Chat log from the meeting on 2018-05-15

From OpenSimulator

Jump to: navigation, search

New Mantis entry about failing TPs upsetting a region in the works
Support for 0.8 grids to be dropped
Purging of map tiles
How to compile branch httptests correctly
minor announcement: the addon OpenSimSearch module requires a database migration
building for release vs debug

[11:07] Kayaker Magic: I'm working on a new Mantis, but don't have the how-to-reproduce quite nailed down yet.
[11:08] Kayaker Magic: The bug is that when osTeleportAgent fails to work, it sometimes leaves the region set as a "hyperlink" and then you can no longer teleport to the region you ran the script in.
[11:09] Ubit Umarov: well failed tps can also do that :(
[11:09] Bill.Blight Yeah it is not just with a script, it can happen with regular TP's as well
[11:10] Ubit Umarov: well set as hyperlink.. not sure what that means
[11:10] Bill.Blight If I have a hard TP fail that avie can no longer tp to that region, but other avies seem able to get there ..
[11:10] Bill.Blight eventually it clears itself though
[11:10] Kayaker Magic: There is something called a "hyperlink", a region that looks like it is in one region but is actually in another.
[11:11] Ubit Umarov: not really like that
[11:11] Ubit Umarov: and not on tps on same grid
[11:11] Bill.Blight a hyperlink region is a region that is temp linked for HG TP's
[11:11] Kayaker Magic: I don't use hyperlinks, but found one of my regions accidentally set as one, right after a failed TP
[11:11] Ubit Umarov: and not on hg either
[11:12] Ubit Umarov: the only thing is that the target is placed on a virtual map location
[11:12] Kayaker Magic: It is the source region that gets corrupted on a failed TP
[11:13] Ubit Umarov: but that is the target region.. not the tp source region
[11:13] Bill.Blight I have seen an issue like that a long time ago, when DNS was hosed and even though I was on the grid, it thought I was from the HG, so everytime I tried to jump to a region it tried to hyperlink it, but I have not seen that for a LONG time
[11:13] Ubit Umarov: but guess that is the avatar, not region
[11:14] Ubit Umarov: and dns issues do cause a lot of issues
[11:15] Ubit Umarov: or problems :)
[11:15] Ubit Umarov: so mk sure machine dns is fine, and using a good dns server
[11:15] Ubit Umarov: google public dns was not such one
[11:16] Ubit Umarov: specially for dynamic dns
[11:18] Kayaker Magic: On another topic: The server I set up httptest on is going away in a few days, so I'm going to have to move my test.
[11:18] Kayaker Magic: Can I run an httptest region on OSGrid?
[11:18] Ubit Umarov: yes
[11:18] Kayaker Magic: great!
[11:20] Ubit Umarov: im just dropping some things for 0.8 hmm or older grids
[11:20] Kayaker Magic: ?
[11:20] Ubit Umarov: 0.9 could run on 0.8 grids
[11:21] Ubit Umarov: i may start dropping that
[11:21] Kayaker Magic: I didn't know that!
[11:22] Ubit Umarov: but for now its var regions case basicly.. 0.8 grids had no clue about those
[11:22] Sheera Khan: Wus? 0.8 did know about var regions
[11:23] Ubit Umarov: no grid code did not
[11:23] Ubit Umarov: at least part of it
[11:23] George Equus: Another topic still: Where / how are non existing regions purged from World Map? Eg do a map search for Belfast and there are lots of them still marked with tiles but not existing, 12 of them at the top.
[11:24] Ubit Umarov: to find a region by postion was a ugly test done region side, testing all 256m spots
[11:24] Ubit Umarov: they are not, you need to do it by hand
[11:25] Kayaker Magic: how by hand? talking to the DB with MySQL?
[11:25] Sheera Khan: the reason for the annual spring-cleanup ^^
[11:25] George Equus: Something Dan need to do then?
