Chat log from the meeting on 2018-12-04

From OpenSimulator

Jump to: navigation, search

[11:01] Bernd.Steinhardt BTW: What causes if I see the Gridname in your name tags? Is this if you are member in the same group like me or the land group?
[11:03] Bill Blight: grid name is HG ... Sometimes if you are friends you don't see it, because the name is directly cached in your viewer
[11:03] Kayaker Magic: I'd like to ask today: Why are HG teleports so unreliable, and what is the right format for specifying one to osTeleportOwner or to FireStormm?
[11:04] Kayaker Magic: (I'm jumping into a question 'cause I have to leave early today)
[11:04] Bill Blight: @kayaker I never use http in the address and it seems to work fine from scripts
[11:05] Kayaker Magic: My regions here on OSGrid are through a dynamic DNS to a home ISP and router, so they are slow and that seems to make HG teleports fail.
[11:06] Kayaker Magic: Jumping to Wrong Plaza, then to my region seems to be the only way to get there.
[11:06] Bill Blight: dns cache
[11:06] Bill Blight: most likely
[11:06] Bill Blight: on other regions when it does the reverse
[11:07] Bill Blight: dyn dns with a short TTL is not liked by a lot of servers
[11:08] Bill Blight: some servers will keep the old cached address until they get restarted and this can cause IP/name mismatch
[11:08] Kayaker Magic: Of course, I cannot write a script to jump to LSBA first and then to my regions, because you can never count on the OSSL functions to work.
[11:12] Ubit Umarov: fear u did broke the merge of master here
[11:13] Ubit Umarov: well here and main osg plazas we are using a new cap to get assets like bodyparts, sounds animations, textures and mesh
[11:13] Ubit Umarov: well ppl with viewers that do use it
[11:13] Ubit Umarov: like fs
[11:14] Andrew.Hellershanks Time for a reminder about the OSCC this weekend. Just 4 days away.
[11:15] Marcus Llewellyn: Yay!
[11:15] Ubit Umarov: ahh you know Bills merges.. i was fearing broken local IMs here :p
[11:15] Bill Blight: yeppers
[11:15] Andrew.Hellershanks The schedule for this years event has been posted.
[11:15] Bernd.Steinhardt Is there anywhere an agenda of the conference?
[11:15] Gavin.Hird link to the agenda?
[11:15] Marcus Llewellyn:
[11:16] Andrew.Hellershanks Marcus beat me to it.
[11:16] Bernd.Steinhardt Perfect, thanks
[11:16] Marcus Llewellyn: I have it open while I'm created "Next up!" video rolls. ;)
[11:16] Andrew.Hellershanks If you miss a panel they are recorded for viewing later, IIRC.
[11:17] Kayaker Magic: Another Question: I have finally got 3 copies of Rubust running, but I recall from previous conversations here that best practice is to only run assets on one of them. Is this still true?
[11:17] Marcus Llewellyn: Yes. All sessions (excpet the OpenSim Dev VIP session) will be streamed live and recorded.
[11:17] Ubit Umarov: guess ill have a few to see by next week :)
[11:17] Ubit Umarov: oh that one not even for offline view ?
[11:17] Andrew.Hellershanks I will be around on Saturday for some of the events. I have a family birthday on Sunday so I may miss most of that days events.
[11:18] Marcus Llewellyn: It's intended to be a relaxed, off the record event for devs and people who funded the event.
[11:18] Kayaker Magic: Yeah, I'm double booked Saturday and will miss half the sessions of OSCC on that day.
[11:18] Ubit Umarov: ill miss both days i fear
[11:18] Marcus Llewellyn: This way, you core devs don't have to sign anything or feel pressured. You just show up. :)
[11:19] Andrew.Hellershanks :)
[11:19] Ubit Umarov: more pressure than some of this meetings?
[11:19] Andrew.Hellershanks I need to check with Melanie if she would like me on the Saturday morning panel with her.
[11:19] Marcus Llewellyn: It will not be a free for all. hehe
[11:19] Marcus Llewellyn: It's at 11:45 slt, so if you can make it, that'd be great. :)
[11:20] Marcus Llewellyn: There is no morning panel this year. No devs were willing to say they were available.
[11:21] Andrew.Hellershanks Marcus, sorry I didn't get back to you about doing a talk. Been a busy week or three for me. I didn't think I had the time to prepare something and I would have had to check with several others to make sure I didn't miss any change of note.
