Chat log from the meeting on 2019-09-10

From OpenSimulator

Jump to: navigation, search

[11:07] Andrew.Hellershanks I can't run the version of the BoM viewer Bill supplied. It uses libraries that are no on my system
[11:08] Ubit Umarov: as we said last meeting he is wearing a simple mesh avatar
[11:08] Kayaker Magic: I was hoping Gavin would be here. I have questions about the Raspberry Pi OpenSim
[11:08] Ubit Umarov: that with BoM will show the system baked textures on those color areas
[11:08] Andrew.Hellershanks Yes, I remember. As soon as I saw the funky colours I remembered about the newer viewers.
[11:09] Ubit Umarov: that thing in center table is the same mesh with the textures with colors
[11:09] Ubit Umarov: a colour per each part that can be replaced by the avatar baked textures
[11:10] Ubit Umarov: that is the basic set..
[11:10] Ubit Umarov: that basic BoM should work on most regions
[11:10] Ubit Umarov: well with proper viewer..
[11:10] Kayaker Magic: Does BOM work with mesh bodies?
[11:11] Bill.Blight That is the main point
[11:11] Ubit Umarov: what do you think the M on BoM means? :p
[11:11] Bill.Blight that is why I'm wearing a mesh avatar
[11:11] Ubit Umarov: thats Bakes On Mesh
[11:12] Ubit Umarov: something ll did deploy on all their regions a few weeks ago
[11:12] Bill.Blight It currently does not work on anything not attached to an avatar
[11:12] Ubit Umarov: walking around busy areas with older viewer we see several like bill is now
[11:13] Ubit Umarov: yeah and silly, to even set the mesh textures one needs to be wearing it
[11:13] Andrew.Hellershanks Hm... Firestorm must have changed the location of their viewer code. I get a 403: forbidden message when I try to do an hg pull,.
[11:13] Ubit Umarov: i mean to set to the BoM magic texture ids
[11:13] Ubit Umarov: its hg.firestorm.. now
[11:14] Ubit Umarov: it was hg.phoenix...
[11:14] Bill.Blight
[11:14] Ubit Umarov: i had same issue
[11:14] Andrew.Hellershanks hg is the command used to pull. I was trying to update my copy of the code repo that I last updated in July of last year.
[11:15] Ubit Umarov: plug as we said the repo url changed
[11:15] Andrew.Hellershanks I figured that out from the 403: forbidden message.
[11:15] Bill.Blight I just posted the link to the new one
[11:15] Bill.Blight see above
[11:15] Ubit Umarov: how nice you did, well done
[11:15] Ubit Umarov: :p
[11:16] Bill.Blight well actually that is a link to all the repos
[11:16] Bill.Blight :P
[11:16] Andrew.Hellershanks Are you using a viewer based on the pre-release code?
[11:16] Kayaker Magic: This is probably a question about MONO: Doesn't it "compile" into an intermediate 'P' code, so anything that runs on an x86 mono system should work on an ARM system?
[11:16] Bill.Blight No
[11:16] Bill.Blight
[11:16] Bill.Blight the most updated
[11:17] Ubit Umarov: that was a big reasoning jump :p
[11:17] Ubit Umarov: ( and failed one ;) )
[11:17] Ubit Umarov: well about opensim..
[11:17] Ubit Umarov: you may had noticed i disabled avatar hover again
[11:18] Andrew.Hellershanks I'm getting the phoenix-firestorm-lgpl-os version as it says it is for OS.
[11:18] Ubit Umarov: a) opensim code wasn't still good
[11:18] Ubit Umarov: b) viewers code is bad
[11:18] Bill.Blight I found a better way for low energy systems and still stay on x86 ... Use old thin clients or embedded systems that have x86 CPUs , still low power but don't have to re-invent the wheel, and also still have plenty of I/O
[11:19] Ubit Umarov: my work copy kinda has a) fixed, but i made several other changes
[11:19] Bill.Blight and they are still a lot faster than PIs
[11:19] Andrew.Hellershanks Bill, will that branch still works on OpenSim?
