[11:03 AM PT] Andrew Hellershanks: Hello, everyone.
[11:03 AM PT] Gavin.Hird @grid.xmir.org:8002: Hi Andrew
[11:05 AM PT] Andrew Hellershanks: I see there was just one commit this past week.
[11:05 AM PT] Andrew Hellershanks: two
[11:06 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Huge commits this week, truly huge, the best. Fixed a potential null ref, it's gone, won't come back. Also a funny typo, typo in code, can you imagine, fixed that. No more typos, all gone.
[11:07 AM PT] Gavin.Hird @grid.xmir.org:8002: typo in comment, yeah?
[11:07 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Yeah, truly a change with much gravity
[11:07 AM PT] Ubit Umarov: yeah important
[11:07 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Documentation is important after all
[11:07 AM PT] Ubit Umarov: "iff" was very bad
[11:07 AM PT] Gavin.Hird @grid.xmir.org:8002: 100%
[11:08 AM PT] Vincent.Sylvester @hg.zetaworlds.com: It's what sets us apart from the javascript monkeys
[11:08 AM PT] Andrew Hellershanks: I have three typo changes I fixed in my copy of OS but still haven't pushed several years later. I was planning to look for additional ones in the ini files first.
[11:08 AM PT] Gavin.Hird @grid.xmir.org:8002: iff = international football federation
[11:08 AM PT] Andrew Hellershanks: hehe
[11:08 AM PT] Gavin.Hird @grid.xmir.org:8002: and you can't have that in coide
[11:09 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I did remove some whitespace also, cause earlier today visual studio decided to just troll me and crash. Error was something to do with whitespace so I figured, might as well do some cleaning, not sure if Ubit is gonna bother merging that though, mostly useless changes
[11:09 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Has been really stable thus far, only second crash this year, knock on wood
[11:10 AM PT] Gavin.Hird @grid.xmir.org:8002: does it crash over whitespace?
[11:10 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Apparently something with syntax highlighter lookahead on lines containing just spaces
[11:11 AM PT] Gavin.Hird @grid.xmir.org:8002: one would think they had things like that figured out by now
[11:11 AM PT] Andrew Hellershanks: Vincent, I have occasionally thought about going through the source code files and cleaning up whitespace. My vim configuration handles that automatically for me.
[11:12 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I did also remove some brackets and shortened some code, cosmetic stuff. Only thing I found was some ossl thing that runs code before threat level check
[11:12 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Whenever I make changes I do look around a bit what might be worth changing
[11:13 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Would hope that most of it gets optimized away anyways, but never really know
[11:13 AM PT] Ubit Umarov: guess his cpu did over degradate...
[11:14 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I have a 10th gen unlike ... ;P
[11:14 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Well testing, hot reloading etc does put some stress on it I guess
[11:14 AM PT] Ubit Umarov: (instaled last microcode and lost like 10% performance )
[11:15 AM PT] Gavin.Hird @grid.xmir.org:8002: ouch
[11:15 AM PT] Ubit Umarov: ( no idea.. some clam no loss.. others 20%.. )
[11:15 AM PT] Ubit Umarov: intel doing a great job...
[11:16 AM PT] Ubit Umarov: guess this i9 is was good as a halt the cost cpu
[11:16 AM PT] Ubit Umarov: ...has...
[11:17 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Good luck with the rma, we'll miss you. Swear intel is screwing this up big time
[11:17 AM PT] Gavin.Hird @grid.xmir.org:8002: Intel and Boeing seems to have hit a few snags lately
[11:17 AM PT] Vincent.Sylvester @hg.zetaworlds.com: a few... yeah
[11:17 AM PT] Ubit Umarov: don't believe intel will do any rma on this
[11:18 AM PT] Vincent.Sylvester @hg.zetaworlds.com: They are supposed to by their own admission, but that's changing daily now too
[11:18 AM PT] Ubit Umarov: like they did replace broken pentiums with the fpu bug back in 90's
[11:18 AM PT] Ubit Umarov: or was still 80's?
[11:19 AM PT] Ubit Umarov: whatever :)
[11:19 AM PT] Ubit Umarov: so how are your viewers, joe?
[11:19 AM PT] Ubit Umarov: gavin?
