Linden: the page I just linked inf?
- [10:04] Infinity Linden: yup
- [10:04] Infinity Linden: but
- [10:04] Tao Takashi: reminds me that I wanted to start my book "OGP explained"
- [10:04] Infinity Linden: auth is going to be changing a bunch
- [10:04] Infinity Linden: lol
- [10:04] Infinity Linden: and
- [10:04] Locklainn Linden: yea,. that hasn't been updated in months
- [10:04] Infinity Linden: everything is going to be specified in terms of LLIDL
- [10:04] Infinity Linden: so
- [10:04] Saijanai Kuhn: wah. Now I have to find a new title
- [10:04] Saijanai Kuhn: OGP for Dummies?
- [10:04] Infinity Linden: headers goes away
- [10:04] Tao Takashi: OGP in a nutshell
- [10:05] Saijanai Kuhn: OGP for the Cmpleat Idiot
- [10:05] Infinity Linden: and at the protocol level, you always get a 200
- [10:05] Tao Takashi: OGP Hacks :)
- [10:05] Infinity Linden: OGP for fun and profit?
- [10:05] Locklainn Linden: :)
- [10:05] Tao Takashi: OGP for the insane
- [10:05] Locklainn Linden: well
- [10:05] Locklainn Linden: eitheer way,
- [10:05] Locklainn Linden: I'm going to put it on that page then
- [10:05] Infinity Linden: "I implemented OGP and all i got was this bear?"
- [10:05] Locklainn Linden: and you guys can move it once I'm good and dead at Linden
- [10:05] Locklainn Linden: :)
- [10:05] Locklainn Linden: hsahaaha
- [10:05] Locklainn Linden: nice
- [10:05] Infinity Linden: lol
- [10:05] Locklainn Linden: but they love it
- [10:05] Saijanai Kuhn: wipes another tear
- [10:05] Tao Takashi: I didn't even get a bear ;-)
- [10:06] Locklainn Linden: its like teh gold medal that people work years to get
- [10:06] Locklainn Linden: its just virtual :)
- [10:06] Tao Takashi: but my OGP avi cannot store the bear anyway ;-)
- [10:06] Infinity Linden: okay
- [10:06] Infinity Linden: gotta
- [10:06] Infinity Linden: run
- [10:06] Tao Takashi: cya Infinity!
- [10:06] Infinity Linden: cheers
- [10:06] Locklainn Linden: ok, me too, cya everyone
- [10:06] Tao Takashi: me too ;-)
- [10:06] Tao Takashi: cya all!
- [10:07] Saijanai Kuhn: K posting transcript