[11:19] Bill.Blight yes
[11:19] Ubit Umarov: still doing basic testing on those
[11:19] Bill.Blight they have not removed the OS stuff
[11:19] Bill.Blight YET
[11:19] Andrew.Hellershanks :)
[11:20] Ubit Umarov: yes and FS team is considering merging it all again
[11:20] Ubit Umarov: they still did not reach a decision, but considering it
[11:20] Bill.Blight or as an option ...
[11:20] Bill.Blight
[11:20] Kayaker Magic: Bill, what are the specifications of your old embedded systems?
[11:20] Ubit Umarov: the split also has own issues
[11:21] Ubit Umarov: the chicken/egg beeing one
[11:21] Bill.Blight currently using
[11:21] Bill.Blight one of those
[11:21] Ubit Umarov: ( good arielle arrived after me telling i did disabled avatar hover
[11:21] Bill.Blight got i for 30$
[11:21] Ubit Umarov: Ohh damm i said that?
[11:22] Arielle.Popstar why you do that?
[11:22] Ubit Umarov: read the meeting log :p
[11:24] Andrew.Hellershanks We talked a bit more about BoM. Who wants to start us on the next topic for todays meeting?
[11:25] Arielle.Popstar what was the result of the BoM talk?
[11:26] Arielle.Popstar still backwardly incompatible in core?
[11:26] Ubit Umarov: basicly a recap of last meeting
[11:26] Ubit Umarov: no way around that
[11:26] Ubit Umarov: ( and still nothing on core )
[11:26] Arielle.Popstar use a different method?
[11:26] Ubit Umarov: stop firing dust into the air..
[11:26] Arielle.Popstar there are 3 different methods now
[11:27] Arielle.Popstar yours, Lanani/Hy and a scripted version
[11:27] Ubit Umarov: mine?
[11:27] Ubit Umarov: well whatever
[11:27] Arielle.Popstar well the one you are proposing
[11:27] Bill.Blight There is only ONE method that is compatible with the viewer
[11:27] Andrew.Hellershanks 3 different methods of BoM??
[11:28] Arielle.Popstar yes andrew
[11:29] Andrew.Hellershanks I would have expected only one version in the OS code repo. Are you talking viewer side, or customized versions of OS?
[11:29] Ubit Umarov: im not proposing anything
[11:29] Ubit Umarov: im seeing a tpv feature that may be good to support
[11:29] Arielle.Popstar forked os and a scripted version
[11:29] Ubit Umarov: it is a TPVs feature, or may be
[11:30] Arielle.Popstar the BoM you mean or a different feature?
[11:30] Ubit Umarov: we where talking about BoM
[11:30] Arielle.Popstar ok
[11:30] Andrew.Hellershanks The phoenix repos are excruiatingly slow (or its an HG issue). 18 minutes to go on downloading the viewer code. :P
[11:31] Bill.Blight Viewers only support the SL version of BOM for editing, so if OS needs to work with existing viewers, then it must go the SL/LL/FS route
[11:31] Arielle.Popstar have you chatted to any tpv devs?