[11:20 AM PT] Gavin.Hird @grid.xmir.org:8002: I published another preview build and people in SL download both the Mac and Win versions
[11:20 AM PT] Ubit Umarov: :)
[11:21 AM PT] Andrew Hellershanks: :)
[11:21 AM PT] Gavin.Hird @grid.xmir.org:8002: depsite not being anywhere near having PBS or even AIS3 support
[11:21 AM PT] Gavin.Hird @grid.xmir.org:8002: OS Mac viewer also got a small refresh
[11:21 AM PT] Ubit Umarov: their services still work without ais ?
[11:21 AM PT] Gavin.Hird @grid.xmir.org:8002: sure
[11:21 AM PT] Gavin.Hird @grid.xmir.org:8002: ais3 that is
[11:23 AM PT] Ubit Umarov: wlel i did look aroung some things for webrtc...
[11:24 AM PT] Ubit Umarov: big fail :P
[11:24 AM PT] Lyr Lobo slips in and grins
[11:24 AM PT] Ubit Umarov: the things ms is using do not compile on win, even on cygwin...
[11:25 AM PT] Ubit Umarov: tested a chinese media controler that kinda did compile..
[11:25 AM PT] Ubit Umarov: and did a simple file stream...
[11:25 AM PT] Andrew Hellershanks: bbiab
[11:25 AM PT] Ubit Umarov: total fail... a lot worse than a test i did in 90's streaming a webcam :)
[11:26 AM PT] Ubit Umarov: pixelization... etc etc
[11:26 AM PT] Ubit Umarov: well this the things writen on a proper computer language..
[11:26 AM PT] Ubit Umarov: rest are java.. blabla thing
[11:26 AM PT] Ubit Umarov: things..
[11:26 AM PT] Vincent.Sylvester @hg.zetaworlds.com: From what I hear through the grapevine the setup might also evolve yet still as apparently some have complained the actual spacial part isn't all that great. Seems it only has doppler shift and can't properly place audio above or below
[11:26 AM PT] Ubit Umarov: and linux domain..
[11:27 AM PT] Ubit Umarov: well nothing has 3d as sl.. that is extra code
[11:27 AM PT] Ubit Umarov: ( like it was for them )
[11:27 AM PT] Ubit Umarov: they use janus as mcu for that..
[11:28 AM PT] Ubit Umarov: that does not compile.. parts of it use low level linux kernel API that cygwin can't do
[11:28 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Which is not hard to setup, but they also have custom data channel, which also has to work to connect and I haven't found how that works yet
[11:28 AM PT] Ubit Umarov: same for coturn
[11:29 AM PT] Ubit Umarov: we do need a win solution for all the stanaalons
[11:29 AM PT] Ubit Umarov: also several grids on windows servers
[11:29 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Unfortunately so
[11:29 AM PT] Ubit Umarov: so a linux only solution is not a good solution
[11:30 AM PT] Gavin.Hird @grid.xmir.org:8002: yeah
[11:30 AM PT] Gavin.Hird @grid.xmir.org:8002: I suppose it cannot be run in a Windows linux subsystem
[11:30 AM PT] Ubit Umarov: insane
[11:31 AM PT] Ubit Umarov: oh fun... seems docker also does not work on win
[11:31 AM PT] Gavin.Hird @grid.xmir.org:8002: convoluated if it worked
[11:31 AM PT] Gavin.Hird @grid.xmir.org:8002: convoluted*
[11:31 AM PT] Ubit Umarov: ( the chinese thing wanted docker.. but does compile on cyg )
[11:31 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Well the thing is. Ideally you wouldn't put this into OpenSim and run it there. You'd run it externally and just connect the parts to each other. To not only locate it better, but also because running a turn server behind nat is super annoying and likely goes over a lot of people's head to setup free port ranges etc.
[11:32 AM PT] Ubit Umarov: linux on win .. imagine all the ppl on standalones having to install that
[11:32 AM PT] Ubit Umarov: and that is also not that good
[11:32 AM PT] Ubit Umarov: the version 2 internet is crap
[11:33 AM PT] Gavin.Hird @grid.xmir.org:8002: someone told me the day it is just a click to get it set up :-)
[11:33 AM PT] Ubit Umarov: yeah ppl say things don't they? :P
[11:33 AM PT] Gavin.Hird @grid.xmir.org:8002: the other day*
[11:33 AM PT] Gavin.Hird @grid.xmir.org:8002: indeed
[11:34 AM PT] Ubit Umarov: and after install it is a linux... know any linux that works out of the box ( or ever...)
