Chat log from the meeting on 2018-10-16

From OpenSimulator

(Difference between revisions)
Jump to: navigation, search
(Created page with "[11:14] Andrew.Hellershanks @hg.osgrid.org: Time to start as it seems like everyone who will be here is already here. <br>[11:14] Andrew.Hellershanks @hg.osgrid.org: I took on...")
 
 
Line 329: Line 329:
 
<br>[12:31] Ubit Umarov: to do such a change i should have a request from viewers with full spec of the new protocol messages
 
<br>[12:31] Ubit Umarov: to do such a change i should have a request from viewers with full spec of the new protocol messages
 
<br>[12:31] Bill Blight: (4 lines in the viewer I should say)
 
<br>[12:31] Bill Blight: (4 lines in the viewer I should say)
 +
 +
[[Category:Office Hour Logs]]

Latest revision as of 11:52, 23 October 2018

[11:14] Andrew.Hellershanks @hg.osgrid.org: Time to start as it seems like everyone who will be here is already here.
[11:14] Andrew.Hellershanks @hg.osgrid.org: I took on what I thought was going to be a big task this week but it turned out not to be as bad as I thought.
[11:15] Ubit Umarov: damm time flyes... last meeting seems was almost yesterday
[11:15] Andrew.Hellershanks @hg.osgrid.org: I updated the OSSL web pages and added a note to say when each function was added starting with any additions between 0.6.9 and 0.7
[11:16] Andrew.Hellershanks @hg.osgrid.org: The OSSL pages are up to date as of the 0.9 version.
[11:16] Gavin.Hird @grid.xmir.org:8002: nice
[11:16] Ubit Umarov: hmm ill fix that soon
[11:16] Leighton.Marjoram @grid.outandproud.life:8002: do you have the link Andrew
[11:16] Ubit Umarov: a few on mantis waiting :p
[11:17] Andrew.Hellershanks @hg.osgrid.org: I thought it was useful information to have in the wiki pages if only for historical information.
[11:17] Ubit Umarov: ( opensim with updated docs.. can't be.. )
[11:17] Andrew.Hellershanks @hg.osgrid.org: I found two deprecated functions that were listed on the main functions page and also found some that weren't taged "new".
[11:17] Ubit Umarov: ( well considering some of the docs are possibly wrong.. normal )
[11:18] Andrew.Hellershanks @hg.osgrid.org: The OSSL list of functions is at http://opensimulator.org/wiki/OSSL_Implemented
[11:19] Leighton.Marjoram @grid.outandproud.life:8002: ty
[11:19] Andrew.Hellershanks @hg.osgrid.org: For example, look at http://opensimulator.org/wiki/OsForceAttachToOtherAvatarFromInventory
[11:19] Andrew.Hellershanks @hg.osgrid.org: At the bottom you will see a note saying it was added in 0.7.4-post-fixes.
[11:19] Ubit Umarov: why do we have such a thing? uff
[11:20] Arielle.Popstar @hg.osgrid.org: what thing?
[11:20] Ubit Umarov: :)
[11:20] Gavin.Hird @grid.xmir.org:8002: Proably to satisfy the RLV crowd
[11:20] Ubit Umarov: forceDaAttach i could understand ;)
[11:20] Andrew.Hellershanks @hg.osgrid.org: Ubit, since you mentioned the mantis and something waiting, there was a patch where you said it is on hold. Not sure what it was waiting for.
[11:21] Sheera Khan: might be interesting for NPCs
[11:21] Ubit Umarov: well yeah i can see a few uses
[11:21] Andrew.Hellershanks @hg.osgrid.org: 8093 has the patch that Ubit said is on hold.
[11:23] Ubit Umarov: well that one was in 2016 :p
[11:23] Andrew.Hellershanks @hg.osgrid.org: I'm wondering about all those Sound related OSSL functions. I think that my next job will be to create pages for all those functions. I currently know nothing about what exactly they do vs any other sound functions we have.
[11:23] Andrew.Hellershanks @hg.osgrid.org: Ubit, true.
[11:24] Bill Blight: Some of those sound functions require viewer support which was apparently removed ..
[11:24] Andrew.Hellershanks @hg.osgrid.org: I was surprised to see only 14 reports. Most are at least two years old.
