Chat log from the meeting on 2018-08-07

Summary:
- MySQL.Data.dll not connecting to a socket
- Parcels being shifted after a region restart
- Jenkins stuff
- map search sometimes shows HG regions
- communication of the move to Dev Outreach
- rules about vars of different size next to each other

[10:51] Sheera Khan: I had a little surprise when I tried to start my 0.9.1 and it wouldn't get to my MariaDB :-/
[10:51] Bill Blight: uhhh oh
[10:52] Sheera Khan: the new connector doesn't work with sockets
[10:52] Sheera Khan: only network connections are supported there
[10:52] Bill Blight: sure it does
[10:52] Bill Blight: you just have to make some changes
[10:52] Bill Blight: they are just not configured by default
[10:53] Sheera Khan: so either I had to reconfigure the DB - and honestly I like the concept of UNIX-sockets better in my case
[10:53] Bill Blight: yep
[10:53] Sheera Khan: or I had to copy the old connector from 0.8.x into the new simulator
[10:53] Sheera Khan: it worked ;-)
[10:54] Bill Blight: I always set it up for sockets, but you don't have to change the connector, you just have to configure maria to use sockets
[10:54] Sheera Khan: Maria _is_ using sockets from the ver beginning
[10:54] Sheera Khan: my old 0.8.x runs on that DB
[10:54] Bill Blight: ohhh
[10:54] Sheera Khan: using sockets for sure
[10:54] Bill Blight: I see what you are saying
[10:55] Sheera Khan: I even copied the connection string from my old simulator to the new one
[10:56] Sheera Khan: but the new connector doesn't find the host /var/run/mysql (or whatever the socket is called)
[10:56] Bill Blight: I've never had to change any connectors, but I never leave anything on the DB setup stock, I always configure the sockets first thing on mariaDB there was one middle version there that had sockets disabled by default
[10:56] Sheera Khan: the new connector tries to resolve that path into an IP address ^^
[10:56] Bill Blight: socket = /var/run/mysqld/mysqld.sock
[10:57] Sheera Khan: yepp, that one ..
[10:57] Bill Blight: which new connector are you speaking of , I'm confused
[10:57] Bill Blight: or lost
[10:57] Bill Blight: or both
[10:57] Sheera Khan: I recited it from the top of my head...
[10:58] Sheera Khan: I replaced the MySql.Data.dllfile
[10:58] Gavin.Hird @grid.xmir.org:8002: Did you upgrade to version 8?
[10:59] Bill Blight: strange
[10:59] Sheera Khan: the one coming with 0.9.1 didn't support the socket
[10:59] Bill Blight: I have never had to do that, I always use sockets
[10:59] Sheera Khan: the one from good old 0.8 did
[10:59] Bill Blight: I use the one built into 9
[11:00] Sheera Khan: [opensim@puck150 bin]$ mono -V
Mono JIT compiler version 5.12.0.301 (tarball Fri Jul 20 15:09:05 UTC 2018)
[11:00] Ubit Umarov: ohh yeah hurliman moved and no signed there
[11:00] Bill Blight: Might go put a sign back, till it gets removed again
[11:00] Bill Blight: ?
[11:01] Ubit Umarov: or ill be there telling ppl to come here
[11:02] Sheera Khan: I compiled the simulator on my machine already up on Mono 5.12...
[11:02] dj phil: He I found you :)
[11:03] Bill Blight: that is strange you are having issues with the mysql connector, I never have to replace it
[11:03] Sheera Khan: and the MySQL.Data.dll didn't connect to the socket, it tried to resolve the pathname to an IP address instead
[11:03] Bill Blight: ohhh
[11:03] Gavin.Hird @grid.xmir.org:8002: are you discussing MySQL version 8?
[11:03] Bill Blight: I think the format did change
[11:03] Bill Blight: yes
[11:03] Bill Blight: the format for specifying a socket did change
[11:04] dj phil: It seems there are good beers here refreshing well
[11:04] Gavin.Hird @grid.xmir.org:8002: Someone told me you have to install it with legacy connection support to use it with OpenSim
[11:04] Sheera Khan: mariadb-server-5.5.56
[11:05] Sheera Khan: so no, I'm on an elder release of mariaDB
[11:05] Andrew.Hellershanks @hg.osgrid.org: Sheera, you are using mariaDB and not MySQL?