[11:31] Ubit Umarov: well actually viewers will need to change code for opensim
[11:31] Ubit Umarov: as it may kill regions also
[11:32] Ubit Umarov: it was a surprise to some tpvs devs that the bakes do actually work without SSA/SSB
[11:33] Ubit Umarov: but the way bakes code was made, the viewer side bakes also do the new layers as soon they where defined
[11:33] Ubit Umarov: just a few glithcs
[11:34] Arielle.Popstar baked textures been accessable in viewers for ages
[11:34] Ubit Umarov: and must check if regions do suport them
[11:34] Arielle.Popstar any cb viewer can see that
[11:35] Ubit Umarov: ok show me your AUX3 bake please
[11:35] Ubit Umarov: bahh
[11:35] Arielle.Popstar not on a cb viewer
[11:36] Arielle.Popstar but opensim has an ossl function for them anyway
[11:36] Ubit Umarov: well and that is irrelevant actually
[11:36] Arielle.Popstar anyone moving on
[11:36] Ubit Umarov: hmm that rememinds me i need to check the npc support functions
[11:36] Ubit Umarov: but those can be a later thing
[11:37] Ubit Umarov: and arielle you are talking about different things
[11:37] Ubit Umarov: if you still don't understand BoM learn about it
[11:38] Arielle Popstar: ok
[11:38] Ubit Umarov: it does more than just putting a static texure ID on a face
[11:38] Arielle Popstar: nod
[11:41] Ubit Umarov: gavin remember me to tell you a viewers bug on any hover suport
[11:41] Ubit Umarov: ( for opensim ie no SSA code paths )
[11:41] Gavin.Hird I just had to watch the Apple announcements
[11:42] Gavin.Hird right Ubit, I just disabled for the next version
[11:42] Ubit Umarov: funny om FS a chain of odd code does let shape hover work fine
[11:43] Arielle Popstar: its not working here'
[11:43] Ubit Umarov: well prob is simple and fix also
[11:43] Gavin.Hird ok
[11:43] Gavin.Hird the SL or OS version of FS?
[11:43] Arielle Popstar: os
[11:43] Ubit Umarov: viewers send the avatar body_size on setappearence packet right ?
[11:43] Arielle Popstar: sl version doesnt work in os
[11:44] Ubit Umarov: all recent opensim versions just use that for physics collision shape
[11:44] Gavin.Hird sorry I was thinking SL or OS repository
[11:44] Ubit Umarov: bc all old code was just a fail..
[11:44] Gavin.Hird as there are a half a year difference between them
[11:44] Bill Blight: I'm using the SL repository for the custom test viewer, FS has not removed OS support from that branch yet
[11:45] Gavin.Hird exactly Bill
[11:45] Ubit Umarov: to get the actual body size one does need to compute all the avatar joints blablabla
[11:45] Ubit Umarov: and viewers do that very well and do tell regions..
[11:45] Royale.Mobian I am using the viewer bill mentioned now actually
[11:45] Ubit Umarov: BUT
[11:45] Kayaker Magic: Oh Gavin: That OpenSim Mantis I reported to you about llListen on the Raspberry Pi version: I could not get it to repeat either, didn't have to reboot it just healed itself.
[11:45] Ubit Umarov: before sending they add the offset, breaking it all
[11:46] Gavin.Hird magic at play Kayaker :-))
[11:46] Ubit Umarov: fs now is sending the avatar hover added to the body_size
[11:46] Ubit Umarov: that means physics can even get a negative size
[11:47] Kayaker Magic: I was running 3 rezzers at once, each one creating objects 5 times a second, each that called llListen. I may have exceeded some other system resource.
[11:47] Arielle Popstar: does physics matter if posing?