[11:22] Ubit Umarov: that about 0.9 version.. hmm im curius to see those news
[11:22] Marcus Llewellyn: We've rolled with it. :) Maria Forolov will be doing her stats, and Joyce will be highlighting 0.9 and the current work going in OpenSim.
[11:22] Ubit Umarov: but since 0.8.2 is so good, even the one used by oscc.. who wants 0.9 ?
[11:22] Ubit Umarov: :p
[11:22] Kayaker Magic: Gahhh!
[11:23] Marcus Llewellyn: Trust me, I am NOT happy that the OSCC grid is still on 0.8
[11:23] Marcus Llewellyn: No one is.
[11:23] Andrew.Hellershanks I hope that the people running the grid used for the OSCC will get around to updating it before next year.
[11:23] Ubit Umarov: well never mind :)
[11:23] Kayaker Magic: Where is Neb?
[11:23] Andrew.Hellershanks Busy with work or travel perhaps. I haven't seen him on IRC in some time.
[11:23] Marcus Llewellyn: I honestly don't know. Crista finally surfaced this week, but that's way to late to do anything about the grid.
[11:24] Arielle.Popstar if Diva had a mind to anyway
[11:25] Arielle.Popstar she mentioned at the conference that she considered Opensim stalled until a Opensim specific viewer
[11:26] Andrew.Hellershanks Diva may have a long time to wait for an OS specific viewer.
[11:26] Gavin.Hird can't she just write one?
[11:26] Marcus Llewellyn: Heh
[11:26] Arielle.Popstar in which case we may wait even longer before any more code submissions?
[11:27] Arielle.Popstar she said it should be from a professional developer
[11:27] Gavin.Hird Speaking of viewer, at the latest Sl TVP developer meeting Jessica said the FS team had to do something drastic about opensim support
[11:27] Andrew.Hellershanks The real indication of whether development has "stalled" is to check the git log.
[11:27] Gavin.Hird it sounded more like dropping it than anything else
[11:28] Kayaker Magic: Last question for me today: Ubit: can the http server in OpenSim handle multiple requests?
[11:28] Arielle.Popstar well since diva is the one for Hypergrid it may mean increasing breakage in that area
[11:29] Ubit Umarov: can't say i understand the question.. multiple requests is a basic req for it since ever
[11:29] Andrew.Hellershanks I was thinking the same about that question, Ubit.
[11:29] Dev Random: Kayaker, do you mean "Is it thread-safe?"?
[11:32] Kayaker Magic: I mean it seems to do single thread requests, will it ever handle multiple requests simultaneously?
[11:33] Ubit Umarov: kay it is multi thread since ever
[11:33] Ubit Umarov: well multi request
[11:34] Ubit Umarov: threading is just a implementation detail
[11:34] Bill Blight: should not need a new thread unless it is really over worked, spawning new threads when not needed is a waste of resources and just creates overhead
[11:35] Ubit Umarov: for now it actually does multithreading.. but just details
[11:35] Kayaker Magic: Then why do things work so much better with multiple Robusts running?
[11:36] Ubit Umarov: some are handle by the receiver thread, others are placed in pools for a limited number or workers to handle
[11:36] Ubit Umarov: there is a think called load
[11:36] Ubit Umarov: more machines, more true cpus
[11:37] Ubit Umarov: specially if also multiple mysql servers
[11:37] Kayaker Magic: Multiple robusts on one machine with one mysql server shows a significant improvement.
[11:37] Ubit Umarov: all on same machine.. don't think they are that better
[11:37] Bill Blight: I think it comes down to system handles more than anything else, more processes, lets you spread those over multiple processes
[11:38] Bill Blight: also now you are spreading over more event queues
[11:43] Andrew.Hellershanks I have been rather busy this past week so I haven't had time to update the wiki page of OSSL functions with the latest additions. This coming week should be less busy and I'll get the wiki updated.
[11:43] Dev Random: I couldn't make it here last week... was there discussion on exactly what constitutes a change in API version?
[11:43] Ubit Umarov: no
[11:44] Andsim Hammerer: oh what is change in it
[11:44] Ubit Umarov: i did add a constant for api version, but that has a low level prespective
[11:44] Andsim Hammerer: k
[11:45] Ubit Umarov: may use it to auto force scripts compile, or not :)
[11:45] Andrew.Hellershanks Andsim, I've been updating the list to make sure all supported OSSL functions were listed, marking which ones were new, noting when OSSL functions had been added.