[11:25] Ubit Umarov: map are disk files
[11:25] Bill.Blight clearing out the mapitiles
[11:25] Ubit Umarov: yeap dan needs to do it
[11:25] Bill.Blight bin/maptiles
[11:25] Bill.Blight default ^^
[11:26] George Equus: so. to make it perfectly clear, this is nothing a user can do
[11:26] Ubit Umarov: yeap
[11:26] George Equus: OK. Thanks
[11:27] Ubit Umarov: my test regions here now are inside a 2km oen i did test once on map :)
[11:28] Ubit Umarov: see ubittest14
[11:28] Ubit Umarov: on map i mean
[11:29] Ubit Umarov: well seems i did test more than one 1km regions :)
[11:30] Ubit Umarov: and that image is now with new warp code
[11:30] Ubit Umarov: well not anymore.. i changed it more :)
[11:31] Ubit Umarov: those brown trees are wrong
[11:31] Ubit Umarov: and the large box was missing a face lol
[11:32] Ubit Umarov: and no one cares about what im saying LOL
[11:33] Ubit Umarov: but i had warp3d messed up
[11:34] Ubit Umarov: i could move a region from my test grid here to show you.. but 2 lazy to do it now :p
[11:34] Ubit Umarov: im changing map
[11:34] Ubit Umarov: warp3d
[11:34] Ubit Umarov: making it a lot slower lol
[11:34] Kayaker Magic: May not actually ever start....
[11:34] Bill.Blight Let me translate that for Ubit, "I'm breaking the map"
[11:35] Leighton.Marjoram good to know Kayaker :)
[11:35] mayLee Jun: bbew nice to fix it instead
[11:35] Ubit Umarov: well not a priority.. but someone asked me about map.. and well i saw it again and started changing it :p
[11:36] Kayaker Magic: What is that saying: "A good engineer can fix something whether it is broken or not!"
[11:37] Ubit Umarov: btw that warp3d.dll must be OUR version
[11:37] Ubit Umarov: ) and not the one on master either )
[11:37] Ubit Umarov: i may rename it to avoid confusion
[11:38] Ubit Umarov: bc thats no longer like the original .net port one
[11:38] Ubit Umarov: ie 0.01% of it is not :p
[11:38] Ubit Umarov: but the good 0,01% ;)
[11:39] Ubit Umarov: but its now slower and eats more peek memory
[11:39] Ubit Umarov: and again, no one cares LOL
[11:40] Sheera Khan: would you mind mentioning the download link for httptests again? So readers of the log get the idea where to get the cool stuff ;-)
[11:40] Ubit Umarov: its our git
[11:40] Ubit Umarov: clone it as usal
[11:40] Ubit Umarov: the do git checkout httptests
[11:40] Ubit Umarov: and you are on it
[11:41] Ubit Umarov: it is also on gitview
[11:41] Ubit Umarov: just a bit confusing to navigate to it
[11:42] Ubit Umarov: gitview is per commit
[11:43] Ubit Umarov: lets see-- start at
[11:44] Ubit Umarov: then on heads select httptests
[11:44] Ubit Umarov: and there should be a zip per commit
[11:44] Ubit Umarov: or tgz
[11:45] Ubit Umarov: and of course says 0.9.0 and its not lol
[11:45] Sheera Khan: did I get it right: git clone git:// - git checkout httptests;  ?
[11:45] Ubit Umarov: well mastes also tells that
[11:46] Ubit Umarov: and you should be on that clone folder of course
[11:47] Ubit Umarov: so guess, clone, cd .. , git checkout :)
[11:48] Ubit Umarov: AND
[11:48] Ubit Umarov: do not use old inis..
[11:48] Sheera Khan: I did the following: git clone git://; cd opensim; git checkout httptests
[11:48] Ubit Umarov: edit its INIs
[11:48] Bill.Blight umm
[11:48] Kayaker Magic: No changes to Region.ini files I hope?