[11:47] Ubit Umarov: and even if not, that means that the final offset is 1.5 what should be
[11:47] Ubit Umarov: and the avatar physics shape is larger than in similar case at sl
[11:47] Ubit Umarov: fix is simple:
[11:48] Ubit Umarov: well when i find the file i tell you :p
[11:48] Arielle Popstar: waits for the simple fix that likely results in breakage
[11:48] Gavin.Hird goody, goody
[11:48] Ubit Umarov: / correct for the collision tolerance (to make it look like the
<Ubit> // agent is actually walking on the ground/object)
<Ubit> // NOTE -- when we start correcting all of the other Havok geometry
<Ubit> // to compensate for the COLLISION_TOLERANCE ugliness we will have
<Ubit> // to tweak this number again
<Ubit> const LLVector3 body_size = gAgentAvatarp->mBodySize;
<Ubit> msg->addVector3Fast(_PREHASH_Size, body_size);
[11:48] Ubit Umarov: this is how code should read
[11:49] Ubit Umarov: on file agent.cpp
[11:49] Ubit Umarov: those coments are just old, not valid anymore
[11:49] Ubit Umarov: there should be no offset added to body_size there
[11:49] Gavin.Hird noted
[11:49] Ubit Umarov: ... ok that is the viewer side problem on avatar hover
[11:50] Ubit Umarov: region side issues, i will fix one day :)
[11:50] Gavin.Hird one day :-)
[11:50] Ubit Umarov: ( actually already have it fixed :p )
[11:50] Ubit Umarov: just in the middle of a lot of other bugs
[11:51] Gavin.Hird ok
[11:52] Andrew Hellershanks: When are you not in the middle of a lot of bugs, Ubit?  :)
[11:52] Ubit Umarov: im changing a bit the appearence code and bakes processing
[11:52] Ubit Umarov: to allow for 11 bakes (and 45 textures on viewers packets )
[11:53] Ubit Umarov: as i said, teleports to regions will be blocked if the destination does not support current outfit
[11:53] Ubit Umarov: user can just change it and try again
[11:54] Ubit Umarov: ( in theory :p )
[11:54] Andrew Hellershanks: The user will receive a message telling them why they can't TP?
[11:54] Ubit Umarov: yes
[11:54] Ubit Umarov: to or crossing
[11:54] Ubit Umarov: tp or..
[11:55] Andrew Hellershanks: Good. I'm sure there will be some who won't read the message or won't understand it but that will be their problem.
[11:55] Ubit Umarov: ( gavin file is of course llagent.cpp not agent.cpp )
[11:56] Gavin.Hird it is always ll, or fs for special code they have added
[11:56] Ubit Umarov: yeap
[11:57] Ubit Umarov: that send of appearance is opensim code only now
[11:57] Ubit Umarov: SSA does not use that of course
[11:58] Ubit Umarov: so no need for conditional testing there
[11:58] Ubit Umarov: well on FS the handling of avatar hover is so confusing... gezz
[11:59] Ubit Umarov: but well viewer devs pains not mine :p
[11:59] Ubit Umarov: ok i made you all fall asleep =
[11:59] Ubit Umarov: ?
[12:00] Ubit Umarov: Andrew understand what osSlerp is now ? :)
[12:00] Andrew Hellershanks: We are at the top of the hour. Any other topic for today before we wrap things up?
[12:00] Gavin.Hird I really never check the FS code, so I have no idea of how they handle it. There are a few small sections of code such as the AO, the profile and search that has FS code in Dayturn.
[12:01] Ubit Umarov: i did add a auxiliar math funtion: osSlerp
[12:01] Ubit Umarov: that is a interpolation thing for rotations
[12:01] Andrew Hellershanks: Ubit, Not really. I did start reading about it but I'm not into graphics theory. I don't know how, or even if, I would ever use it.
[12:01] Ubit Umarov: it is just that.. interpolation
[12:02] Kayaker Magic: linear or angular?
[12:02] Ubit Umarov: linear interpolation will not work well on rotations
[12:02] Ubit Umarov: Spherical
[12:02] Ubit Umarov: it could be done with other LSL
[12:02] Andrew Hellershanks: I just set rotation on an object when I need to rotate it, or use omega to make it spin.
[12:02] Gavin.Hird yes, the documentation and osslEnable.ini records are missing on that one
[12:02] Kayaker Magic: Yeah, I have done it linear in LSL, that was good enough for my use at the time.
[12:02] Ubit Umarov: well i may take all those out
[12:03] Ubit Umarov: always forget to update it :)
[12:03] Gavin.Hird ;-)
[12:03] Ubit Umarov: ( but also forgot to update syntaxe.xml )
[12:03] Ubit Umarov: i did add it to our Wiki :p
[12:03] Gavin.Hird that is what I meant by documentation
[12:04] Gavin.Hird Andrew does the website, right?