[11:45] Andrew.Hellershanks Ubit recently added a number of new string handling functions.
[11:45] Bill Blight: I'll try to get back at that too Andrew, just been busy getting ready for OSCC
[11:45] Andsim Hammerer: ah
[11:45] Dev Random: Ubit, didn't you also mention something like "if(API < 11.0) quit;" in user scripts
[11:45] Ubit Umarov: but any addition or major chage to current fucntions is a API change
[11:46] Ubit Umarov: well yes, but if the api is diferent it will not run anyways
[11:46] Ubit Umarov: so will not reach that if :)
[11:46] Andrew.Hellershanks Due to the number of string functions that were added I was trying to decide if they belong in their own category on the page along with a few other existing string manipulation functions.
[11:46] Dev Random: lol.. good answer, Ubit
[11:46] Ubit Umarov: unless the change was on how a funtion works
[11:47] Ubit Umarov: ie does "link" fine but gives dif results
[11:47] Andrew.Hellershanks np, Bill.
[11:48] Ubit Umarov: well i got distracted from script part :)
[11:48] Andrew.Hellershanks I'll offer up that question here. Should string functions be in a separate category on the list of OSSL functions or just lumped in to the Miscellaneous category?
[11:48] Ubit Umarov: like by that problem with inventory and fs
[11:48] Bill Blight: My opinion they should Andrew
[11:49] Bill Blight: String Manipulation is kind of taking on a life of it's own
[11:49] Andrew.Hellershanks Bill, I'm leaning in that direction also.
[11:50] Andrew.Hellershanks IIRC, about a dozen string functions were added along with three that we had already.
[11:50] Marcus Llewellyn: Sounds sensible to me.
[11:50] Ubit Umarov: ohh where not that many
[11:50] Andrew.Hellershanks Ubit, not too far off that. I forget the exact count.
[11:50] Dev Random: 10+ seems like enough to justify a category
[11:52] Selby.Evans Good to put string functions in their own category.
[11:52] Andrew Hellershanks: Back to the API version issue, if the API version was bumped every time new functions were added it could get bumped multiple times during a dev cycle.Updating the API version
[11:52] Bill Blight: I'm thinking the api version might be for when the underlying function format is changed or the result of a function changes
[11:53] Dev Random: the big problem with such things is that an API version may not actually change the function I want to call.. It may change may others.. so mine is still effectively at $version - 1
[11:53] Ubit Umarov: API is and will go on being a issue
[11:54] Ubit Umarov: almost ever grid or stanalone is on its own api
[11:54] Ubit Umarov: and also some modules can extend it
[11:54] Ubit Umarov: so.. making a script that works every where.. is.. lets say hard
[11:54] Dev Random: ha
[11:54] Andrew Hellershanks: Bumping the API version when a new release is made available, whether that is a full version release or an rc, could be done if there were additions or changes to the API.
[11:55] Ubit Umarov: ( and on top of that confusiong diferent ossl threat levels )
[11:56] Dev Random: but if changes are not always backward compatible. then my check for (API > 11) might pass, but the script will break anyway.
[11:56] Andrew Hellershanks: If you can test the version of API available it might alert a person that the API has changed but won't tell them what changed.
[11:57] Andrew Hellershanks: If the behaviour of a previously existing function has changed that would be a reason to bump the API version number.
[11:57] Ubit Umarov: about that viewers talk
[11:58] Ubit Umarov: if i remember last yr oscc i also said opensim was kinda stalled bc of that
[11:58] Ubit Umarov: BUT
[11:58] Ubit Umarov: i was talking about new features
[11:58] Ubit Umarov: and how things look and work
[11:58] Ubit Umarov: bc all of that is defined by viewers
[11:59] Ubit Umarov: i was not talking is is done.. far from it
[11:59] Andrew Hellershanks: Ubit, Be careful what you say. People like sound bites and it is too easy to just take the first part of that statement. :)
[11:59] Ubit Umarov: it does things some does them very badly, very slow..
[12:00] Ubit Umarov: so lots of things need work, a lot of it
[12:00] Gavin.Hird make a list
[12:00] Ubit Umarov: viewers have a another problem
[12:00] Ubit Umarov: they change.. most time just with c&p from ll
[12:01] Andrew Hellershanks: I was just about to say that some people have been asking about a roadmap to see a list of what needs to be done and what is (or will) be worked on.
