Chat log from the meeting on 2016-04-05

[11:02] Andrew Hellershanks: Afternoon, everyone.
[11:04] Andrew Hellershanks: small group today
[11:04] Gavin.Hird @grid.xmir.org:8002: yep
[11:05] Simulator Version v0.5 ruft: OpenSim 0.9.0.0 Dev 2cfe848: 2016-03-24 16:43:27 -0400 (Unix/Mono)
[11:06] Kayaker Magic: I want someone to explain My Suitcase to me, and tell me why it is disabled in OSgrid.
[11:07] Gavin.Hird @grid.xmir.org:8002: oh, that. I'm listening :-))
[11:07] Kayaker Magic: Uh oh, blind leading the blind again?
[11:07] Billy.Bradshaw @hg.viewtwo.net:8600: I saw the comments on the mantis
[11:08] Kayaker Magic: So who is in charge of setting up things like this in OSgrid?
[11:09] Andrew Hellershanks: I don't even have a "My Suitcase" folder in the viewer I'm using at the moment.
[11:09] Equus Staedler: There is and never was a Suitcase on OSgrid
[11:09] Kayaker Magic: For those who don't know, I'm talking about mantis #7868
[11:09] Kayaker Magic: I never noticed before, usually I HG into OSgrid from elsewhere.
[11:11] Kayaker Magic: OK, so assuming that no My Suitcase was done on purpose, should I submit a second mantis that says it displays (unavailable) in inventory when it should not?
[11:12] Billy.Bradshaw @hg.viewtwo.net:8600: When HG started you had to HG somewhere and acquire an itemm then when you returned home the my suitcase folder set was created.. bit odd at the time
[11:12] Gavin.Hird @grid.xmir.org:8002: when is it unavailable?
[11:12] Kayaker Magic: If you log into OSgrid, then HG to another grid, everything in your inventory says (unavailable) even though it is available.
[11:13] Billy.Bradshaw @hg.viewtwo.net:8600: The orehn made it into a service that HG enabled the folder set was created when you logged in
[11:13] Gavin.Hird @grid.xmir.org:8002: yes
[11:16] Gavin.Hird @grid.xmir.org:8002: The most updated viewers only recognize folders of type 100 as My Suitcase, but grids like OSgrid send them as type 9 to the viewer
[11:16] Andrew Hellershanks: I rarely HG TP anywhere so I'm not used to having, or seeing, a suitecase folder.
[11:16] Gavin.Hird @grid.xmir.org:8002: if they send anything at all
[11:17] Andrew Hellershanks: I do remember there was a change in the number for one of the folders. That may be it. i'd have to dig up the specifics of it.
[11:17] Billy.Bradshaw @hg.viewtwo.net:8600: I got the impression in the early days of HG that many people did not see it becoming as popular as it has
[11:17] Gavin.Hird @grid.xmir.org:8002: There was a discussion last fall and some commits
[11:17] Kayaker Magic: So who maintains the central servers for OSgrid?
[11:17] Kayaker Magic: Nebadon?
[11:18] Ken Savage: I think Dan does most of it now
[11:18] Billy.Bradshaw @hg.viewtwo.net:8600: According to recent posts Dan Banner heads up OSGrid these days
[11:18] Gavin.Hird @grid.xmir.org:8002: As I have said before, the problem with OSGrid as test grid, it does not fully implement out-of-the-box opensimulator, but something else too
[11:18] Gavin.Hird @grid.xmir.org:8002: so it is hard to test everything against
[11:19] Sheera Khan: Metropolis has a rather plain vanilla testgrid, would you be interested in an account there?
[11:19] Andrew Hellershanks: Major grids won't want to run with an "out of the box" configuration. There are always changes and tweaks applied to get the best performance
[11:20] Kayaker Magic: Well, I have enough things to test as it is, I could throw up my hands and let someone else decide if this is a bug or a feature.
[11:20] Gavin.Hird @grid.xmir.org:8002: Yes of course, but people who make viewers need a standard to test against
[11:20] Andrew Hellershanks: Gavin, what features do you feel you can't test due to osgrids current configuration?
