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?