[12:01] Ubit Umarov: with things that we do not want, at least in that time frame
[12:01] Ubit Umarov: so they keep getting broken for us
[12:01] Ubit Umarov: like that fs inventory issue
[12:01] Gavin.Hird as long as there is no roadmap viewers will take functionality from SL
[12:01] Ubit Umarov: well we are not SL
[12:01] Gavin.Hird I have the same code in my viewer as FS have
[12:02] Andsim Hammerer: when i tested my version os oon singularity is broken on my grid
[12:02] Gavin.Hird and threading the caps to the simulator fixed the folder copy
[12:02] Andsim Hammerer: for inventory
[12:02] Gavin.Hird Singularity is dead code
[12:02] Dev Random: What became of Alchemy viewer?
[12:02] Ubit Umarov: and even if consider to add a feature identical to sl, will be done on a diferent time frame
[12:02] Gavin.Hird it has not recevied a sensible update for years
[12:02] Ubit Umarov: yes but we will not add caps just because SL did
[12:02] Gavin.Hird it does not even run on my system andy more, let alone compile
[12:03] Ubit Umarov: that cap is basiclly worse then older udp
[12:03] Andrew Hellershanks: I haven't used Singularity that often but I don't recall seeing problems using it in an OS grid.
[12:03] Ubit Umarov: except on one detail
[12:03] Ubit Umarov: and i mean for us
[12:03] Gavin.Hird the cap has been in the viewer code forever
[12:03] Dev Random: Gavin, Sianna Gearz started a Patreon "should I fix Singularity" to judge interest.....
[12:03] Gavin.Hird they just fixed a bug around it
[12:03] Ubit Umarov: yes but fs lost the udp code
[12:03] Ubit Umarov: so our code path got broken
[12:04] Bill Blight: FS was not even falling back to UDP, it was http or bust
[12:04] Andsim Hammerer: yea that why i issues a ban on my grid soo cant be use any more
[12:04] Selby.Evans My proposal for educational viewer:
[12:04] Ubit Umarov: that is our main issue
[12:04] Ubit Umarov: singu 1.8.7 seems to have a few issues
[12:04] Gavin.Hird In any case there is a new non and RLV build of the viewer still hot from the compiler if anyone wants to test it
[12:04] Ubit Umarov: all viewers do
[12:04] Ubit Umarov: in fact fs does seem the one working better for now
[12:05] Ubit Umarov: even hating us, and doing very silly things at low level
[12:05] Bill Blight: When FS goes full HTTP, there is hope that .9x will work in the future but .8x will be dead to it ..
[12:05] Ubit Umarov: some moves to http do mk sense, others don't
[12:05] Gavin.Hird FS will probably drop opensim support for the next release
[12:05] Andrew Hellershanks: The topic of a good (OS specific) viewer comes up now and then. We will keep talking about it until someone steps up and puts out such a viewer.
[12:06] Ubit Umarov: well fs is telling that for 3yrs at least
[12:06] Gavin.Hird there is so much new code from LL they cannot integrate all in one viewer any more
[12:06] Andsim Hammerer: i had a viewer i used taken from fs
[12:06] Selby.Evans development status:
[12:07] Bill Blight: stock singu
[12:07] Bill Blight: oh it has a new readme
[12:07] Arielle Popstar: noone ever says what an opensim specific viewer should look like. What should it be able to do differently from the FS one
[12:07] Gavin.Hird Commits since last meeting:
[12:07] Andsim Hammerer: that version of singulartiy is broken in my grid since it need http
[12:07] Andrew Hellershanks: Selby, that URL points to a Windows only version.
[12:08] Ubit Umarov: does not matter what a opensim viewer does as long it does it in sync with opensim
[12:08] Andsim Hammerer: right
[12:08] Ubit Umarov: SL viewers are not independent from the rest
[12:08] Selby.Evans right, andrew -- Drvelopment status for now
[12:08] Ubit Umarov: they are just part of all
[12:08] Arielle Popstar: it could at the very least be stated what would be sync currently
[12:09] Andrew Hellershanks: Selby, good luck with it. I won't be able to help with testing.
[12:09] Ubit Umarov: back in the day opensim could just ride the wave following sl viewer changes
[12:09] Gavin.Hird There is one litte detail you forget Ubit: Viewers also need to run on recent clients
[12:09] Gavin.Hird to keep them running takes a heck of a lot of code changes
[12:09] Ubit Umarov: well things got complex.. goals changed.. not that possible anymore
[12:09] Gavin.Hird flüstert: and that is what we get from LL
[12:10] Gavin.Hird they also need to build on recent development tools
[12:10] Ubit Umarov: you new os and os versions?
