Chat log from the meeting on 2014-12-30
From OpenSimulator
(Difference between revisions)
m |
|||
Line 1: | Line 1: | ||
− | <pre | + | <pre> |
− | + | ||
− | + | ||
− | + | ||
− | + | ||
[10:59] Mic Bowman: is diva planning to be here today? | [10:59] Mic Bowman: is diva planning to be here today? | ||
[10:59] Seth.Nygard @refugegrid.com:8002: Hello everyone | [10:59] Seth.Nygard @refugegrid.com:8002: Hello everyone | ||
Line 496: | Line 492: | ||
[11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ++ | [11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ++ | ||
</pre> | </pre> | ||
+ | |||
+ | |||
+ | |||
+ | [[Category:Office Hour Logs]] |
Latest revision as of 15:42, 25 October 2015
[10:59] Mic Bowman: is diva planning to be here today? [10:59] Seth.Nygard @refugegrid.com:8002: Hello everyone [10:59] Nebadon Izumi: doubt it Mic [11:00] Shez.Oyen @mirage.duckdns.org:9038: Hi Seth :) [11:00] Nebadon Izumi: i could try to ping her [11:00] Mic Bowman: it would be good to hear about the changes for mono [11:00] Mic Bowman: not necessary [11:00] Mic Bowman: i'll find out from her later... i think i know the changes my modules need to make [11:00] Aine.Caoimhe @refugegrid.com:8002: hehe...this will be fun.....more than half of you show as "loading..." for me [11:00] BlueWall.Slade @gateway.bluewallvirtual.net:8002: I think she hasn't commit her work yet? [11:00] Robert.Adams @home.livingroomcam.us:14559: if you have a module, time to freshen it up is here [11:00] Robert.Adams @home.livingroomcam.us:14559: I'm totally loaded for everyone [11:01] Nebadon Izumi: well I pinged Diva [11:01] Nebadon Izumi: no response yet [11:01] Aine.Caoimhe @refugegrid.com:8002: [11:00] Loading...: I'm totally loaded for everyone <-- like that [11:01] Nebadon Izumi: she may show up if she sees it [11:01] Mic Bowman: would just like to start migrating my modules [11:01] Nebadon Izumi: she responded [11:01] Nebadon Izumi: she will probably be here shortly [11:01] Mic Bowman: and having a better sense of the layout she has in mind would help [11:02] Crista.Lopes @nile.ics.uci.edu:9000: hi [11:02] Mic Bowman: hi crista! [11:02] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Helo [11:02] Nebadon Izumi: hello [11:02] Seth.Nygard @refugegrid.com:8002: Hello [11:02] Richardus.Raymaker @hypergrid.org:8002: hi all [11:02] Aine.Caoimhe @refugegrid.com:8002: ah ha....suddenly you all just de-loading-ed for me [11:02] Crista.Lopes @nile.ics.uci.edu:9000: is this over voice? [11:02] Nebadon Izumi: voice is enabled [11:02] Crista.Lopes @nile.ics.uci.edu:9000: I mean the meeting [11:02] Nebadon Izumi: but we usually do chat [11:02] Mic Bowman: ok... why are name labels red on the left side of the hall and yellow on the other? [11:02] Crista.Lopes @nile.ics.uci.edu:9000: ok [11:02] Marcus.Llewellyn @grid.verymad.net:8002: Text has traditionally been prefered. [11:02] Nebadon Izumi: so it can be preserved in logs [11:03] Nebadon Izumi: i am logging in my radegast viewers to keynotes [11:03] Nebadon Izumi: so bit disttracted for a few more minutes here [11:04] Shez.Oyen @mirage.duckdns.org:9038: wow, they are diff colors on left and right [11:04] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Dahlia will probably be a few minutes. She would probably want to hear about the module changes. [11:04] Nebadon Izumi: Justin are you going to fire up the Avacon bots? [11:04] Marcus.Llewellyn @grid.verymad.net:8002: I don't see different name tag colors. [11:04] Justin Clark-Casey: mic: I am seeing console messages "Could not find packet with sequence number <n> to ack" [11:04] Shez.Oyen @mirage.duckdns.org:9038: I would have prob never noticed that [11:04] Mic Bowman: justin: good [11:05] Mic Bowman: that means that some of the acks are late coming in [11:05] Mic Bowman: how many do you see? [11:05] Richardus.Raymaker @hypergrid.org:8002: Just like RL :) [11:05] Mic Bowman: and whats show queues look like [11:05] Mic Bowman: for resent packets [11:05] Aine.Caoimhe @refugegrid.com:8002: while waiting then....I have a question re HG 1.,5 vs 2.0.....does 1.5 expose the (home) server to attacks or only the person travelling? [11:05] Justin Clark-Casey: on here, I'm seeing about 40. It's a pity they're not identified to the user [11:06] Justin Clark-Casey: 122 2420 6128 561 1884 0 0 0 0 3 0 0 Region (Staff Zone 1) # show throt User Region Type Max Target Actual Resend Land Wind Cloud Task Texture Asset kb/s kb/s kb/s kb/s kb/s kb/s kb/s kb/s kb/s kb/s Marcus.Llewellyn @ Staff Zone 1 Rt 1000 1152 1000 100 95 19 19 314 314 137 Nebadon Izumi Staff Zone 1 Rt 1000 4608 1000 100 68 13 13 339 339 124 Billy.Bradshaw @hg Staff Zone 1 Rt 1000 768 768 76 87 17 17 228 228 112 Mic Bowman Staff Zone 1 Rt 1000 2304 1000 100 77 15 15 331 331 128 Robert.Adams @home Staff Zone 1 Rt 1000 1459 1000 100 88 17 17 321 321 134 BlueWall.Slade @ga Staff Zone 1 Rt 1000 1152 1000 100 95 19 19 314 314 137 Shez.Oyen @mirage. Staff Zone 1 Rt 1000 3072 1000 100 73 14 14 335 335 126 Aine.Caoimhe @refu Staff Zone 1 Rt 1000 768 768 76 87 17 17 228 228 112 Justin Clark-Casey Staff Zone 1 Rt 1000 5120 1000 100 68 13 13 340 340 124 Seth.Nygard @refug Staff Zone 1 Rt 1000 1536 1000 100 86 17 17 322 322 133 Richardus.Raymaker Staff Zone 1 Rt 1000 2304 1000 100 77 15 15 331 [11:06] Justin Clark-Casey: Region (Staff Zone 1) # show q User Region Type Since Pkts Pkts Pkts Bytes Q Pkts Q Pkts Q Pkts Q Pkts Q Pkts Q Pkts Q Pkts Last In In Out Resent Unacked Resend Land Wind Cloud Task Texture Asset Marcus.Llewellyn @ Staff Zone 1 Rt 107 8561 7206 0 0 0 0 0 0 0 0 0 Nebadon Izumi Staff Zone 1 Rt 53 9049 6970 1 0 0 0 0 0 0 0 0 Billy.Bradshaw @hg Staff Zone 1 Rt 41 8786 11157 9 0 0 0 0 0 0 0 0 Mic Bowman Staff Zone 1 Rt 162 7401 6914 1 0 0 0 0 0 0 0 0 Robert.Adams @home Staff Zone 1 Rt 12 8200 8769 1 0 0 0 0 0 0 0 0 BlueWall.Slade @ga Staff Zone 1 Rt 166 4547 7630 382 0 0 0 0 0 0 0 0 Shez.Oyen @mirage. Staff Zone 1 Rt 37 7241 6801 0 0 0 0 0 0 0 0 0 Aine.Caoimhe @refu Staff Zone 1 Rt 14 6667 9498 21 -5 0 0 0 0 0 0 0 Justin Clark-Casey Staff Zone 1 Rt 57 6299 6504 23 0 0 0 0 0 0 0 0 Seth.Nygard @refug Staff Zone 1 Rt 69 5317 7627 0 0 0 0 0 0 0 0 0 Richardus.Raymaker Staff Zone 1 Rt 111 3870 4787 1 0 0 0 0 0 0 0 0 Crista.Lopes @nile Staff Zone 1 Rt 2817 1529 5289 10 0 0 0 0 0 0 0 0 Kayaker.Mag [11:06] Justin Clark-Casey: @nile Staff Zone 1 Rt 1000 3072 1000 100 73 14 14 335 335 126 Kayaker.Magic @gri Staff Zone 1 Rt 1000 256 256 11 16 11 11 25 25 15 [11:06] Justin Clark-Casey: not sure how helpful that is.... [11:06] Aine.Caoimhe @refugegrid.com:8002: ie if some malicious person wanted to do something to a grid running 1.5, would they have an easier time doing so than they would with 2.0? [11:06] Crista.Lopes @nile.ics.uci.edu:9000: yes [11:07] Mic Bowman: ok... that looks like a few packets being lost [11:07] Crista.Lopes @nile.ics.uci.edu:9000: but the differences affect mostly the inventory data [11:07] Aine.Caoimhe @refugegrid.com:8002: something that affects the entire grid or only the person traveling ? [11:07] Crista.Lopes @nile.ics.uci.edu:9000: the entire inventory data of the grid [11:07] Aine.Caoimhe @refugegrid.com:8002: drat [11:07] Billy.Bradshaw @hg.viewtwo.net:8600: Quick query before throttle testing discussed. OpenMetaverse.dll updated in core around 25th Dec, but cannot see a corresponding update to libopenmetaverse [11:08] Crista.Lopes @nile.ics.uci.edu:9000: there wsn't a release of libomv, just several commits [11:08] Crista.Lopes @nile.ics.uci.edu:9000: current libomv in opensim core corresponds to master head of libomv in github [11:09] Billy.Bradshaw @hg.viewtwo.net:8600: k [11:09] Justin Clark-Casey: this is where it's helpful to have the libomv commit level in the commit message [11:10] Mic Bowman: justin: can you post show queues and show throttles into the IM [11:11] BlueWall.Slade @gateway.bluewallvirtual.net:8002: show queues > mic bowmanim [11:12] BlueWall.Slade @gateway.bluewallvirtual.net:8002: :) [11:12] BlueWall.Slade @gateway.bluewallvirtual.net:8002: If we could use bash for the console [11:12] Mic Bowman: would be nice [11:12] Mic Bowman: i think i need to keep working to figure out the stats collectors [11:13] Mic Bowman: (all three of them) [11:13] BlueWall.Slade @gateway.bluewallvirtual.net:8002: would it make sense to consolidate them int o the monitoring module? [11:13] Robert.Adams @home.livingroomcam.us:14559: I'm for StatsManager.... it has a web interface as well as console interface [11:13] Justin Clark-Casey: the monitoring module is old. [11:14] Justin Clark-Casey: it would make sense to cnosolidate everything. The only reason it hasn't been done is lack of time [11:14] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ++ [11:14] Robert.Adams @home.livingroomcam.us:14559: a monitoring app can query for just the stats it wants [11:14] Mic Bowman: justin/nebadon: looks like everyone is getting their requested throttles right now given that there are per-client limits in place [11:14] Mic Bowman: client limits are 1mbps [11:14] BlueWall.Slade @gateway.bluewallvirtual.net:8002: I like the idea of using a rrdb to store the data for generic use [11:14] Justin Clark-Casey: mic: apart from kayaker. Throttle keeps changing there because of non-acked packets [11:14] Richardus.Raymaker @hypergrid.org:8002: stats in json fdormat, if its not the same we already ahve [11:15] Mic Bowman: they should change [11:15] Mic Bowman: but should generally stay around 1mbps [11:15] BlueWall.Slade @gateway.bluewallvirtual.net:8002: The dotnet lib for that has a funky license though. [11:15] Nebadon Izumi: ok 60 clients logged in [11:15] Mic Bowman: which is what it looks like [11:15] Nebadon Izumi: actually i must say, compared to the OSCC load testing [11:15] Nebadon Izumi: my clients had a much easier time logging in [11:15] Nebadon Izumi: already feels like improvement to me [11:15] Robert.Adams @home.livingroomcam.us:14559: StatsManager will return web queries in JSON (and optionally callback form) [11:16] Mic Bowman: there is one more thing i want to add... a "grace period" [11:16] Justin Clark-Casey: right, I'm just saying that kayaker is not getting requested for that reason [11:16] BlueWall.Slade @gateway.bluewallvirtual.net:8002: snmp FTW :) [11:16] Mic Bowman: justin: kayaker looks to have target at 256kbps [11:16] Robert.Adams @home.livingroomcam.us:14559: http://www.virtualportland.org/SimStats/ is a simple stats display for my standalone region [11:16] Robert.Adams @home.livingroomcam.us:14559: when I was running DSG, it included graphs of the bandwidth between the servers [11:17] Mic Bowman: that is a viewer throttle i would guess [11:17] BlueWall.Slade @gateway.bluewallvirtual.net:8002: nice Robert [11:17] Nebadon Izumi: ya very nice [11:17] Robert.Adams @home.livingroomcam.us:14559: SimStats does a GET query to get the stats and gets JSON back which the JS tableifies [11:18] Aine.Caoimhe @refugegrid.com:8002: is there any possible mechanism that can be used to "rebake" attachments (yes, I know they aren't baked in the first place, but what I mean is to have the ability to ask the sim to re-send your attachments) that isn't going to involve a viewer change? I've found increasingly that since every new HG tp times out and requires a second or even third attempt, some or all of me doesn't display correctly to pre-existing people in the region I tp to. They only see me complete when they relog (me relogging and returning doesn't trigger the update in their viewer) [11:18] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: BlueWall, IIRC, I would second rrdb format for stats. There are some nice tools to turn that data in to a chart. [11:18] Richardus.Raymaker @hypergrid.org:8002: rewear is in SL the normal way. or change outfit and then back [11:18] Robert.Adams @home.livingroomcam.us:14559: there is a console command to force resending, isn't there? [11:18] Aine.Caoimhe @refugegrid.com:8002: cam [11:18] BlueWall.Slade @gateway.bluewallvirtual.net:8002: I updated thsi old stats app a few days ago : http://gateway.bluewallvirtual.net:8000/~jamesh/simstats.html [11:19] Aine.Caoimhe @refugegrid.com:8002: can't do that with HG2.0 without having everything in suitcase [11:19] BlueWall.Slade @gateway.bluewallvirtual.net:8002: That suitcase is your firewal. [11:19] Aine.Caoimhe @refugegrid.com:8002: the console command only resends scene (region) objects, not attachments [11:19] Seth.Nygard @refugegrid.com:8002: And also a console command is of no use to the many people that have their regions hosted for them. [11:20] Crista.Lopes @nile.ics.uci.edu:9000: what's the problem with attachments? [11:21] Aine.Caoimhe @refugegrid.com:8002: if you are wearing anything that takes more than 15 seconds to be sent from your source region to your destination region the tp times out and then 90% of the time on re-tp they don't appear for people who were already in the region at the time of the initial failed attempt [11:21] Aine.Caoimhe @refugegrid.com:8002: (15 seconds total for the entire tp process to finish) [11:21] Aine.Caoimhe @refugegrid.com:8002: or might be 20 [11:22] Nebadon Izumi: ya i am not sure rebake is the correct term there [11:22] Justin Clark-Casey: there were improvements to that made in master. I wouldn't suggest trying to make any changes until the next release, personally [11:22] Nebadon Izumi: more like reattach [11:22] Aine.Caoimhe @refugegrid.com:8002: yeah...reattach without actually having to remove [11:22] Aine.Caoimhe @refugegrid.com:8002: (so it can be done for something you're wearing that isn't in suitcase) [11:23] Nebadon Izumi: i have made it a habbit now to always keep my appearance