[11:05] Sheera Khan: yes I use mariaDB
[11:05] Bill Blight: I do too
[11:06] Sheera Khan: and the connector from the old 0.8.2+ does connect without an issue
[11:06] Andrew.Hellershanks @hg.osgrid.org: I have one grid running that uses mariadb.
[11:06] Bill Blight: Loves me some MariaDB
[11:06] Ubit Umarov: ohh i did added some osSound things .. had forgotten that :)
[11:06] Sheera Khan: and replacing the DLL from the 0.9.1 with the old one from 0.8.2 worked :-)
[11:07] Sheera Khan: now OS0.9.1 connects to the socket
[11:07] Ubit Umarov: hmm ?
[11:08] Sheera Khan: I seem to have missed a setting though, my viewer didn't get an URL to upload stuff ^^
[11:08] Andrew.Hellershanks @hg.osgrid.org: I have a grid using mono 4.6.2 and mariadb and I didn't have to do anything special to access the DB's for 0.9.1
[11:09] Bill Blight: we are talking about a socket connection not a standard tcp connection that most use
[11:09] Ubit Umarov: hmm ok
[11:10] Ubit Umarov: that was removed no ?
[11:10] Ubit Umarov: by mysql i mean..
[11:10] Ubit Umarov: ?
[11:10] Bill Blight: no
[11:10] Bill Blight: but the connection format changed
[11:11] Bill Blight flüstert: instead of "Data Source" you use
[11:11] Bill Blight: "server"
[11:11] Bill Blight: if you want a socket connection not a TCP one
[11:12] Ubit Umarov: ( no one at hurliman )
[11:12] Bill Blight: for instance .ConnectionString = "database=myDB;server=/var/lib/mysql/mysql.sock;user id=myUser; pwd=myPass";
[11:12] Ubit Umarov: you should use other pass Bill that does not seem very strong
[11:13] Bill Blight: HAHA
[11:13] Gavin.Hird @grid.xmir.org:8002: he can always use yourPass
[11:13] Sheera Khan: Sorry I was a bit distracted as I just tried it.. Nope, doesn't work for me
[11:14] Sheera Khan: ConnectionString = "Server=/var/lib/mysql/mysql.sock;Database=opensim;User ID=whoever;Password=******;Old Guids=true;"
[11:14] Bill Blight: very strange
[11:14] Bill Blight: ohh
[11:14] Sheera Khan: the result is: Unable to connect to any of the specified MySQL hosts. ---> System.Net.Sockets.SocketException: Could not resolve host '/var/lib/mysql/mysql.sock'
[11:15] Bill Blight: make sure that path is right to the socket on your system
[11:15] Bill Blight: mine is in var/run/yada yada yada socket here, crap
[11:15] Sheera Khan: The path definitly is correct as my old 0.8.2+ runs on exactly those parameters
[11:15] Bill Blight: ok strange
[11:15] Sheera Khan: with the exception of server vs. Data Source
[11:16] Sheera Khan: and when I replace the MySQL.data.dll from 0.9.1 by the one from 0.8.2 it works
[11:18] Bill Blight: well works for me, but I only have one system that uses sockets, the rest connect remotely
[11:18] Ubit Umarov: well thats mysql thingy
[11:19] Sheera Khan: temporary fix is: Copy the old Connector into the new directory ;-)
[11:19] Bill Blight: yeah the mysql thing pretty much off the shelf code,
[11:20] Bill Blight: Ok Ubrokeit , you looked at that mantis about the parcel issues yet?
[11:20] Sheera Khan: There was a mantis about that I think...
[11:21] Gavin.Hird @grid.xmir.org:8002: what Mantis is that?