[11:20] Gavin.Hird @grid.xmir.org:8002: same with plugins
[11:21] Gavin.Hird @grid.xmir.org:8002: profiles
[11:21] Gavin.Hird @grid.xmir.org:8002: trash
[11:21] Gavin.Hird @grid.xmir.org:8002: groups
[11:21] Gavin.Hird @grid.xmir.org:8002: my suitcase
[11:21] Gavin.Hird @grid.xmir.org:8002: no postgres
[11:21] Andrew Hellershanks: If you need an out of the box set up you can always set up a copy of OpenSim for use in testing.
[11:22] Gavin.Hird @grid.xmir.org:8002: yes, and we do
[11:22] Billy.Bradshaw @hg.viewtwo.net:8600: Calling cards appearing within the suitcase and somehow duplicated to the root level folders
[11:22] Gavin.Hird @grid.xmir.org:8002: BUT, a lot of bugs only show up under load
[11:22] Gavin.Hird @grid.xmir.org:8002: with many testers
[11:22] Sheera Khan: yes, but the problem would be the missing avatars in such a lonely constellation
[11:23] Andrew Hellershanks: Testing OpenSim itself is also difficult unless it is tested in a more real world situation.
[11:23] Ken Savage: I have tried FSAssets twice now with problems each time.
[11:23] Ken Savage: Anyone here had any luck with it?
[11:23] Gavin.Hird @grid.xmir.org:8002: but then the out of the box needs to be put through its phases
[11:24] Billy.Bradshaw @hg.viewtwo.net:8600: No Ken
[11:24] Sheera Khan: @Ken, I didn't try yet
[11:24] Gavin.Hird @grid.xmir.org:8002: how do you even know if the asset server works properly when the standard is not tested on OSGrid?
[11:24] Equus Staedler: All info about OSGrid can be found here: http://wiki.osgrid.org/index.php/Main_Page
[11:25] Sheera Khan: Most of the assets are at their right place the day after ...
[11:25] Andrew Hellershanks: Gavin, right out of the box would mean using SQLite as the database.
[11:25] Gavin.Hird @grid.xmir.org:8002: no
[11:25] Ken Savage: OSGrid is using FSAssets, but I am not sure the ini instructions on using it are all complete and correct
[11:25] Gavin.Hird @grid.xmir.org:8002: right out of the box supports 3 databases
[11:25] Gavin.Hird @grid.xmir.org:8002: SQLite, MySQL and Postgres
[11:25] Gavin.Hird @grid.xmir.org:8002: all needs to be tested
[11:25] Gavin.Hird @grid.xmir.org:8002: under load
[11:26] Gavin.Hird @grid.xmir.org:8002: with all code changes where releavant
[11:26] Kayaker Magic: No need to test SQLite under load, it should never be used under load!
[11:26] Gavin.Hird @grid.xmir.org:8002: we kindoff agree on that :-)
[11:26] Billy.Bradshaw @hg.viewtwo.net:8600: Why do you say that Kaymaker?
[11:26] Andrew Hellershanks: Gavin, Right out of the box it is only set for SQLite. You have to change the settings to use one of the other databases. If you are concerned about osgrid not being "out of the box", is the problem with testing that it isn't using core versions of some features?
[11:27] Andrew Hellershanks: Kayaker, agreed
[11:27] Gavin.Hird @grid.xmir.org:8002: right out of the box in grid mode is not SQLite?
[11:27] Billy.Bradshaw @hg.viewtwo.net:8600: I am testing SQLite, works well with a few changes
[11:27] Billy.Bradshaw @hg.viewtwo.net:8600: Things like groups (core) require implementation
[11:27] Gavin.Hird @grid.xmir.org:8002: I mean right out of the box is not standalone
[11:28] Ken Savage: Since SQLLite is the default, that one gets tested more by others than any other one, next MySQL.
[11:29] Andrew Hellershanks: Out of the box you have to decide what mode you want to run in. It isn't set for anything until you uncomment one of the lines in the Architecture section of OpenSim.ini
[11:29] Ken Savage: Since no one is really doing any coding right now, testing won't do much good because any bugs found won't get fixed.