[11:34 AM PT] Ubit Umarov: :)
[11:34 AM PT] Ubit Umarov: nah not a option
[11:34 AM PT] Ubit Umarov: simpler to have a linux solution and send all to osgrid :p
[11:35 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I'd say be simpler and probably also much better performance and latency just to get a decent vps somewhere close to where the users are and set things up on there
[11:36 AM PT] Vincent.Sylvester @hg.zetaworlds.com: At the end of the day having all that stuff on the same box... or behind nat... total pain
[11:36 AM PT] Ubit Umarov: well that is not the point of standalones..
[11:36 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Yeah, but it's not above requiring external stuff either, look at search or money
[11:37 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Can't provide everything inside of it, especially not something as heavy and complex as webrtc can be
[11:37 AM PT] Ubit Umarov: and we are talking just abotu have a platform to start with.. not even interface to opensim :)
[11:37 AM PT] Ubit Umarov: well many say that webrtc is crap
[11:37 AM PT] Ubit Umarov: ofc.. :)
[11:38 AM PT] Ubit Umarov: to start the force of secure connections
[11:38 AM PT] Ubit Umarov: that is a big waste of resources, that many just do not need...
[11:39 AM PT] Ubit Umarov: simple p2p could be ok.. but now ppl all scared abotu having ip public..
[11:39 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I'm just saying. You can't shoehorn everything into OpenSim or require it to be just as easy to setup, cause in some areas that's just not a thing and might never be. In case of webrtc there is already the part of the custom data plugin, the OpenSim module and the interconnect between plugin and OpenSim to consider. Trying to also find a solution that "just works" everywhere is picking the wrong battle in my opinion
[11:40 AM PT] Ubit Umarov: so all want at least a TURN server...
[11:40 AM PT] Vincent.Sylvester @hg.zetaworlds.com: As much as being platform independent is a key point, if then selecting an inferior solution based on it just brings the whole thing down
[11:41 AM PT] Ubit Umarov: but with a turn server on a multi user room... all one tells is copied and sent to all others
[11:41 AM PT] Andrew Hellershanks: I'm back.
[11:41 AM PT] Ubit Umarov: very nice for bandwidth..
[11:41 AM PT] Andrew Hellershanks: What is a turn server?
[11:42 AM PT] Ubit Umarov: a thing that gets what you send and sends to others, fixing NAT issues
[11:42 AM PT] Ubit Umarov: or going around NAT issues..
[11:44 AM PT] Andrew Hellershanks: um... ok
[11:44 AM PT] Andrew Hellershanks sighs
[11:44 AM PT] Andrew Hellershanks: My cat has decided she wants to play during the day time (for a change) and picked the wrong day/time.
[11:44 AM PT] Vincent.Sylvester @hg.zetaworlds.com: You really don't want to run that behind nat, because you have to open a large port range for that to work and that in.. turn.. heh.. exposes your network. Turn servers are chatty buggers. Someone probes your network for it, it can be quite bad as well
[11:45 AM PT] Ubit Umarov: well standalones are behind nat :)
[11:45 AM PT] Ubit Umarov: original basic webrtc was p2p
[11:46 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Like I said, pick battles and don't lose the war. Focus on the parts we can fix and leave the rest up to users to work out. I mean it's not like linux is like this ancient language no one understands either
[11:46 AM PT] Ubit Umarov: a extra server do help the negotiation... then direct UDP between both parties
[11:46 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I'd rather help 100 people to install their vps and turn server on it, than settle for a shitty one that happens to somehow poorly work on windows too
[11:46 AM PT] Gavin.Hird @grid.xmir.org:8002: what, you don't need a PDP/11 to run Linux?
[11:47 AM PT] Ubit Umarov: well a pdp 11 is like 900W or power :)
[11:47 AM PT] Ubit Umarov: more..