[11:48] Bill.Blight not quite
[11:49] Ubit Umarov: you can at most copy bin/assetcache from a 0.9 region to speedup first startup
[11:49] tx.0h you after cd you do a git checkout httptest
[11:49] Bill.Blight git clone git://; cd opensim;git checkout httptests
[11:49] tx.0h s
[11:49] tx.0h what bill said
[11:49] Bill.Blight oops
[11:50] Bill.Blight I did put an extra s on there
[11:50] Sheera Khan: ok, now I've got it :-)
[11:50] tx.0h no, but you misspelled http
[11:50] tx.0h no, all fine
[11:50] Ubit Umarov: ./
[11:50] tx.0h all right.
[11:51] Ubit Umarov: msbuild /p:Configuration=Release
[11:51] Ubit Umarov: fix ini files and regions.ini
[11:51] tx.0h i'm back on master now. and the sim rezzes under 3 minutes, ubit. i'm ready to test whatever you like.
[11:51] Kayaker Magic: MSBUILD?
[11:51] Ubit Umarov: cd bin
[11:52] Ubit Umarov: ./
[11:52] Kayaker Magic: does it still work with xbuild?
[11:52] Ubit Umarov: yes kayaker msbuild is the official mono5 compiler
[11:52] tx.0h it's better to use msbuild, i've learned
[11:52] Ubit Umarov: xbuild will work
[11:53] Kayaker Magic: Xbuild on linux keeps telling me it will go away and I should use msbuild,
[11:53] Ubit Umarov: it will say:
[11:53] Kayaker Magic: but msbuild does not work on the pre-httptest version.
[11:53] Bill.Blight msbuild does seem to work better, but my thoughts for not using xbuild, may have only applied to a certain build of mono
[11:54] Ubit Umarov: >>>> xbuild tool is deprecated and will be removed in future updates, use msbuild instead <<<<
[11:54] tx.0h nods.
[11:55] Ubit Umarov: kay you also did not know that mono is now a ms product ?
[11:55] Sheera.Khan yaaaayyy, it works - 182 warnings ;-)
[11:55] Bill.Blight eek
[11:55] Ubit Umarov: WHAT WARNINGS ???
[11:55] Bill.Blight I only get 3 warnings
[11:55] Ubit Umarov: it has 3 warnings !!
[11:56] Ubit Umarov: 182 is master i guess
[11:56] tx.0h yes, i also had 3 warnings
[11:56] Kayaker Magic: Yes, I have heard about the migration of mono to MS sources.
[11:56] Ubit Umarov: well i get 3 with msbuild
[11:57] Andrew Hellershanks: I just had one very minor announcement.
[11:57] Andrew Hellershanks: For those of you using the addon OpenSimSearch module there was a database migration this week that affects the popularplaces table.
[11:57] Andrew Hellershanks: It was storing the maturity information as a number instead of as a string.
[11:58] Sheera.Khan I did this:
[11:58] Sheera.Khan git clone git://
cd opensim/
git checkout httptests
[11:58] Bill.Blight msbuild should really be used with httptest
[11:58] Sheera.Khan and got:
[11:58] Sheera.Khan . . .
Shared/Tests/LSL_ApiUserTests.cs(97,27): warning CS0219: The variable `sp' is assigned but its value is never used
XEngine/Tests/XEngineCrossingTests.cs(189,35): warning CS0219: The variable `scriptItemSceneB' is assigned but its value is never used

182 Warning(s)
0 Error(s)
[11:58] Ubit Umarov: well its 3 warnings with msbuild
[11:59] Ubit Umarov: nopes
[11:59] Ubit Umarov: that must be master
[11:59] Bill.Blight ohh
[11:59] Bill.Blight there is no "s" on the end of httptest
[12:00] Bill.Blight nevermind
[12:00] Bill.Blight there is
[12:00] Ubit Umarov: sheera type git branch
[12:00] Bill.Blight ignore me
[12:00] Ubit Umarov: one has a *
[12:00] Sheera.Khan git checkout httptest -> error: pathspec 'httptest' did not match any file(s) known to git.
[12:01] Sheera.Khan git checkout httptests -> Bereits auf 'httptests' Ihr Branch ist auf dem selben Stand wie 'origin/httptests'.
