Chat log from the meeting on 2019-03-12

From OpenSimulator

Jump to: navigation, search

[11:00] Andrew Hellershanks: Hello, everyone.
[11:00] Gavin.Hird Hi Andrew
[11:00] Selby.Evans hi andrew
[11:01] Andrew Hellershanks: Hm... no Sheera here today. That makes me the keeper of the log for today.
[11:01] Bill Blight: yeah it does
[11:01] Gavin.Hird Perhaps not aware of the US summer time
[11:02] Andrew Hellershanks: oh, yes. That's right. The clocks changed this past weekend. That always makes for a certain amount of confusion for a couple of weeks until everyone gets back in sync.
[11:04] Gavin.Hird The PI is still running a VAR. I managed to compile bulletsim for it too
[11:05] Andrew Hellershanks: I didn't get the updates to the OSSL pages done this past week. I got a few other things done so that task is now at the top of my list. I'll be creating the new OSSL pages after this meeting.
[11:05] Bill Blight: I've been hiding most of the week, and may hide for another one
[11:06] Andrew Hellershanks: Hello, Leighton.
[11:06] Leighton.Marjoram Hello Andrew
[11:06] Andrew Hellershanks: Bill, forced vacation or hiding from someone? ;)
[11:06] Leighton.Marjoram hello everyone
[11:06] Bill Blight: Self induced Exile
[11:06] Gavin.Hird I have updated the keywords for a new viewer built yet to be release, but tooltip texts are still welcome
[11:07] Vincent.Sylvester I don't have much this week, figured out that compiling under mono below 5.12 can result in stuck linker, upgrading to 5.18 so far seems more stable, but time will tell if that is true. Also setup a mirror for creating binaries of master when commits happen, because why not eh. Spent a bit of effort fiddling with shaders, but nothing yet on getting that pesky water fog below Z=0 taken care of. Getting decent shaders for ocean exploration down to Z=-100 would really enhance some terrains.
[11:08] Gavin.Hird I think the issue Vincent is the coordinate system for the renderer at negative Z
[11:09] Vincent.Sylvester I fear that it may be something hardcoded and not easily changed in the shader files
[11:10] Gavin.Hird not in the shader, but in the renderer code
[11:11] Andrew Hellershanks: There have been problems in the past when dealing with something where Z is negative. Not too surprising the renderer(?) may not like negative Z values.
[11:12] Andrew Hellershanks: Vincent, which viewer are you working on?
[11:12] Vincent.Sylvester Well the fact the density of the fog changes the deeper you go has me wondering if there maybe a shader responsible for it still
[11:12] Gavin.Hird I would think you'd have to move the renderer coordinate system to have 0 at the -100 level and then translate to word coordinates
[11:13] Gavin.Hird but it gets really wonky when you have to regions adjacent to each other with different negative Z
[11:13] Bill Blight: world already supports -100 but opengl might have issues with negative z
[11:13] Gavin.Hird which is why I say you have to transform it for the render
[11:14] Bill Blight: yeah I was just translating
[11:14] Bill Blight: LOL
[11:14] Vincent.Sylvester Well I did notice that turning Basic Shaders off solves the problem, so the problematic shader must be in that package
[11:15] Gavin.Hird In any case I think I will remove Aurora sim support from the viewer
[11:15] Gavin.Hird does anyone use that or whitehall?? any more?
[11:16] Ubit Umarov: well that is var regions support
[11:16] Ubit Umarov: basicly
[11:16] Gavin.Hird yes, and no
[11:16] Gavin.Hird there is a lot of other cruft
[11:16] Vincent.Sylvester Support for infinite regions I am guessing
[11:16] Gavin.Hird which is set by a special cap
[11:16] Andrew Hellershanks: I know someone who used to use WhteStar, or WhiteCore, whichever it was called. Don't know if they still use it or not.
[11:17] Gavin.Hird Whitehall - I am watching Brexit too much, haha
[11:17] Andrew Hellershanks: :)
[11:18] Gavin.Hird DispatchopenRegionSettings is the name of the cap
[11:18] Gavin.Hird it sets params such as enable_teen_mode
[11:19] Gavin.Hird allow_render_water
[11:19] Vincent.Sylvester I got an unrelated question I want to bark in for readback since I gotta leave early: What is the status of Postgres support? Is it safe to use? Actively maintained?
[11:20] Gavin.Hird As I said I use it and there is nothing wrong with it
[11:20] Gavin.Hird it performs pretty good
[11:20] Ubit Umarov: oops RL ppl
[11:20] Andrew Hellershanks: IIRC, the last time there were changes made to the database related code for MySQL the PostgreSQL code also got changed (eventually).
[11:20] Ubit Umarov: have to leave this meeting soon
[11:21] Ubit Umarov: sorry, cya
[11:21] Andrew Hellershanks: ok, tc Ubit.
[11:21] Ubit Umarov is offline.
[11:21] Gavin.Hird how would you use it Vincent?