[11:24] Dev Random: Has somebody done an audit on "ThreatLevel" for OSSL? e.g. osMakeNoteCard is listed as "High" for some reason. http://opensimulator.org/wiki/OsMakeNotecard
[11:25] Bill Blight: yes
[11:25] Bill Blight: because it can abuse your asset server
[11:25] Dev Random: ah.
[11:25] Bill Blight: if someone was to use it in a loop to create notecards
[11:25] Dev Random: got it
[11:25] Andrew.Hellershanks @hg.osgrid.org: IIRC, osMakeNC had a higher threat level than it does now.
[11:25] Andrew.Hellershanks @hg.osgrid.org: That doesn't mean it is any safer to use than before.
[11:25] Gavin.Hird @grid.xmir.org:8002: We're making a whole book here!
[11:26] Ubit Umarov: some that in past had a minor level now just have none
[11:27] Bill Blight: People may not see the reason for the threat levels until it is too late and someone has trashed their region/grid
[11:27] Andrew.Hellershanks @hg.osgrid.org: Dev, There was a review of the threat levels and some adjustments were made.
[11:27] Arielle.Popstar @hg.osgrid.org: cant put a limit on how often it loops?
[11:27] Ubit Umarov: they are listed at end osslenable.ini i think
[11:27] Bill Blight: then just use the data server arielle
[11:27] Bill Blight: would end up being the same thing
[11:27] Andrew.Hellershanks @hg.osgrid.org: Arielle, Careful. That sounds like you are asking for a throttle.  ;)
[11:27] Arielle.Popstar @hg.osgrid.org: question is Bill, were there know cases of it being abused?
[11:28] Ubit Umarov: tricky that memory arielle
[11:28] Bill Blight: yes there are known cases
[11:28] Bill Blight: just nobody talks about it because nobody wants to admit to being "taken down by a notecard"
[11:28] Ubit Umarov: we will not tell you about abuses like that
[11:28] Ubit Umarov: ( or anyone :p )
[11:28] Andrew.Hellershanks @hg.osgrid.org: Some functions are limited through built-in delays. Such delays I have added to the function description.
[11:29] Ubit Umarov: you all know that mantis does have a privacy option right?
[11:29] Ubit Umarov: that should be used on all security related issues
[11:29] Sheera Khan: sure I know :-)
[11:29] Andrew.Hellershanks @hg.osgrid.org: Bill, death by multiple paper cuts from the edges of all the NC's?  ;) hehe
[11:30] Dev Random: It does make sense from that PoV.
[11:31] Arielle.Popstar @hg.osgrid.org: well sometimes good for people to know the why of threat levels
[11:31] Bill Blight: Most of the threat levels do from that point of view
[11:32] Andrew.Hellershanks @hg.osgrid.org: Bottom line is if someone wants to grief a region they will find a way.
[11:32] Bill Blight: Maybe the specifics yes, but the crowd that wants everything enabled for everybody can't say they were not warned ..
[11:32] Ubit Umarov: and many times turns a minor issue, a larger one
[11:33] Arielle.Popstar @hg.osgrid.org: its a pain when getting script issues on every script because of threat levels
[11:33] Ubit Umarov: yeah they are a pain
[11:34] Bill Blight: Not for someone who understands them ;)
[11:34] Andrew.Hellershanks @hg.osgrid.org: You shouldn't have them on *every* script.
[11:34] Andrew.Hellershanks @hg.osgrid.org: If you do, you are doing something wrong.
[11:34] Arielle.Popstar @hg.osgrid.org: cant imagine the frustration it must be for those who have no control over the levels
[11:34] Bill Blight: working on a jet engine is a mystery to someone who does not know how to do it, that is why we have the ability to learn, or at least most of us ..
[11:35] Arielle.Popstar @hg.osgrid.org: not everyone aspires to be an opensim dev
[11:35] Dev Random: heh... I have a HUD that won't work here because it uses Dynamic Texture.
[11:35] Bill Blight: refusing or the inability to learn is not the fault of the subject matter
[11:35] Andrew.Hellershanks @hg.osgrid.org: It is one of the reasons I made sure to include the permission and threat level details to the OSSL wiki pages. Someone who sees a function listed as available will be able to see that it might not be available for them.
[11:39] Arielle.Popstar @hg.osgrid.org: what about V2 groups? Osgrid admins say they cant use it because it too laggy
[11:40] Bill Blight: V2 Groups does not scale well, not sure it was ever meant to
[11:40] Andrew.Hellershanks @hg.osgrid.org: I would have thought it was less likely to be laggy vs the phpxmlrpc based group system.