[12:04] Andrew Hellershanks: I wouldn't know how to create a web page for osSlerp that could provide an example. I can only put it in the wiki as is.and say the function exists.
[12:04] Ubit Umarov:
[12:04] Andrew Hellershanks: Gavin, I have been updating the wiki pages now and then as new things get added.
[12:04] Gavin.Hird yes
[12:05] Gavin.Hird thanks Ubit, then I can update the viewer too
[12:05] Ubit Umarov: well one day someone will translate that wiki entry to english :p
[12:05] Andrew Hellershanks: ok. That page just needs a minor update to point out exactly when it was added.
[12:06] Ubit Umarov: there is no need for more than what is there already
[12:06] Ubit Umarov: speciall no GIT commits
[12:06] Ubit Umarov: those pages are for final users
[12:06] Andrew Hellershanks: I prefer to point out which version of OS includes the change, a date, and/or commit #.
[12:07] Ubit Umarov: yeah.. useless info
[12:07] Ubit Umarov: commit # at least
[12:07] Kayaker Magic: I can write an example LSL program to demonstrate it!
[12:07] Ubit Umarov: final users don't care about git
[12:07] Gavin.Hird it is actually already in the PI code Kayaker
[12:08] Andrew Hellershanks: There are people writing code that is being used on different versions of OS. It can be useful as it can help them know if the version of OS they are using supports the function they want to use.for them to know
[12:08] Ubit Umarov: well its just another aux function
[12:08] Ubit Umarov: ppl did the same using other LSL functions
[12:08] Andrew Hellershanks: well, the OS version number and date are useful.
[12:08] Ubit Umarov: this is just a bit cleaner and faster
[12:08] Gavin.Hird absolutely Andrew
[12:09] Ubit Umarov: yeah i forgot the os version :p
[12:09] Kayaker Magic: 0.9.1
[12:09] Andrew Hellershanks: and a more specific date than 2019.
[12:10] Ubit Umarov: it does tell month :p
[12:10] Ubit Umarov: but whatever change as you like, just do not add commit #
[12:11] Andrew Hellershanks: Oh, Set is supposed to be "September"?
[12:11] Ubit Umarov: i said that needs translation to engish didn't i ? :p
[12:11] Ubit Umarov: gezz
[12:11] Andrew Hellershanks: I didn't realize that is what you were referring to when you said it needed a translation as the rest is in English.
[12:12] Gavin.Hird LL just released 137 changesets of bugfixes, which may come in handy
[12:13] Ubit Umarov: unless some of those fixes turn out to be a bug on opensim context :)
[12:13] Bill Blight: You must translate it from portUbitguese to english
[12:13] Andrew Hellershanks grins at Bill
[12:13] Ubit Umarov: yeah Setembro :p
[12:13] Gavin.Hird that is of course always something that must be tested before using them in the viewer
[12:14] Ubit Umarov: seems EEP is a big change
[12:14] Andrew Hellershanks: Gavin, that should keep you busy for a day or two.
[12:14] Ubit Umarov: will break windlight and a few more other things
[12:14] Gavin.Hird that is a major owne
[12:14] Ubit Umarov: guess we will not have that
[12:14] Gavin.Hird indeed :-))
[12:15] Gavin.Hird also the VS2017 rewrite is a major change
[12:15] Ubit Umarov: EEP may be the reason for FS to keep the split
[12:15] Gavin.Hird one of the reasons
[12:15] Ubit Umarov: ohh my local compile of FS failed
[12:15] Ubit Umarov: bc of a bool return type
[12:16] Ubit Umarov: strange it did work before and works for bill
[12:16] Gavin.Hird I just love bool
[12:16] Ubit Umarov: i had to chage the type to int to mk it compile here
[12:16] Ubit Umarov: ( on 2 offending functions )
[12:16] Gavin.Hird the LL defined BOOL is an int
[12:16] Ubit Umarov: linker just failed to find them
[12:17] Gavin.Hird I have split it into a separate file llbool.h
[12:17] Ubit Umarov: linker was looking for int .. and they where bool ..