[11:47 AM PT] Gavin.Hird @grid.xmir.org:8002: nice space heater in winter
[11:47 AM PT] Gavin.Hird @grid.xmir.org:8002: noisy I bet
[11:47 AM PT] Ubit Umarov: yeah
[11:48 AM PT] Ubit Umarov: back then fans where real fans..
[11:48 AM PT] Gavin.Hird @grid.xmir.org:8002: lol
[11:48 AM PT] Gavin.Hird @grid.xmir.org:8002: the TTY was even louder
[11:48 AM PT] Ubit Umarov: :)
[11:50 AM PT] Gavin.Hird @grid.xmir.org:8002: I wonder if the TTY keyboard is cat proof - you need real point pressure to type a character
[11:50 AM PT] Andrew Hellershanks: Depends on the weight of the cat
[11:50 AM PT] Gavin.Hird @grid.xmir.org:8002: sure, if you have a tiger
[11:51 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I have some good news to report on the maptile front as well. After wrecking my tiny monkey brain over a bunch of math and bit shifts I finally got a working system to parallel the zoom level generation. It now creates the zoom levels with a lot of tiles in parallel and then combines the resulting images on the larger end, which is a lot faster than adding each tile to them. 2-6 as parallel "copy each tile to image" and 7-11 as "combine previous tiles to larger one". This results in less cpu usage for the larger levels and overall speeds up processing again by a factor of 2 or more
[11:51 AM PT] Andrew Hellershanks: Doesn't need to be that heavy
[11:52 AM PT] Vincent.Sylvester @hg.zetaworlds.com: [MAP IMAGE WORKER]: Finished generating 5120 zoom level tiles of 13046 maptiles. Took 52.4396577 seconds!
[11:52 AM PT] Gavin.Hird @grid.xmir.org:8002: sweet
[11:52 AM PT] Andrew Hellershanks: Vincent, Sounds good. What type of machine was generating the tiles?
[11:53 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Kept the old routines as well, so when it detects a new region has places tiles it simply adds them to the existing zoom levels, which should be quite light
[11:53 AM PT] Cuga.Rajal @rajal.org:9000: Nice!
[11:53 AM PT] Vincent.Sylvester @hg.zetaworlds.com: My computer. I recently got some ssd sata raid to remove the io bottleneck I had on the disks
[11:54 AM PT] Cuga.Rajal @rajal.org:9000: Still working on the new maptile format option?
[11:54 AM PT] Vincent.Sylvester @hg.zetaworlds.com: This is just for robust making the levels. I did mostly finish the updates to the TIFF loader as well, going full 32 bit floating point on that
[11:54 AM PT] Vincent.Sylvester @hg.zetaworlds.com: So it should be producing the same level of granular control as r32 now
[11:55 AM PT] Vincent.Sylvester @hg.zetaworlds.com: I left the 16 bit int routines in there, but can probably remove those now heh
[11:56 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Still have to test that some more and check that all data is saved and loaded 1 to 1 or as close to as possible
[11:56 AM PT] Vincent.Sylvester @hg.zetaworlds.com: PNG was on my list too, but it's a major pain due to the different compressions it can have and there is no option to have it uncompressed either, so I shelved that
[11:57 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Since TIFF is technically just writing bytes now I even thought maybe just making a custom .terrain format or something to store values without any conversion, but 32 bit floats should be enough anyways so no real need for that
[11:58 AM PT] Vincent.Sylvester @hg.zetaworlds.com: So far it seems those files load in GIMP and you can even view them directly, but programs seem to sometimes convert them down to less precision as they tend to look differently
[11:58 AM PT] Vincent.Sylvester @hg.zetaworlds.com: On windows the photo viewer seems to "flatten" the image to ints
[11:59 AM PT] Vincent.Sylvester @hg.zetaworlds.com: Noticeable difference in the grayscale shade
[11:59 AM PT] Ubit Umarov: oops
[11:59 AM PT] Joe Magarac: Gotta go, next meeting.
[11:59 AM PT] Ubit Umarov: have fun
[11:59 AM PT] Ubit Umarov: our pet decided to came closer
[12:00 PM PT] Andrew Hellershanks: ok, Joe. Thanks for dropping by
[12:00 PM PT] Gavin.Hird @grid.xmir.org:8002: I wondered what the sound was
[12:00 PM PT] Gavin.Hird @grid.xmir.org:8002: sounded like distant shots at first
[12:00 PM PT] Ubit Umarov: yeah he did sniff your head even
[12:00 PM PT] Gavin.Hird @grid.xmir.org:8002: he did :-)
[12:00 PM PT] Andrew Hellershanks: I haven't noticed any sounds out of the ordinary. In fact, I usually don't hear any noises here.