[12:01] Ubit Umarov: ld@linbox:~/opensim/git/opensim$ git branch
* httptests
[12:01] tx.0h this means you are on httptests
[12:01] Ubit Umarov: its git checkout httptests
[12:01] Sheera.Khan schleppi:~/OpenSim/opensim/opensim> git branch
* httptests
[12:02] tx.0h schleppi :-)
[12:02] Ubit Umarov: ok so you are on it
[12:02] Ubit Umarov: odd
[12:02] Andrew Hellershanks: Hm... I just tried building httptests and got 5 errors.
[12:02] tx.0h make clean
[12:02] Ubit Umarov: errors ??
[12:03] Andrew Hellershanks: Ah. They are in addon modules. Something must have changed in the core code.
[12:03] Ubit Umarov: of course it did :p
[12:03] Andrew Hellershanks: Something to do with versioning.
[12:03] Ubit Umarov: ( most on master actually )
[12:03] Ubit Umarov: you did clean?
[12:04] Ubit Umarov: modules need to be compiled for .net4.6 also
[12:04] tx.0h sheera: try 'make clean' then ./
[12:04] Ubit Umarov: make clean? we do not have makefile do we ?
[12:05] tx.0h yes, should be there
[12:05] Ubit Umarov: no there should not be
[12:05] Sheera.Khan I'm just starting in a complete fresh run ...
[12:05] tx.0h my first try was make, which triggered nant
[12:05] tx.0h ah, ok. then you don't need make clean
[12:06] Ubit Umarov: ohh there is
[12:06] tx.0h but ?
[12:06] Ubit Umarov: i should delete it LOL
[12:07] Ubit Umarov: yeap that is wrong now
[12:07] tx.0h bye
[12:08] Sheera.Khan no run anymore?
[12:08] Ubit Umarov: well it does run prebuild
[12:08] Sheera.Khan drats, I did that just some seconds ago
[12:08] tx.0h it?
[12:08] TG.Lucan can use ./ clean
[12:08] Ubit Umarov: Sherea do ./
[12:08] Sheera.Khan the
[12:09] Sheera.Khan I just did
[12:09] TG.Lucan then repeat ./
[12:09] tx.0h and now msbuild or xbuild
[12:09] Sheera.Khan thenI started xbuild as I don't have msbuild
[12:09] tx.0h me neither
[12:09] Sheera.Khan and that's running ATM
[12:10] tx.0h m$ tries to get you a sdk if you like to get msbuild
[12:10] Ubit Umarov: well i don't compile debug version on linux
[12:10] Sheera.Khan and I get quite some warnings like this: The variable `data1' is assigned but its value is never used
[12:11] Ubit Umarov: on tests ?
[12:11] Sheera.Khan as you saw it
[12:11] Sheera.Khan I use mono 4.6.1
[12:11] Ubit Umarov: hmm what is the xbuild comand to set release ?
[12:11] tx.0h ohoh
[12:11] TG.Lucan can use xbuild /p:Configuration=Release
[12:11] Sheera.Khan Shared/Tests/LSL_ApiUserTests.cs(97,27): warning CS0219: The variable `sp' is assigned but its value is never used
XEngine/Tests/XEngineCrossingTests.cs(189,35): warning CS0219: The variable `scriptItemSceneB' is assigned but its value is never used

182 Warning(s)
0 Error(s)
[12:12] Bill.Blight install mono-complete, you will get msbuild
[12:12] Ubit Umarov: grrrr you should use mono 5 on httptests
[12:12] Bill.Blight ohh
[12:12] Bill.Blight yeah httptest requires mono 5x
[12:12] Bill.Blight to build right
[12:12] TG.Lucan Mono 5 has xbuild 14
[12:12] Bill.Blight mono-complete has msbuild
[12:12] Sheera.Khan ah, .net 4.6.1 would be ok, but not mono 4.6.1
[12:13] Sheera.Khan my bad, sorry
[12:13] Ubit Umarov: ok i did not cleaned the tests warnings
[12:13] tx.0h yes, xbuild 14, thats what i have too
[12:13] Ubit Umarov: i never compile debug on linux, its useless for me :)
[12:14] Ubit Umarov: and there is a performance impact now
[12:14] TG.Lucan I run debug on the full grid, on a laptop
[12:15] tx.0h hmmm....