[11:22] Vincent.Sylvester Last I checked there seemed to be missing fields or methods and migration resulted in having to manually adjust some parts so I am simply wondering if it is a good idea to migrate or not.
[11:22] Gavin.Hird migrate from what?
[11:22] Vincent.Sylvester mysql
[11:22] Gavin.Hird how much data?
[11:23] Vincent.Sylvester 80gb roughly
[11:23] Andrew Hellershanks: Hm... that is a different type of migration.
[11:23] Gavin.Hird the small tables you can do pretty easy by sql export and import
[11:24] Gavin.Hird for the regions and inventories I'd do export iar / oar and reimport
[11:24] Vincent.Sylvester I have a conversion tool to read in the migrations, question is are they up to date and is postgres supported enough that I won't run into missing methods for certain transactions
[11:24] Andrew Hellershanks: If the creation of the tables when using PostgreSQL is missing some fields you should open a mantis to report what is missing and in what tables.
[11:24] Gavin.Hird it does not really miss any fields, no
[11:25] Gavin.Hird except for one particular function where there is no support at all and that is for the ban list
[11:25] Vincent.Sylvester That is what I was wondering, so that's not good knowing that I have people that use that function
[11:26] Andrew Hellershanks: Gavin, what is missing regarding support for the ban list? The database table, or something in the grid code?
[11:26] Gavin.Hird if you plan on migrating a grid, I'd set it up and do two regions / users
[11:26] Gavin.Hird both
[11:26] Andrew Hellershanks: ok.
[11:27] Andrew Hellershanks: Seems like an odd oversight.
[11:27] Gavin.Hird Melanie does not think anything but mysql exist
[11:27] Gavin.Hird same with Diva
[11:28] Andrew Hellershanks: Vincent, MySQL gets most of the attention. When changes are made that affect database operations and/or tables, the Postgres side should be updated but doesn't always happen right away. There aren't as many people using PostgreSQL compared to the number using MySQL.
[11:29] Vincent.Sylvester Thing is, the option to use it still sits in the configs without warning that it may not have all features, so ideally adding them at some point would be good.
[11:29] Vincent.Sylvester If need be I can probably source someone to do it, but I would first need to know what all is broken.
[11:30] Andrew Hellershanks: True. Someone needs to setup a system with Postgres to test it.
[11:31] Gavin.Hird The other thing that has been an issue is that NPGSQL version 3 only runs on .NET 4.6 or higher, so that was a problem for adding better functionality
[11:31] Gavin.Hird now that dev use 4.6 it can possibly be sorted
[11:31] Gavin.Hird as I said I run my grid on it
[11:31] Vincent.Sylvester Aight I gotta head out, see you guys next week.
[11:32] Gavin.Hird so any assistance
[11:32] Andrew Hellershanks: ok, Vincent. See you next week.
[11:32] Gavin.Hird cheers Vince
[11:32] Leighton.Marjoram tc Vince
[11:33] Andrew Hellershanks: What else do we have to talk about this week?
[11:34] Gavin.Hird there is also an issue in libopenmetaverse for running NPGSQL version 3
[11:35] Andrew Hellershanks: Is NPGSQL some (recent?) fork of PostgreSQL?
[11:36] Gavin.Hird NPGSQL is alibrary that makes Postgres talk to .NET
[11:36] Gavin.Hird a library*
[11:36] Andrew Hellershanks: ok. I looked at PGSQL a long long time ago. Haven't tried using it since.
[11:37] Gavin.Hird and Postgres version 10 and 11 have implemented a whole range of new functionality that isn't implemented in anything but NPGSQL 3 or higher
[11:37] Gavin.Hird such as parallell queries for instanve
[11:37] Gavin.Hird instance
[11:38] Andrew Hellershanks nods
[11:38] Gavin.Hird native data types for JSON and XML
[11:39] Gavin.Hird the upsert command (update and insert)
[11:39] Gavin.Hird native fulltext search
11:43] Andrew Hellershanks: upsert reminds me of a MySQL function that can automatically insert or update a record without the user having to first check if a record exists then use the appropriate command.
[11:44] Gavin.Hird it is used in the general table handler in opensim
[11:45] Gavin.Hird so the PGSQL code tries to write around it as it wasn't supported till version 10
[11:46] Andrew Hellershanks: Anything else for today or do we get to head out early?
[11:47] Andrew Hellershanks: Any time I ask that type of question something comes up and we go to the top of the hour. One of these days we will really end a meeting early. :)
[11:47] Gavin.Hird maybe today is the day!
[11:48] Andrew Hellershanks: Yes, I'm thinking it might be. I won't get to leave early. I'll stick around to see if anyone else shows up late.
[11:49] Andrew Hellershanks: Today is the day. That will do it for today. See you all again next week.
[11:50] Andrew Hellershanks: I'll stick around until a little past the hour in case some people turn up who didn't realize the clocks have changed.

Personal tools
About This Wiki