[11:21] Bill Blight: http://opensimulator.org/mantis/view.php?id=8349
[11:21] Gavin.Hird @grid.xmir.org:8002: I had that one too
[11:21] Gavin.Hird @grid.xmir.org:8002: it screwed up parcesl on over 20 sims
[11:22] Gavin.Hird @grid.xmir.org:8002: I reverted thsi commit
[11:22] Gavin.Hird @grid.xmir.org:8002: http://opensimulator.org/viewgit/?a=commit&p=opensim&h=0731220a9dca5cc2107246c2bd769d002d921239
[11:22] Ubit Umarov: hmm no had not seen it
[11:22] Gavin.Hird @grid.xmir.org:8002: and I had to redo all the parcesl, but it now works as expected
[11:23] Bill Blight: well they come back if you do back to back restarts or clear all caches after a shutdown
[11:23] Ubit Umarov: err and it does get back to correct position ??
[11:23] Gavin.Hird @grid.xmir.org:8002: parcels got skewed and many split in two. Some also seemed to be rotated in veired ways
[11:23] Bill Blight: they come back correctly
[11:24] Bill Blight: because they are still in the database correctly
[11:24] Bill Blight: it is only local to the sim they are messed up
[11:24] Bill Blight: even a land show , shows them correctly
[11:24] Bill Blight: but by the time they are rendered they are screwed up
[11:26] Gavin.Hird @grid.xmir.org:8002: reverting that bit mask commit fixed it for me
[11:26] Ubit Umarov: hmm ill check it
[11:30] Ubit Umarov: does not look like it.. but well
[11:33] Andrew.Hellershanks @hg.osgrid.org: I'm looking at the change. The problem might be in the change to the calculation of bit. I don't think "== bitmask" is needed there and might be causing the issue.
[11:34] Ubit Umarov: ok think again
[11:34] Ubit Umarov: hihiih
[11:36] Andrew.Hellershanks @hg.osgrid.org: It depends on how == is handled. Before you had 0 or 1. Now you have 0 or not zero. That might be 0 or 1 it might not be. Depends on how C# defines the possible results of ==.
[11:37] Ubit Umarov: hmm ??
[11:39] Andrew.Hellershanks @hg.osgrid.org: == returns either true or false. False is probably 0 but how is true defined? Is it 1 or some other non-zero value?
[11:40] Ubit Umarov: oops lol
[11:40] Andrew.Hellershanks @hg.osgrid.org: I don't think you need the "== bitmask" part of the expression to do what you want to do there.
[11:41] Ubit Umarov: im exactly making clear int to boll convertion
[11:41] Ubit Umarov: but think i see the bug.. so silly lol
[11:43] Andrew.Hellershanks @hg.osgrid.org: ok, good. You'll have it fixed in no time. :)
[11:43] Ubit Umarov: hope so
[11:43] Ubit Umarov: but is not what you where telling that is ok
[11:44] Gavin.Hird @grid.xmir.org:8002: I'm gonna let someone else test it first this time, hehe
[11:44] Ubit Umarov: for c# at least :)
[11:44] Ubit Umarov: lol gavin.Hird
[11:44] Ubit Umarov: well it should not break the parcels.. much
[11:45] Ubit Umarov: lol
[11:46] Bill Blight: somebody needs to get the core's jenkins a registered nickname so it can post to the channel again
[11:46] Ubit Umarov: no need to pray for end of times ( ref to current song )
[11:46] Bill Blight: and mantis
[11:47] Ubit Umarov: err they are out ?
[11:47] Bill Blight: yep
[11:47] Ubit Umarov: good
[11:47] Bill Blight: +r is keeping them out
[11:47] Ubit Umarov: i dont like them anyways
[11:47] Bill Blight: but keeping the spam bots out too
[11:47] Ubit Umarov: jenkins keeps tellin gmy code id bad
[11:47] Andrew Hellershanks: Will Jenkins be able to issue the password to IRC?
[11:47] Bill Blight: Jenkins is pretty smart
[11:47] Ubit Umarov: i don't like him
[11:48] Bill Blight: mine does
[11:48] Bill Blight: LOL
[11:48] Bill Blight: they can auth
[11:49] Andrew Hellershanks: ok.
[11:49] Andrew Hellershanks: Should be no problem to get Jenkins to report on code compiles once again in IRC.
[11:51] Andrew Hellershanks: Bill, you mentioned to me in IRC a while ago that search results are now including HG related results. Are those HG results from the grid the person came from?