[11:29] Gavin.Hird @grid.xmir.org:8002: Right out of the box in both standlone and grid mode needs to be tested to its full capabilities
[11:30] Ken Savage: It's my understanding that Melanie and UBit are deep in a new viewer project for opensim.
[11:30] Gavin.Hird @grid.xmir.org:8002: otherwise nobody knows if it breaks or some other randomly added component breaks it
[11:30] Billy.Bradshaw @hg.viewtwo.net:8600: Ken, there are some projects underway, will take time for people to become familiar with the code structure
[11:31] Gavin.Hird @grid.xmir.org:8002: Ken does that mean that 0.9 is hanging there in thin air while a new viewer is developed?
[11:31] Ken Savage: It sounds like major changes will be required in opensim for the new viewer
[11:31] Ken Savage: That is my suspect, but I do not know that for sure
[11:32] Gavin.Hird @grid.xmir.org:8002: Maybe they should fork it for that project
[11:32] Gavin.Hird @grid.xmir.org:8002: and complete 0.9
[11:32] Andrew Hellershanks: Ken, There are other people who make changes to the code other than Melanie and Ubit. They are the main two who work on it.
[11:32] Ken Savage: Yes, but no one else has done anything to speak of in a long time now.
[11:33] Andrew Hellershanks: If by long time you mean a couple of weeks, it would appear so as March 23 was the last time a change was pushed to master.
[11:33] Ken Savage: Yes, and that was a minor tweak
[11:34] Ken Savage: Code changes haven't been this slow coming in many many years
[11:34] vegaslon plutonian: but for a few months the code changes came fast
[11:35] Ken Savage: That is why I suspect they are waiting to see what happens with a new viewer.
[11:35] Andrew Hellershanks: Other than the 3,000 or so changes that got dumped in at one time and the almost 500 changes made since then.
[11:35] Billy.Bradshaw @hg.viewtwo.net:8600: Alicia Ravern did the core commit for FSAsset, maybe a good source of info.
[11:35] Ken Savage: Yes
[11:36] Ken Savage: Neb's farewell address on his blog seemed to indicate he was backing off because there was almost no development being done after the big avn merge.
[11:37] Billy.Bradshaw @hg.viewtwo.net:8600: Alicia if you read these notes can you join us next week?
[11:37] Ken Savage: I am sure there were other reasons, but that was mentioned
[11:37] Gavin.Hird @grid.xmir.org:8002: Would it be possible to back out to 0.8.3 (pre avination merge) and make smaller changes from there and let the big merge have more time and resources before release?
[11:38] Andrew Hellershanks: Ken, what issue(s) are you having with FSAssets?
[11:38] Ken Savage: Nooooo, the avn merge is way better
[11:38] Gavin.Hird @grid.xmir.org:8002: I feel it is disruptive with so many things not working properly and it kindoff is pushed on everyone by the sheer size of OSgrid?
[11:39] Ken Savage: I get a lot of Robust startup error messages and av appearance is not working
[11:39] Andrew Hellershanks: Gavin, There should not be any major issues in master. If you are finding something isn't working you should report it.
[11:39] Ken Savage: But thats only with fsassets, with normal MySQL it is fine
[11:39] Billy.Bradshaw @hg.viewtwo.net:8600: Kev, with FSAssets or in general with 0.8?
[11:39] Ken Savage: I run master and other than FSAssets it runs great
[11:39] Billy.Bradshaw @hg.viewtwo.net:8600: 0.9?
[11:39] Ken Savage: yes
[11:40] Billy.Bradshaw @hg.viewtwo.net:8600: Ken, windows or linux or iOS
[11:40] Ken Savage: I honestly think the issue with FEAssets is incorrect or incomplete info in the ini instructions
[11:40] Ken Savage: Windows, but most are running it on Linux
[11:41] Gavin.Hird @grid.xmir.org:8002: the biggest issue is region crossings
[11:42] Gavin.Hird @grid.xmir.org:8002: where many have reported this
[11:42] Equus Staedler: It is...
[11:43] Andrew Hellershanks: Gavin, What is the problem with region crossings?