[11:40] Arielle.Popstar @hg.osgrid.org: then core needs a better group thing?
[11:40] Bill Blight: was meant for the "easy to use" crowd
[11:40] Ubit Umarov: v2 never had must testing
[11:41] Ubit Umarov: but no one can expect 5000 notices to be sent fast
[11:41] Ubit Umarov: and still having other services working
[11:41] Andrew.Hellershanks @hg.osgrid.org: Arielle, Is there a problem with all groups or just some? A group with a large membership is likely to have problems.
[11:41] Bill Blight: Why ? need a new one, the XMLRPC groups work fine for those willing to put in the work
[11:41] Arielle.Popstar @hg.osgrid.org: how do they do 50,000 in s/l then?
[11:41] Bill Blight: they use something more like the XMLRPC not a contained module
[11:42] Ubit Umarov: don't remember sl groups beeing that fast
[11:42] Bill Blight: they have servers dedicated to nothing but groups
[11:42] Gavin.Hird @grid.xmir.org:8002: SL groups are struggling too
[11:42] Ubit Umarov: groups never scale well
[11:42] Ubit Umarov: they are exponencial by nature
[11:42] Arielle.Popstar @hg.osgrid.org: well faster then the sort of delays i was hearing was experienced by osgrid when they tested it
[11:42] Gavin.Hird @grid.xmir.org:8002: it comes up as an issue on every TPV developer meeting
[11:43] Andrew.Hellershanks @hg.osgrid.org: Arielle, An additional followup question is when is the group system laggy? Is it during chats or when sending offline IMs (ie. for group notices)?
[11:43] Ubit Umarov: hte complain i had from paela was notices if i remember
[11:43] Bill Blight: You are kind of out of the loop andrew, this has been a known issue with V2 groups for a while with big groups sending notices
[11:43] Gavin.Hird @grid.xmir.org:8002: The other question is how much resources was thrown at it during testing?
[11:43] Andrew.Hellershanks @hg.osgrid.org: Ubit, that makes sense. It takes time to prep and send all the offline IMs.
[11:44] Ubit Umarov: IMs should have a central dispacher
[11:44] Ubit Umarov: groups even more
[11:44] Andrew.Hellershanks @hg.osgrid.org: I was just thinking of something along that line.
[11:45] Andrew.Hellershanks @hg.osgrid.org: Sending group notices should be sent out for some other thread or process to handle.
[11:45] Arielle.Popstar @hg.osgrid.org: what about some sort irc thing?
[11:45] Ubit Umarov: just having each region get the list of members online is ... yeack..
[11:45] Arielle.Popstar @hg.osgrid.org: doesnt that work wquicker?
[11:45] Andrew.Hellershanks @hg.osgrid.org: Who is going to volunteer to write such a thing?  :)
[11:45] Bill Blight: the slow down issue I'm pretty sure comes from checking presence to see who is online and who is offline and the hand off to the offline IM, but just guessing
[11:45] Ubit Umarov: they are done async
[11:46] Andrew.Hellershanks @hg.osgrid.org: There are some ugly looking SQL queries in the groups code.
[11:46] Ubit Umarov: that is why ppl complain they are slow
[11:46] Ubit Umarov: but just think about what is needed region side
[11:46] Ubit Umarov: and with groups with more than 5000 members
[11:47] Andrew.Hellershanks @hg.osgrid.org: There are also some other missing features in the groups code.
[11:47] Ubit Umarov: do you expect that to be fast ?
[11:47] Gavin.Hird @grid.xmir.org:8002: maybe some old tech like UUCP could come to the rescue for distribution
[11:47] Leighton.Marjoram @grid.outandproud.life:8002: *thought about it Ubit and my head hurts*
[11:47] Ubit Umarov: irc does it better
[11:47] Ubit Umarov: ( but also does not scale well )
[11:47] Gavin.Hird @grid.xmir.org:8002: IRC does not handle offline very well?
[11:48] Ubit Umarov: offline is not irc standard
[11:48] Bill Blight: would like to see XMPP come back to opensimulator
[11:48] Gavin.Hird @grid.xmir.org:8002: no but IM and group notices needs to be delivered offline too
[11:48] Ubit Umarov: xmpp ?? why ?