[11:52] Bill Blight: seems to be when a person HG's out
[11:52] Bill Blight: and it creates a temp HG link to the system
[11:52] Andrew Hellershanks: I'm referring to searches for regions using the Map.
[11:52] Bill Blight: they show up in the map search
[11:52] Bill Blight: but only on the patched viewer
[11:53] Andrew Hellershanks: I'm still not clear what you are saying.
[11:53] Andrew Hellershanks: It creates a temp HG link to which system?
[11:54] Ubit Umarov: gavin ready to test?
[11:54] Bill Blight: Ok .. When a person HG's out there is a temporary link created to your home grid/standalone in a random high map position, those normally don't show up in a standard map search, they are NOW showing up
[11:54] Gavin.Hird @grid.xmir.org:8002: hehe
[11:54] Gavin.Hird @grid.xmir.org:8002: not right now
[11:54] Ubit Umarov: well ill test it first :)
[11:54] Gavin.Hird @grid.xmir.org:8002: I have enought testing the BOOL conversions in the viewer
[11:55] Gavin.Hird @grid.xmir.org:8002: only done about 2400 of the 13800 so far
[11:55] Ubit Umarov: nahh its not bool conversion :)
[11:56] Ubit Umarov: its a bit more silly lol
[11:56] Gavin.Hird @grid.xmir.org:8002: :-)
[11:57] Bill Blight: So as an example Andrew, if I have a region named Welcome Center and someone HG's out to "yada yada center" and I search for center, I see both those regions return in the map search
[11:58] Andrew Hellershanks: Bill, the code is checking if HG search is allowed and whether the search string contains a . before it goes in to the section of search that appears to be HG related.
[11:58] Ubit Umarov: was George
[11:58] Andrew Hellershanks: I'm not sure what that is all about.
[11:58] Andrew Hellershanks: Jim, The meeting is open to everyone.
[11:59] Bill Blight: open to everyone but about over for the day, unless we keep rambling
[11:59] Jim Jackson: if its going to be held here from now on, should probably let people know?
[11:59] Sheera.Khan @hypergrid.org:8002: the logs are posted on: http://opensimulator.org/wiki/Office_hours#Tuesday_-_Office_Hour
[12:00] Bill Blight: Good point Jim
[12:00] Jim Jackson: yes the link says its held at hurliman
[12:00] Sheera.Khan @hypergrid.org:8002: This meeting is scheduled for 1100 PST, 1400 EST, 1900 UTC, 2000 CET on Hurliman Plaza at OSGrid (osgrid.org). http://hg.osgrid.org:80:Dev Outreach
[12:00] Andrew Hellershanks: Someone left a big arrow pointing this direction at the other meeting location. An actual text based sign might help clarify the latest move.
[12:00] Sheera.Khan @hypergrid.org:8002: I just looked it up...
[12:00] Bill Blight: changed the address just not the name
[12:00] Bill Blight: name still says hurliman
[12:00] Bill Blight: lol
[12:01] Andrew Hellershanks: Oh. The web page. :)
[12:01] Ubit Umarov: you forget the sign was returned :)
[12:01] Bill Blight: Yeah the signs got returned when Hurliman got moved
[12:01] Sheera.Khan @hypergrid.org:8002: the web page is updated
[12:01] Ubit Umarov: ty
[12:01] Andrew Hellershanks: Sheera, ty
[12:02] Ubit Umarov: so jim.. any question for andrew  ?
[12:02] Sheera.Khan @hypergrid.org:8002: well, I did it today, though... so maybe it was in the cache still
[12:02] Andrew Hellershanks: I was thinking of a sign in the region next door. One at Hurliman would also have been useful.
[12:02] Bill Blight: nextdoor was hurliman
[12:02] Ubit Umarov: no region next door andrew .. was moved
[12:02] Bill Blight: but it got moved
[12:03] Gavin.Hird @grid.xmir.org:8002: I supposet the web page is the most important place to keep current
[12:03] Ubit Umarov: guess lbsa still points to wright.. didn't check
[12:03] Andrew Hellershanks: We've moved several times. I'm losing track. Right. It was Projet Secret then Hurliman, then here.