[11:44] Billy.Bradshaw @hg.viewtwo.net:8600: You can always visit me Gavin, but I have a single micro grid with external access
[11:44] Gavin.Hird @grid.xmir.org:8002: there are problems with crossing both on var regions and standard regions
[11:44] Gavin.Hird @grid.xmir.org:8002: particularly if regions are running on instances on the same IP
[11:45] Gavin.Hird @grid.xmir.org:8002: it has been discussed in this meeting the previous weeks
[11:45] Ken Savage: The issue of TPing out of a large heavy prim var is a viewer issue common to all viewers.
[11:45] Billy.Bradshaw @hg.viewtwo.net:8600: Can you be more specific with the nature of the problems?
[11:46] Gavin.Hird @grid.xmir.org:8002: it reports that the region you cross into cannot be found
[11:46] vegaslon plutonian: I have found that can be a problem with the loopback on some routers, where the loopback works but packets are lost when you have more then one instance behind it
[11:46] Kayaker Magic: I have two instances running on one server, if they are two vars next to each other, an avatar cannot fly back and forth across the border without getting errors and becoming locked up and unable to TP.
[11:46] Gavin.Hird @grid.xmir.org:8002: 2016-03-04 16:15:18,815 DEBUG (Heartbeat-(Skomrak)) - OpenSim.Region.Framework.Scenes.ScenePresence [SCENE PRESENCE] CheckForBorderCrossing: position outside region. Gavin Hird in Skomrak at pos <5.131564, 256.0646, 35.84481>
[11:47] Andrew Hellershanks: Ken, Regarding FSAssets, Alicia worked on the db connector and HG part. Melanie has made fixes since then so she doesn't know if the instructions are 100% or not.
[11:47] Gavin.Hird @grid.xmir.org:8002: 2016-03-04 16:15:18,828 DEBUG (Heartbeat-(Skomrak)) - OpenSim.Services.Connectors.Simulation.SimulationServiceConnector [REMOTE SIMULATION CONNECTOR]: QueryAccess to http://sim1.xmir.org:8301/agent/97d02b42-9119-4708-804e-4c7d699162ae/644e1e53-76c9-4ae3-8146-7624831b65cd/ returned True, reason , version 0/0
2016-03-04 16:15:18,832 DEBUG (Heartbeat-(Skomrak)) - OpenSim.Region.CoreModules.Framework.EntityTransfer.EntityTransferModule [ENTITY TRANSFER MODULE] GetDestination: region not found. Old region name=Skomrak at <1000,1002> of size <256,256>. Old pos=<5.131564, 256.0646, 35.84481>
2016-03-04 16:15:18,834 DEBUG (Heartbeat-(Skomrak)) - OpenSim.Region.Framework.Scenes.ScenePresence [SCENE PRESENCE] CheckForBorderCrossing: Crossing failed. Restoring old position.
[11:47] Andrew Hellershanks: Kayaker, are the two vars the same size?
[11:48] Kayaker Magic: Yes
[11:48] Billy.Bradshaw @hg.viewtwo.net:8600: Gavin version 0/0 looks strange
[11:48] Gavin.Hird @grid.xmir.org:8002: yes
[11:48] Gavin.Hird @grid.xmir.org:8002: but the crossing works on direct TP
[11:48] Billy.Bradshaw @hg.viewtwo.net:8600: that should be 6/6 if both using 0.8
[11:48] Gavin.Hird @grid.xmir.org:8002: so fly /walk fails
[11:48] Billy.Bradshaw @hg.viewtwo.net:8600: 0.9
[11:49] Billy.Bradshaw @hg.viewtwo.net:8600: or maybe it is a different mechanism
[11:49] Kayaker Magic: When it works, (only on 1x1 vars for me) border crossing in 0.9 is GREAT!
[11:50] Gavin.Hird @grid.xmir.org:8002: probably but it should conceptually be the same lookups / calculations being done
[11:50] Billy.Bradshaw @hg.viewtwo.net:8600: version is normally associated with inter region comms during HG
[11:50] Ken Savage: TPs between older sims and newer ones will always be an issue until more people upgrade
[11:50] Andrew Hellershanks: Ken, I was just talking with Alicia. If you can tell me what mantis covers the issues you are having with FSAssets I will pass along the number and she will have a look in about an hour or two.