[11:48] Andrew.Hellershanks @hg.osgrid.org: Bill, when did it go away from OS?
[11:48] Leighton.Marjoram @grid.outandproud.life:8002: Jabbers @ Bill
[11:48] Ubit Umarov: well that is optional on opensim
[11:48] Bill Blight: can handle messages a lot faster and would allow external chat client
[11:48] Ubit Umarov: and forget all those are heavy
[11:49] Ubit Umarov: you want thing text chat.. think irc
[11:49] Ubit Umarov: want think..
[11:49] Bill Blight: yes have used the IRC module a bit
[11:49] Ubit Umarov: ( guess some grids did used a irc server )
[11:49] Andrew.Hellershanks @hg.osgrid.org: I have seen one grid that had an IRC gateway in use.
[11:50] Bill Blight: actually made some changes to the IRC module, to make it a little more chat friendly
[11:50] Sheera Khan: In Metropolis we had a webbased text chat - it wasn't used very often but when it failed it was sorely missed by some peole on mobiles
[11:50] Andrew.Hellershanks @hg.osgrid.org: It was mainly used in a grid user support capacity.
[11:50] Ubit Umarov: not telling there isn't something wrong with v2 on that
[11:50] Arielle.Popstar @hg.osgrid.org: some grids seem to even be able to send notices externally
[11:50] Ubit Umarov: just it will never scale well
[11:50] Arielle.Popstar @hg.osgrid.org: which osgrid cant
[11:51] Bill Blight: My group notices go out to the HG, but ONLY if the user is online
[11:51] Arielle.Popstar @hg.osgrid.org: with G+ disappearing, groups would be a handy alternative
[11:51] Andrew.Hellershanks @hg.osgrid.org: Arielle, that would be more easily doable with the v1 groups that uses an external website for some of the functionality.
[11:51] Gavin.Hird @grid.xmir.org:8002: Both Discord and Telegram have chat APIs
[11:53] Arielle.Popstar @hg.osgrid.org: has Diva been around? She did a lot of that IM and V2 code didnt she?
[11:53] Ubit Umarov: she made v2
[11:53] Gavin.Hird @grid.xmir.org:8002: Groups work up to a certain point, but are in reality quite limited
[11:53] Andrew.Hellershanks @hg.osgrid.org: Arielle, I haven't seen any sign of Diva in a long time.
[11:53] Ubit Umarov: in fact v2 is basicly derivative of xml
[11:53] Arielle.Popstar @hg.osgrid.org: well it is something of an alternative as G+ groups are often used for notices of events and stuff
[11:54] Ubit Umarov: just with opensim grid side
[11:54] Arielle.Popstar @hg.osgrid.org: is that code offlimits to anyone but her?
[11:54] Andrew.Hellershanks @hg.osgrid.org: There are a couple of improvements in the v2 groups code I was thinking about porting to v1 (if possible).
[11:54] Ubit Umarov: but since most kept using xmçl one it did not got much testing
[11:54] Ubit Umarov: even code
[11:54] Ubit Umarov: i did made changes to xml one
[11:55] Ubit Umarov: not as much to v2 for that reason ( was testing at osgrid )
[11:55] Bill Blight: Since most grids run their own websites, and have proper web servers the XMLRPC groups help handle the load
[11:55] Andrew.Hellershanks @hg.osgrid.org: I need to check on the latest available version of the php groups system. It has been ages since I installed a copy of it.
[11:55] Gavin.Hird @grid.xmir.org:8002: LL have fixed a bunch of bugs in groups in the viewer code the last few weeks, which is maybe also one reason why it has not worked so well
[11:56] Ubit Umarov: well all our IMs need revision
[11:56] Ubit Umarov: a lot of things are just wrong
[11:56] Andrew Hellershanks: Would be worth filing a mantis about it so the issue doesn't get lost.
[11:57] Arielle Popstar: yes, getting increasin failures on HG IM's
[11:57] Andrew Hellershanks: One thing I've thought about doing is updating the v1 groups code to include HG support.
[11:57] Arielle Popstar: increasing failures on HG jumps too for that matter
[11:58] Ubit Umarov: well not that clear reasons for HG fails
[11:58] Arielle Popstar: no
[11:58] Ubit Umarov: there can be tons of reasons
[11:58] Ubit Umarov: internet is not beening easier..