[12:03 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I did also uncouple the maptile renderer region side from the process so that it would run without blocking the console
[12:03 PM PT] Vincent.Sylvester @hg.zetaworlds.com: That irked me a lot, especially warp3d taking 20 minutes and you can't do anything in that time
[12:04 PM PT] Andrew Hellershanks: Vincent, that's good. It used to take a while to generate the nice looking tiles and case a delay in region startup.
[12:04 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I still have to come up with a way to "reserve" the tiles so it doesn't look empty despite a region being there
[12:04 PM PT] Vincent.Sylvester @hg.zetaworlds.com: At the moment just sends water and then does maptile later effectively re-registering to grid, which is how that has always worked anyways
[12:05 PM PT] Vincent.Sylvester @hg.zetaworlds.com: It also means you can switch between renderers on the fly. Don't like the mapimage, switch to warp3d, no restart required
[12:05 PM PT] Andrew Hellershanks: Vincent, Does the map tile generator immediately ignore any objects that are under water?
[12:06 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I don't think they do. Mapimage just puts boxes for prims wherever they are. warp3d might have internal cutoffs, but that code is too abstract for my brain
[12:06 PM PT] Vincent.Sylvester @hg.zetaworlds.com: There are upper limits for them
[12:06 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Might have lower limits too
[12:07 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I didn't touch the generators themselves, just stuffed them into their own threads
[12:07 PM PT] Andrew Hellershanks: Won't make much difference for most regions but would for a water region where everything is below sea level. Searching for objects with Z >= sea level would be a simple way to save a little bit of time.
[12:09 PM PT] Vincent.Sylvester @hg.zetaworlds.com: I have some regions with terrain height of 0 and corals on the ground. In mapimage generator that stuff still shows up. Have not tried warp3d. It would mean getting the water height or something too, extra step
[12:09 PM PT] Andrew Hellershanks: hm... that might not work either if some object had its center just below sea level but stuck up out of the water.
[12:10 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Let alone you'd maybe want to see some of it on the map. I think with warp3d the terrain is rendered and the water above as well, cause you can kinda see the terrain when the water is pretty shallow
[12:11 PM PT] Vincent.Sylvester @hg.zetaworlds.com: warp3d is mostly slow due to large complex objects. You can help it out with making simpler meshes or reducing textures etc. It valiantly tries to render what it can, cause it probably isn't aware of complexity before it starts parsing an object
[12:13 PM PT] Vincent.Sylvester @hg.zetaworlds.com: In that whole revamp I also included removal of the terrain_image assets that would get uploaded. Instead saving those on scene directly, so when viewer asks for them, the region delivers directly instead of being the "proxy" for asset server it normally is
[12:13 PM PT] Ubit Umarov: yes it does render both
[12:13 PM PT] Ubit Umarov: (optionally )
[12:14 PM PT] Vincent.Sylvester @hg.zetaworlds.com: That means less spam to assets on that, though it does mean retrieving a maptile of another region may not work... easier to ask robust directly anyways in my opinion, but yeah
[12:15 PM PT] Vincent.Sylvester @hg.zetaworlds.com: Sort of outweighs the negative of that to not have the asset server full of crap
[12:15 PM PT] Gavin.Hird @grid.xmir.org:8002: OK guys, I think I need to go.
[12:15 PM PT] Gavin.Hird @grid.xmir.org:8002: have a good one
[12:15 PM PT] Vincent.Sylvester @hg.zetaworlds.com: See ya :)
[12:15 PM PT] Andrew Hellershanks: ok, Gavin. See you next week.
[12:15 PM PT] Ubit Umarov: cya
[12:15 PM PT] Andrew Hellershanks: Any last minute items for today?
[12:16 PM PT] Andrew Hellershanks: Alrighty then. That will do it for this week.
[12:16 PM PT] Andrew Hellershanks: Thank you all for coming. See you again next week.