[12:03] Selby.Evans @grid.kitely.com:8002: There is a box at Hurliman that says Dev outreach
[12:03] Jim Jackson: yeah we had to move it because this var was lagging it and we had a big meeting on sunday
[12:04] Bill Blight: lol
[12:04] Ubit Umarov: oh the box is still there ?
[12:04] Sheera.Khan @hypergrid.org:8002: it severely lagged before ^^
[12:04] Andrew Hellershanks: I was surprised to see a regular size region up against a var. I didn't think that was possible without causing problems.
[12:04] Jim Jackson: it did cause problems
[12:04] Selby.Evans @grid.kitely.com:8002: a box is there, Ubit
[12:05] Bill Blight: that region lagged before this was ever here, but ok
[12:05] Ubit Umarov: hmm don't make bill remember certain grid with that mixed sized regions
[12:05] Sheera.Khan @hypergrid.org:8002: come in and take a seat, if you like, Jim
[12:05] Bill Blight: uggg
[12:05] Bill Blight: that was a night mare Ubit
[12:05] Ubit Umarov: thx Selby
[12:05] Andrew Hellershanks: Is it considered safe to have regions of differing sizes up against each other or is that still going to lead to problems?
[12:06] Ubit Umarov: i made a mantis on that
[12:06] Bill Blight: as long as there is only ONE per side, it is no diff than singles being next to each other
[12:07] Ubit Umarov: yeap things seem to work, if one per side only
[12:07] Gavin.Hird @grid.xmir.org:8002: so same size vars next to each other is good too?
[12:07] Andrew Hellershanks: ok. I can see some people forgetting that little detail.
[12:07] Bill Blight: the , "This var caused lag", is Bull, sorry just is
[12:08] Ubit Umarov: yes gavin.Hird should be good
[12:08] Andrew Hellershanks: AFAIK, if regions of the same size are up against each other all will be well. It is when you try to mix sizes you run in to problems.
[12:08] Bill Blight: yeah var next to var of the same size works fine, var's of different size, the ONE PER SIDE, rule applies
[12:08] Ubit Umarov: hmm think a grid is going to do its "mainland" all 4x4 regions
[12:09] Bill Blight: when I bring my mainland up either going to be 2x2 or 4x4
[12:09] Ubit Umarov: be carefull if you violate that rule, things still seem to work..
[12:10] Bill Blight: until they don't
[12:10] Andrew Hellershanks nods
[12:10] Ubit Umarov: but on tps Kabumm
[12:10] Ubit Umarov: or crossings out of range
[12:10] Andrew Hellershanks: Last time I remember discussions about different sizes of regions up against each other the issue was more that viewers might not handle it well.
[12:10] Gavin.Hird @grid.xmir.org:8002: right Andrew
[12:10] Bill Blight: yes it is viewer side issue mostly, it seems
[12:11] Gavin.Hird @grid.xmir.org:8002: the viewer code leaves a lot to be desired
[12:11] Ubit Umarov: viewers suport for var was never finished
[12:11] Andrew Hellershanks: Yet another reason to get an OS specific viewer some day. :)
[12:11] Ubit Umarov: on some minimap is bad
[12:11] Ubit Umarov: and local tps, etc
[12:13] Andrew Hellershanks: Bill, I'll do an HG TP to some other grid and see what sort of search results I get. It will let me see if HG results can be filtered out.
[12:13] Bill Blight: it is not that big of a deal but it is annoying
[12:14] Andrew Hellershanks: It has me wondering why that code has all the HG stuff in that one routine.
[12:14] Andrew Hellershanks: I'll know more once I do some more tests.
[12:14] Bill Blight: but it is interesting , I can see what places my users are HG'ing to with random searches ... LOL
[12:14] Andrew Hellershanks: :)
[12:15] Andrew Hellershanks: That makes it a feature, not a bug. ;)
[12:15] Bill Blight: well no such thing as bugs in Opensimulator, they are all "Support Challenged Features"
[12:15] Andrew Hellershanks: Bill, thanks for mentioning it. I don't do much involving HG and I don't have an HG enabled test setup on my machine.
[12:16] Andrew Hellershanks: We have passed the one hour mark once again. Any last minute items before we head off in different directions?