[11:59] Ubit Umarov: " you suposed to pay for a server"
[11:59] Ubit Umarov: " you are suposed to be big to house servers"!
[11:59] Arielle Popstar: but one issue i see is that on a failure message, enough of the avi has made it that on a another attempt, the server responds that one is already logged into the desitinations and needs to be logged out first
[11:59] Gavin.Hird @grid.xmir.org:8002: I updated to mono version 5.16.0.179, and things are running more stable
[12:00] Arielle Popstar: my connection today is same speed as the server i used to rent
[12:00] Ubit Umarov: well about mono and -net changes i have some reports that may point to something odd on threading
[12:00] Ubit Umarov: but not sure..
[12:01] Andrew Hellershanks: Before some of you start disappearing I want to change topic for just a moment. We can come back to this very shortly.
[12:01] Bill Blight: Home networking equipment is a far cry from enterprise stuff used in data centers, wire speed is not the only thing to consider
[12:02] Andrew Hellershanks: Has anyone noticed that particle effects work well in some regions but not others? This can be so even when the regions are mostly empty. Any idea of a possible cause?
[12:02] Bill Blight: other particles in the region you can see
[12:02] Bill Blight: viewer has a limit on the particles it will render
[12:02] Bill Blight: first come first serve
[12:02] Andrew Hellershanks: Bill, I know. That's why I said it happens in regions that are mostly empty.
[12:03] Bill Blight: I have not noticed it
[12:03] Andrew Hellershanks: In one grid I had a particle effect that worked well. I TP to the grid sandbox and it doesn't work properly.
[12:03] Bill Blight: and as particles are rendered viewer side, could be a viewer code change
[12:03] Gavin.Hird @grid.xmir.org:8002: I only see it when approaching the viewer limit
[12:03] Andrew Hellershanks: Same viewer. Just TP between regions to see difference.
[12:03] Bill Blight: willing to bet if you looked around you might find a particle emmiter there somewhere that is slamming into the limit
[12:04] Bill Blight: just beause a region appears empty does not mean it is
[12:04] Andrew Hellershanks: I am aware of the particle limit. I will check the Sandbox region for particle emitters but I bet I will find none.
[12:04] Gavin.Hird @grid.xmir.org:8002: smoke emitters can use a very large number of particles for instance
[12:05] Ubit Umarov: wel if regions are on same opensim version, script is the same, points to viewer thing
[12:05] Bill Blight: other thing there is the .001 bug with particles , I might be missing a 0 there, where an emitter with a fast burst rate will trump other particle emittters
[12:05] Ubit Umarov: on that case test not tp, but relog on target region
[12:06] Ubit Umarov: ( so viewers does not have any particles in memory =
[12:06] Andrew Hellershanks: At least Ubit hasn't said it is one of those issues that only I seem to have.  :)
[12:06] Bill Blight points to Andrew, welllllllll
[12:06] Ubit Umarov: ( no but did think it :p )
[12:06] Arielle Popstar: wait for it....
[12:07] Andrew Hellershanks: Its an annoying problem. As I want to sell my particle generator via Kitely it would be a problem if some buyers find it doesn't work properly.
[12:07] Andrew Hellershanks: Arielle, :)
[12:07] Ubit Umarov: ahh kitely.. 0.8.x
[12:07] Ubit Umarov: we don't know about 0.8 remember?
[12:07] Arielle Popstar: them and metro last 2 large grids
[12:08] Andrew Hellershanks: Ubit, you can sell items via Kitely MP that get delivered to grids with different versions of OS.
[12:08] Gavin.Hird @grid.xmir.org:8002: Oz Linden just sent a message that http assets will be mandatory from January 6, 2019
[12:08] Bill Blight: I can't count below 9 anymore, do numbers still exist below that
[12:08] Ubit Umarov: ( i actually don't.. never used it :) )
[12:08] Gavin.Hird @grid.xmir.org:8002: so many existing viewers will not be able to log on to SL after that
[12:08] Ubit Umarov: Gavin we are not SL
[12:08] Arielle Popstar: maybe thats why Gavin
[12:09] Arielle Popstar: will halt a fair bit of copybot
[12:09] Andrew Hellershanks: Gavin, that will be a problem for Linux users as LL no longer cares about supporting it in their viewer.