[12:15] Ubit Umarov: ok sheera.Khan
[12:15] Ubit Umarov: now fix the damm ini files LOL
[12:15] Ubit Umarov: and region.ini
[12:15] Ubit Umarov: then cd bin
[12:15] TG.Lucan held up ok with safari
[12:15] Ubit Umarov: ./
[12:15] Sheera.Khan but the workflow would be ok now :-) I just have to get me a mono-complete 5.x
[12:16] tx.0h just..
[12:16] tx.0h what linux are you on?
[12:16] Ubit Umarov: uff one of this days i need to clean also the tests warnings
[12:17] Sheera.Khan atm I'm on an old openSuse 42.2
[12:17] Andrew Hellershanks: hm... the .csproj files for the addon modules need to be updated in order to allow them to compile with mono 5 in the httptests branch. I'll work on that later.
[12:17] Sheera.Khan but I'll try it tomorrow on a more recent machine
[12:17] Ubit Umarov: mono 5 as also a improved JIT
[12:17] Ubit Umarov: it does a git better optimization
[12:18] Ubit Umarov: but only on release mode
[12:18] TG.Lucan latest mon has switched gc to coop
[12:18] TG.Lucan mono
[12:18] tx.0h hmmm.. i hope there is mono-5 for opensusi
[12:18] Sheera.Khan how do I set the release mode?
[12:18] Ubit Umarov: its a optiom
[12:19] Sheera.Khan @tx: usually there is and if not I've gotto compile it myself^^
[12:19] tx.0h :-)
[12:19] Ubit Umarov: msbuild /p:Configuration=Release on msbuild
[12:19] tx.0h thats what i did too
[12:19] TG.Lucan xbuild /p:Configuration=Release /t:rebuild
[12:19] Ubit Umarov: don't remember on x
[12:19] Ubit Umarov: ahh try that now :)
[12:20] Ubit Umarov: same option nice
[12:20] Andrew Hellershanks suspects that information isn't in the wiki.
[12:20] Ubit Umarov: ( well xbuild was a clone of msbuild i guess )
[12:20] tx.0h but until we need to find out whats wrong with httptests on arm cpu
[12:21] Ubit Umarov: on your arm cpu, you mean :p
[12:21] tx.0h i hope i can help ubit on that or other ways
[12:21] Andrew Hellershanks: tx.0h: It won't compile or it won't run?
[12:21] tx.0h on my thinkerboard, yes
[12:21] Ubit Umarov: ill ask neb to try on his rasp
[12:21] tx.0h yeah!
[12:21] Andrew Hellershanks: tx.0h, yes to which part?
[12:21] tx.0h good idea.
[12:22] Ubit Umarov: i think he already is running httptests
[12:22] Andrew Hellershanks: I have Pi's including a Pi 3.
[12:22] Ubit Umarov: btw you can talk with him tx
[12:22] tx.0h andrew, with a clean cache it rezzes one hour
[12:22] Ubit Umarov: he also has experience compiling mono etc
[12:23] tx.0h i tried now on master, with mono 5, and it rezzes in about 3 minutes
[12:23] Bill.Blight I'm willing to bet if httptests ran that bad for Neb, ubit would have heard about it by now
[12:23] tx.0h but what is his time on irc?
[12:23] Sheera.Khan release compiles in my mono 4.6.1 too, but I'm not sure it will run
[12:24] Ubit Umarov: how man warnings ?