[11:52] Andrew Hellershanks: Gavin, are the regions you are flying/walking into running the same version of OpenSim as the region you are leaving?
[11:52] Gavin.Hird @grid.xmir.org:8002: the changes between 0.8.2 and 0.9 for entity transfer and scene presence is staggering so lots of things can go wrong there :-)
[11:52] Kayaker Magic: Ken, I have started over 3 or 4 times with fresh copies of 0.9, set up two vars next to each other with identical code, still cannot cross more than once. It is not old-vs-new problem.
[11:52] Gavin.Hird @grid.xmir.org:8002: yes they are
[11:52] Ken Savage: There is no mantis about the FSAssets issue yet. I am still hoping she or someone can send me a robust ini file configured for HG and FSAssets that is working on another grid, to make sure its not just my config.
[11:53] Ken Savage: Var to var borders crossings are always hit and miss, not sure why but it's been that way forever
[11:54] vegaslon plutonian: never really had problems with my three 2x2 var regions, crossing always work fine
[11:55] Gavin.Hird @grid.xmir.org:8002: are they all running in the same simulator instance?
[11:55] Gavin.Hird @grid.xmir.org:8002: or 3 different?
[11:55] vegaslon plutonian: yes they are
[11:55] Andrew Hellershanks: Ken, if you are using settings from the wiki it should be fine. Alicia is asking you to create a mantis with those errors and your asset config section and she will take a look later tonight
[11:55] Gavin.Hird @grid.xmir.org:8002: try run them in 3 instances and see how it goes?
[11:55] Kayaker Magic: I have not tried 2x2, had the problem with crossings on 3x3s and 4x4s
[11:56] Ken Savage: I will have to reconfig for FSAssets on my test grid and try it again tonight and save the error messages
[11:57] Gavin.Hird @grid.xmir.org:8002: there are no problems crossing regions in the same instance, only between instances
[11:57] vegaslon plutonian: it used to only work with them in seperate instances only recent that same instance works
[11:57] Andrew Hellershanks: Ken, ok. when you have a mantis number I can pass it along to Alicia
[11:58] Gavin.Hird @grid.xmir.org:8002: did you change that before or after taking the 0.9 code?
[11:59] vegaslon plutonian: been broke a few times before 0.9 and then when the avination merge happened it got broke again and had to be fixed
[11:59] Gavin.Hird @grid.xmir.org:8002: right
[11:59] Andrew Hellershanks: Ken, Alicia has to go but will ping you when she returns.
[11:59] Ken Savage: I am in IRC chat with Alicia now :)
[12:00] Andrew Hellershanks: :)
[12:00] Andrew Hellershanks: Wow. The hour has flown by
[12:00] Kayaker Magic: Another meeting to go to, gota run.
[12:00] Gavin.Hird @grid.xmir.org:8002: Perhaps the code now thinks that crossing to a different instance is a HG transfer, but it isn't so it fails
[12:00] Ken Savage: She is busy right now, but will maybe be able to send me a robust.ini for FSAssets in about an hour
[12:00] Andrew Hellershanks: ok, tc, Kayaker
[12:00] Kayaker Magic: Well, it has been scary guys....
[12:01] vegaslon plutonian: moving from too connected regions has always been the more complex of the region transfers
[12:01] Andrew Hellershanks: I have a test grid running using 0.9 that has one region per instance so I can try region crossings.
[12:01] Gavin.Hird @grid.xmir.org:8002: that'd be great!
[12:02] Equus Staedler: Region crossings are a major issue for sure, would be terrific if that could be resolved. Never been perfect but was not an issue before 0.9
[12:03] Gavin.Hird @grid.xmir.org:8002: region crossings on a grid is pretty fundamental
[12:03] Equus Staedler: Yes it is.
[12:03] Gavin.Hird @grid.xmir.org:8002: there is perhpas more acceptance that HG transfers fails
[12:03] vegaslon plutonian: when they work in 0.9 they work better just some corner case here where they are not working
[12:03] Equus Staedler: Very annoying when you are forced to relog because you get stuck
[12:03] Andrew Hellershanks: vegas, logins and region crossing both involve a lot of things happening.