[12:10] Gavin.Hird that is also code changes we get from LL
[12:10] Ubit Umarov: yeap
[12:10] Andrew Hellershanks: Let's hope that there is some good news on the viewer development front made at this years OSCC.
[12:10] Ubit Umarov: we are also having that on opensim
[12:10] Selby.Evans I understand andrew -- Mac version inwork --
[12:10] Ubit Umarov: seen no viewer dev or talk on shedule
[12:11] Andrew Hellershanks: Selby, I run Linux. Mac and Windows version don't help me. :)
[12:11] Andsim Hammerer flüstert: oh
[12:11] Andsim Hammerer: oh*
[12:11] Andrew Hellershanks: Selby, Mac users will be glad to know a viewer version for them is coming soon.
[12:11] Ubit Umarov: we did achive some compatibilty with a particular generation of viewers
[12:12] Ubit Umarov: ( with a lot of effort also from viewer devs )
[12:12] Selby.Evans This version is for students to use in education -- may not have a linux version. -- I understand -- cuts out techs -- but techs would hate it.
[12:13] Ubit Umarov: and that happened also bc some grids did pay for viewer changes and made them public
[12:13] Ubit Umarov: most current grids that do that, its changes for internal use, and keep it private also
[12:14] Gavin.Hird Selby where is the souce and commits in the repo you linked?
[12:15] Selby.Evans Sorry, Gavin --that is all I have now -- more will come in OSCC
[12:15] Ubit Umarov: but also a lot was also "opensim devs" "fixing things for viewers and viewer devs fixing things for opensim, all ending up on a mess
[12:16] Gavin.Hird ok...
[12:16] Ubit Umarov: viewers var regions suport is still defective for example
[12:16] Ubit Umarov: something damm old now ( since aurora )
[12:16] Andsim Hammerer: never had issue with var
[12:17] Gavin.Hird flüstert: why don't you make a list of what is not working and how to test it?
[12:17] Andsim Hammerer: yea
[12:17] Andrew Hellershanks: Andsim, try putting vars of different sizes up against each other.
[12:17] Andsim Hammerer: oh
[12:17] Andsim Hammerer: that
[12:17] Andsim Hammerer: hmm ok
[12:17] Bill Blight: put multiple regions on the same side touching a var
[12:17] Bill Blight: and wait for the boom
[12:17] Ubit Umarov: yeap that is a example
[12:17] Arielle Popstar: Robert mentioned not to do that quite a while ago
[12:18] Ubit Umarov: minimap is a gamblig bet
[12:18] Andsim Hammerer: i have test on my computer
[12:18] Andrew Hellershanks: I've seen a 2x2 or 3x3 var with one regular sized regions next to it on three sides.
[12:18] Bill Blight: Single region per side, regardless of size
[12:18] Arielle Popstar: and like mega's, dont put the northeast corner abutting the southwest corner of another Var
[12:18] Bill Blight: this has been confirmed at least in most part to be a viewer side issue, by another viewer dev
[12:18] Bill Blight: Megas are dead
[12:19] Ubit Umarov: local tps on some viewers.. a gamblig bet :)
[12:19] Ubit Umarov: well several issues
[12:19] Andsim Hammerer: os have http cap yet?
[12:19] Ubit Umarov: i do understand well viewer devs problems.. but yeap
[12:20] Andrew Hellershanks: Those issues could become the start of a roadmap, or at least a list of known issues to be worked on.
[12:20] Ubit Umarov: yeackkk
[12:20] Ubit Umarov: And.. caps are http and we have tons since ever
[12:20] Andsim Hammerer: hmm
[12:21] Andsim Hammerer: k
[12:21] Ubit Umarov: mesh never worked on lludp
[12:22] Andrew Hellershanks: For those of you attending the OSCC, I hope you find it useful and informative.
[12:22] Andrew Hellershanks: I might see some of you there on Saturday.
[12:23] Arielle Popstar: hopes it isnt going to be depressing but after diva's talk, i think it will be
[12:23] Bill Blight: It is easy for people to sit in ivory towers and preach what is wrong, when they are not actually in the fight, personally I'll stand by the people in the trenches .
[12:23] Andrew Hellershanks: For now, thank you all for coming. See you next week. We can talk about things we heard or learned about from panels at the OSCC.
[12:23] Ubit Umarov: diva said nothing new

Personal tools
About This Wiki