[12:24] Bill.Blight it might run but would be missing all the advantages of the newer .NET
[12:24] Sheera.Khan and the amount of warnings stays at 182
[12:24] Ubit Umarov: cant be
[12:24] Ubit Umarov: grrr
[12:25] Bill.Blight compiling something targeted to 4.6 on lower should cause a lot of errors .. LOL
[12:25] TG.Lucan Sheera.. do #mono -V (# is prompt0
[12:25] Sheera.Khan LSL_Api.cs(17688,17): warning CS0162: Unreachable code detected
LSL_Api.cs(14783,37): warning CS0219: The variable `data' is assigned but its value is never used
OSSL_Api.cs(3230,21): warning CS0472: The result of comparing value type `OpenMetaverse.UUID' with null is always `false'
[12:25] Bill.Blight she has 4.6, we determined that
[12:26] Sheera.Khan Mono JIT compiler version 4.6.1 (Stable Thu Oct 20 14:33:14 UTC 2016)
Copyright (C) 2002-2014 Novell, Inc, Xamarin Inc and Contributors.
TLS: __thread
SIGSEGV: altstack
Notifications: epoll
Architecture: amd64
Disabled: none
Misc: softdebug
LLVM: supported, not enabled.
GC: sgen
[12:26] tx.0h good old novell stuff
[12:26] Sheera.Khan I can't update this machine right now..
[12:26] TG.Lucan will that work with dotnet 4..6.1?
[12:27] Bill.Blight mono 4.6.x is not .NET 4.6 compliant
[12:27] Sheera.Khan jepp, I got that by now
[12:27] Bill.Blight 4.6 compatibility was added in 5x
[12:28] Bill.Blight that was at TG to answer his question
[12:28] TG.Lucan seems to be compiling something, odd
[12:28] Ubit Umarov: odd that data warning is valid
[12:29] tx.0h i guess, i will make a new compile tomorrow.
[12:29] Ubit Umarov: bc diferent warnings ignore settings btw xbuild and msbuild
[12:29] tx.0h it can't be, that it takes that long.
[12:29] Ubit Umarov: i guess..
[12:30] tx.0h on master it used full bandwidth to load everything
[12:30] tx.0h and on httptests it was so slow. can't be
[12:31] TG.Lucan btw, I see 2 warnings with the latest httptests
[12:31] tx.0h 2 or 3, just a few.
[12:31] TG.Lucan Scripting/Minimodule/MRMModule.cs(200,43): warning CS0612: 'SecurityManager.PolicyHierarchy()' is obsolete
[12:31] TG.Lucan lsl.lexer.cs(19729,7): warning CS0162: Unreachable code detected
[12:31] Sheera.Khan I could give you the output of the xbuild run... but I guess it's better to first redo that on a mono5 system
[12:32] Ubit Umarov: lsl.lexer.cs(19729,7): warning CS0162: Unreachable code detected is by design
[12:32] TG.Lucan kk
[12:32] tx.0h nods
[12:33] TG.Lucan To be complete the region I arrived from Version: OpenSim Snail Dev f872199 (SIMULATION/0.3 - SIMULATION/0.6)
[12:34] Ubit Umarov: OpenSim Snail Dev is httptests
[12:34] tx.0h ubit, bill, thanks for your passion
[12:34] Sheera.Khan compiling takes me real 2m40.226s
[12:35] Andrew Hellershanks: I'll have a look and see if I can figure out what changes are required in .csproj files to allow the addon modules to build.
[12:35] Sheera.Khan I'll report back when I updated the mono system
[12:35] Ubit Umarov: update their prebuild.xml
[12:38] Ubit Umarov: vs2017 also ignores those warnings
[12:39] TG.Lucan I have seen that, run an add on region from win 10 desktop
[12:39] Ubit Umarov: but warning CS0219: The variable `data' is assigned but its value is never used is valid
[12:40] Ubit Umarov: LSL_Api.cs(17688,17): warning CS0162: Unreachable code detected does not make much sense
[12:41] Sheera.Khan well Ubit, I guess it's better waiting for me toupdate the mono toolchain to 5.x
[12:41] Sheera.Khan is there any 5 ish mono to avoid?
[12:41] TG.Lucan stay with even number subs
[12:42] Ubit Umarov: im on 5.12 now
[12:42] TG.Lucan odd numbers, development
[12:42] TG.Lucan I compile mono daily for other projects, now on 5.15
[12:43] Bill.Blight the only major reported issues were with 5.10

Personal tools
About This Wiki