[12:04] Billy.Bradshaw @hg.viewtwo.net:8600: Thing is with a grid your neighbors could anywhere using a whole host of different environments
[12:04] Billy.Bradshaw @hg.viewtwo.net:8600: non deterministic
[12:05] Gavin.Hird @grid.xmir.org:8002: but it should work all things being equal
[12:05] Simulator Version v0.5 shouts: OpenSim 0.9.0.0 Dev 2cfe848: 2016-03-24 16:43:27 -0400 (Unix/Mono)
[12:05] Lucy Afarensis: And if they are offline causes problems
[12:05] Andrew Hellershanks: BTW, AFAIK, regarding folder numbers the root folder had the wrong number and is now using #9 and 100 had been agreed upon for the HG suitcase folder.
[12:06] Gavin.Hird @grid.xmir.org:8002: exactly Andrew
[12:06] Gavin.Hird @grid.xmir.org:8002: I think the fear on OSGrid is that people have My Suitcase folder with all kinds of folder numbers, so it is not trivial to migrate them to type 100
[12:07] Gavin.Hird @grid.xmir.org:8002: Anyway, thanks for a good meeting!
[12:08] Andrew Hellershanks: I don't think there should be "all kinds of folder numbers". I only know of those two changes re: folder numbers being discussed a while back. Any required migration in osgrid should not be that hard.
[12:08] Gavin.Hird @grid.xmir.org:8002: The reason why they have all kinds of numbers are it was a user defined folder, and it can have all kinds of numbers
[12:09] Billy.Bradshaw @hg.viewtwo.net:8600: I see what you are saying Gavin
[12:09] Gavin.Hird @grid.xmir.org:8002: OSgrid has data way back. I have data back to 2008 there
[12:10] Andrew Hellershanks: If there are HG folders with different numbers the name could be used to set them with the correct folder number.
[12:10] Billy.Bradshaw @hg.viewtwo.net:8600: I think if the system created a new my suitcase, it would duplicate the name, you could then distinguish by the fact the new one was empty
[12:12] Gavin.Hird @grid.xmir.org:8002: We tried that and it did not work very well because the viewer code only accepts one with that name now
[12:12] Gavin.Hird @grid.xmir.org:8002: so you basically get an empty folder
[12:12] Billy.Bradshaw @hg.viewtwo.net:8600: ah
[12:12] Gavin.Hird @grid.xmir.org:8002: but all the data still exist in the database
[12:12] Andrew Hellershanks: That would be something to ask the people who run osgrid.
[12:13] Billy.Bradshaw @hg.viewtwo.net:8600: can you rename a folder from the simulator?
[12:13] Gavin.Hird @grid.xmir.org:8002: you can rename with a query in the database
[12:13] Andrew Hellershanks: Folder renames would be done in the database, or in a viewer if it isn't a system folder.
[12:13] Billy.Bradshaw @hg.viewtwo.net:8600: k
[12:14] Andrew Hellershanks: Any other topics before we wrap up this weeks meeting?
[12:14] Gavin.Hird @grid.xmir.org:8002: We have a release build of Kokua 4.0.2 ready with a lot of bug fixes
[12:14] Gavin.Hird @grid.xmir.org:8002: for all platforms
[12:14] Andrew Hellershanks: ok. I'll update my copy of the viewer when its out.
[12:15] Sheera.Khan @hypergrid.org:8002: just a short question off topic: do var-regions have to be square still?
[12:15] Gavin.Hird @grid.xmir.org:8002: we are just waiting for LL to place the update versions on their repo. The autoupdate only works on SL logins, and we are dependent on them for that to work
[12:16] Andrew Hellershanks: Gavin, downloading it now
[12:16] Gavin.Hird @grid.xmir.org:8002: very nice
[12:16] Andrew Hellershanks: Sheera, yes. They still need to be square. I think it may be a viewer issue if they aren't.
[12:17] Sheera.Khan @hypergrid.org:8002: ok, ty for that :-)
[12:19] Andrew Hellershanks: ok, that's it for this week. See you next week.