[12:17] Ubit Umarov: no idea how that was working fine and now failed for me
[12:17] Ubit Umarov: and does work for bill
[12:17] Ubit Umarov: well c++ things
[12:18] Andrew Hellershanks: I can't get autobuild to run in order to start trying to compile latest FS code.
[12:18] Gavin.Hird did you update anyting since last build?
[12:18] Ubit Umarov: i did pull yes
[12:18] Ubit Umarov: but that file is from 2013 no changes since then
[12:18] Gavin.Hird probably some header elsewhere has changed
[12:19] Ubit Umarov: did compile fine etc
[12:19] Ubit Umarov: but linker said NO
[12:19] Ubit Umarov: well i only compiled to test the change on that hover thing i told about
[12:20] Ubit Umarov: had no real need 2 the code was clear
[12:20] Ubit Umarov: and seeing my avatar 4m tall on region data also :)
[12:21] Ubit Umarov: or -0.3m with other set of values
[12:22] Ubit Umarov: min avatar height is around 1.6m with -2 of hover.. boinf .. -0.4m tall.. not funny
[12:24] Andrew Hellershanks: Surprisingly we are almost half past the hour and haven't had a lot of people having to head out yet.
[12:25] Ubit Umarov: ppl did fall asleep
[12:25] Andrew Hellershanks: This is a good time to wrap up the meeting for today.
[12:25] Andrew Hellershanks: We usually lose a number of people just after the hour mark.
[12:25] Andrew Hellershanks: Thank you all for coming and staying this long today. See you all again next week.
[12:28] Ubit Umarov: well testing my changes alone is a pain
[12:28] Ubit Umarov: guess ill just commit what i have :p
[12:29] Ubit Umarov: log with several viewers, 3 versions.. yeack
[12:29] Ubit Umarov: 3 avatars online
[12:29] Ubit Umarov: better let you do it :p
[12:29] Bill Blight: same here but from the other side, testing 3 diff versions of the viewer
[12:30] Ubit Umarov: and tell me what smokes :p
[12:30] Ubit Umarov: forgot to tell that tags with viewers versions will be gone
[12:31] Ubit Umarov: as side effect of changes
[12:31] Ubit Umarov: they depend on non standard usage of data
[12:31] Ubit Umarov: data slots
[12:31] Ubit Umarov: they are also a privacy issue
[12:31] Ubit Umarov: so gone
[12:32] Ubit Umarov: ( sl did forbid them ages ago )
[12:32] Arielle Popstar: privacy issue?
[12:32] Bill Blight: yeah I'm gonna bail , need to rebuild a viewer
[12:32] Ubit Umarov: yes sl did not want ppl to see that fs is used a lot more than their own viewer :)
[12:33] Ubit Umarov: 80% share ( according to fs )
[12:33] Arielle Popstar: i see
[12:33] Arielle Popstar: not surprising
[12:33] Ubit Umarov: but there are more privacy issues
[12:34] Ubit Umarov: but well i did changed the code then noticed they where gone :)
[12:35] Ubit Umarov: a pain to support them again, + privacy issues == gone
[12:36] Arielle Popstar: ok
[12:36] Arielle Popstar: oh well
[12:36] Ubit Umarov: also telling others what software we are running is not a good idea
[12:37] Ubit Umarov: see i have this bugged thing that has a exploit over my head.. nahhh
[12:37] Ubit Umarov: it already happened to me in IRC ages ago... :)
[12:38] Ubit Umarov: a damm mirc version with a exploit on dcc
[12:39] Arielle Popstar: nod

Personal tools
About This Wiki