[12:09] Gavin.Hird @grid.xmir.org:8002: it will be a problem for many
[12:09] Ubit Umarov: will not do any blind change on opensim
[12:10] Bill Blight: What is this SL you speak of? Is it some ancient beast from the time of legend that no longer has teeth?
[12:11] Ubit Umarov: and really take the LLUDP payload and send it via HTTP is a great improvement?
[12:11] Gavin.Hird @grid.xmir.org:8002: they seem to think so
[12:11] Ubit Umarov: it mb for them, since they move load out of regions to cdn(?)
[12:11] Bill Blight: to me it seems to be a move to allow them to move to cheaper cloud hosting
[12:12] Gavin.Hird @grid.xmir.org:8002: I guess the CDN providers are not too keen on supporting LLUDP
[12:12] Ubit Umarov: no one likes udp does not fit well on their paying methods
[12:12] Bill Blight: well considering LLUDP is not a "real" protocol, just one they made up ..
[12:13] Ubit Umarov: lludp is a protocol
[12:13] Andrew Hellershanks: I like one persons signature block in their email. It says "I know a joke about UDP but you might not get it."  :D
[12:13] Bill Blight: I was being sarcastic
[12:13] Ubit Umarov: several like it where proposed at that time
[12:13] Ubit Umarov: it does fill a gap on protocols
[12:13] Ubit Umarov: TCP is a stream oriented prococol
[12:14] Ubit Umarov: lludp like othes are Message oriented protocol
[12:14] Ubit Umarov: but its is a protocol ( and it is not udp )
[12:14] Ubit Umarov: it you say it is udp, they you can also say tcp is also udp
[12:14] Bill Blight: I know it is a protocol, just not an "Industry standard one"
[12:15] Ubit Umarov: ( actually is not udp and tcp have a common parent )
[12:15] Ubit Umarov: it is a industry standard
[12:15] Ubit Umarov: it is defined.. etc etc
[12:15] Ubit Umarov: if only one company uses it it's a detail ;)
12:16] Ubit Umarov: but several others where proposed, guess none in large use, because http paranoia
[12:17] Andrew Hellershanks: TCP is like having a two-way conversation. With UDP you may just be talking to yourself.
[12:17] Ubit Umarov: of course seems some lindens also have no idea what lludp is now
[12:17] Ubit Umarov: no
[12:17] Ubit Umarov: tcp is creating a wire
[12:17] Ubit Umarov: that stays there
[12:18] Andrew Hellershanks: Ubit, I was only giving a very rough analogy to someone talking with their voice.
[12:18] Ubit Umarov: message protocols are just sending one message
[12:18] Bill Blight: LLUDP does not have an RFC so it is not an industry standard .. (not offically)
[12:18] Ubit Umarov: rfc are just requests
[12:18] Andrew Hellershanks: TCP expects some form of reply to messages sent out.
[12:18] Gavin.Hird @grid.xmir.org:8002: so UDP is like a message in a bottle. Someone might receive it.
[12:19] Ubit Umarov: they are not norms either :)
[12:19] Ubit Umarov: hmm? gavin ?
[12:19] Ubit Umarov: udp does have a destination
[12:19] Dev Random: I like that analogy, Gavin.
[12:19] Arielle Popstar: so udp requires less bandwidth?
[12:19] Ubit Umarov: its like tcp on that
[12:19] Andrew Hellershanks: Gavin, right. UDP also doesn't guarantee that the order sent is the same as received.
[12:19] Arielle Popstar: overall?
[12:19] Ubit Umarov: udp is not multicast
[12:19] Bill Blight: UDP is the shotgun of network protocols, point it in the right direction , fire, and hope it hits something
[12:20] Leighton.Marjoram @grid.outandproud.life:8002: I always remember UDP as a "Fire and forget" protocol.
[12:20] Ubit Umarov: even multicast has a list of receivers
[12:20] Arielle Popstar: as not destination response is required
[12:20] Gavin.Hird @grid.xmir.org:8002: you could say that a message in a bottle has a detination too. It migh just never reach it
[12:20] Ubit Umarov: and tcp is also like that
[12:20] Bill Blight: but TCP sends an ACK
[12:20] Andrew Hellershanks: Group chat is done via UDP. Sometimes text you entered in to the group chat window comes out in a different order when you finally what you typed echoed back through the group.
[12:20] Ubit Umarov: just tcp adds a ton of more things to clear it all
[12:20] Bill Blight: UDP you never know if it got there even if it does
[12:20] Ubit Umarov: ludp also has acks
[12:21] Ubit Umarov: tcp is a LOT more complex than this
[12:21] Bill Blight: LLUDP yes, not UDP
[12:21] Ubit Umarov: well i said lludp is not udp :)
[12:21] Bill Blight: LLUDP is a bastard child of TCP/UDP
[12:21] Andrew Hellershanks: hehe
[12:21] Ubit Umarov: no
[12:21] Ubit Umarov: it is a protocol  :p
[12:22] Ubit Umarov: not that bad one actually
[12:22] Ubit Umarov: just tricky
[12:22] Ubit Umarov: and got broken in time by ll themselves
[12:23] Ubit Umarov: some things are just broken bc the linden that made them had no idea what lludp is
[12:23] Ubit Umarov: happens :)
[12:23] Ubit Umarov: bc it is tricky
[12:23] Bill Blight: Until I start seeing LLUDP in the list of names for IPTABLES, Cisco Routing, Foundry routers and other BIG IRON names, it is not a STANDARD ...
[12:23] Gavin.Hird @grid.xmir.org:8002: just like the renderer. Nobody in LL knows how it works any more
[12:24] Ubit Umarov: and industry always hated udp
[12:24] Bill Blight: but they supported it
[12:24] Ubit Umarov: cisco routers just refused it
[12:24] Bill Blight: LOL
[12:24] Bill Blight: no
[12:24] Bill Blight: badly configured ones refused it
[12:24] Bill Blight: LOL
[12:24] Ubit Umarov: it was the audio things what changed that
[12:24] Ubit Umarov: no cisco oficially refused to suport udp
[12:24] Gavin.Hird @grid.xmir.org:8002: every time they try and fix it it falls over and renders something something white
[12:25] Ubit Umarov: when audio streams started to show up and then video, they where forced to change
[12:25] Sheera.Khan @hypergrid.org:8002: @Arielle: yes UDP uses less bandwidth as the header is 12 Byte smaller than the TCP header, and you don't have the 3 way handshake to start and stop a connection
[12:25] Ubit Umarov: but they hate it anyways :p
[12:26] Ubit Umarov: on a perfect line udp is superion to tcp
[12:26] Bill Blight: When they started supporting it is a moot point because they did
[12:26] Bill Blight: LOL
[12:26] Ubit Umarov: just there are no perfect lines :)
[12:26] Bill Blight: nope
[12:26] Bill Blight: the network I used to manage had over 3000 Cisco routers and a thousand or so other brands, and it was a monster to make them all speak to each other
[12:26] Dev Random: unrelated note.... I'm hearing Christmas music streaming here.
[12:27] Ubit Umarov: in fact there was a move into use of udp for a lot of things back in the day
[12:27] Andrew Hellershanks: Dev, poor you.
[12:27] Ubit Umarov: where the error rate of comms did improved
[12:27] Andrew Hellershanks: Way too early for Christmas music.
[12:27] Bill Blight: haha
[12:27] Bill Blight: my chirstmas playlist got activated
[12:27] Bill Blight: oops
[12:27] Leighton.Marjoram @grid.outandproud.life:8002: *clicks mute* "not ready for Christmas yet"
[12:28] Bill Blight: me either on my way to fix that
[12:28] Ubit Umarov: and ll did worse
[12:28] Ubit Umarov: they don't use tcp
[12:28] Andrew Hellershanks: Fortunately my viewer is not set to autoplay audio streams in regions.
[12:28] Gavin.Hird @grid.xmir.org:8002: Apropos music, "Forensics expert ‘cut to pieces’ #Saudi journalist’s body as colleagues listened to music"...
[12:28] Ubit Umarov: they use http
[12:28] Ubit Umarov: that is layers on top of tcp
[12:29] Ubit Umarov: i have no reason to not add suport for assets
[12:29] Ubit Umarov: but no viewer asked for it
[12:30] Bill Blight: and I doubt stock firestorm every will in opesnim
[12:30] Bill Blight: as they seem to force disable it
[12:30] Bill Blight: (4 lines of code to fix that , btw)
[12:31] Ubit Umarov: to do such a change i should have a request from viewers with full spec of the new protocol messages
[12:31] Bill Blight: (4 lines in the viewer I should say)

Personal tools
General
About This Wiki