in my suticase [11:23] Crista.Lopes @nile.ics.uci.edu:9000: yeah, that sounds like a bug that can be fixed eventually with some tweaking on timings. Shouldn't need manual intervention [11:24] Nebadon Izumi: things seem to work much better if you do that [11:24] BlueWall.Slade @gateway.bluewallvirtual.net:8002: maybe link the curent outfit under the suitcase? [11:24] Marcus.Llewellyn @grid.verymad.net:8002: First time visitors to my grid have failed TPs more often than not due to asset transfer taking longer than the TP timeout. :/ [11:24] Aine.Caoimhe @refugegrid.com:8002: yes [11:24] Aine.Caoimhe @refugegrid.com:8002: it's a very short time out [11:25] Seth.Nygard @refugegrid.com:8002: That issue is common with all grids [11:25] Aine.Caoimhe @refugegrid.com:8002: and all assets have to be sent in their entirety in that window [11:25] Nebadon Izumi: ya I have seen that too [11:25] Nebadon Izumi: first TP fails [11:25] Nebadon Izumi: second one is fine [11:25] Aine.Caoimhe @refugegrid.com:8002: yeps [11:25] Aine.Caoimhe @refugegrid.com:8002: 2nd almost always is instant [11:25] Nebadon Izumi: ive never experienced it on my own grids [11:25] Nebadon Izumi: usually when someone asks me to visit theirs [11:25] Nebadon Izumi: even on my European grids [11:25] Nebadon Izumi: havent experienced it [11:26] Aine.Caoimhe @refugegrid.com:8002: but if anyone was in the region when you failed the 1st attempt, it's hit and miss as to which of your attachments they see [11:26] Aine.Caoimhe @refugegrid.com:8002: if they relog, they'll see everything, and anyone new who enters is sent everything...it's only existing avi where it's hit and miss [11:27] Marcus.Llewellyn @grid.verymad.net:8002: That can be especially annoying if you're mesh. It's one thing to be missing hair. But your whole body is even more vexing. Heh. [11:27] Aine.Caoimhe @refugegrid.com:8002: exactly [11:27] Nebadon Izumi: ya Mesh seems to have some trouble wiht hypergrid [11:27] Aine.Caoimhe @refugegrid.com:8002: and worse when it's mesh clothing over mesh body [11:27] Nebadon Izumi: there are more problems than just that [11:27] Nebadon Izumi: like if i have the same oar loaded on 2 different grids [11:27] Nebadon Izumi: with the same exact assets [11:27] Aine.Caoimhe @refugegrid.com:8002: it's because mesh often will make the total tp time take more than the 15 second window [11:27] Nebadon Izumi: when i teleport from one to the other [11:27] Nebadon Izumi: i end up with 1000 yellow pyramids [11:27] Crista.Lopes @nile.ics.uci.edu:9000: Mic about version number of modules... I'm looking into it now, and I think that the best is to use the following convention [11:27] Nebadon Izumi: for all the mesh [11:27] Nebadon Izumi: prims are fine [11:28] Nebadon Izumi: only a cache clear fixes it [11:28] Crista.Lopes @nile.ics.uci.edu:9000: if my module is being compiled against opensim 0.8.1, my modules version numbers are 0.8.1.something [11:28] Crista.Lopes @nile.ics.uci.edu:9000: that way ppl will know that my module has been cpmpiled for opensim 0.8.1 [11:28] Aine.Caoimhe @refugegrid.com:8002: it's issues like that which would be really nice to fix considering how many people are HGing these days and the general increase in the use of mesh [11:29] BlueWall.Slade @gateway.bluewallvirtual.net:8002: hmm, how about the version of the module's code? [11:30] Crista.Lopes @nile.ics.uci.edu:9000: that's the .something [11:30] Crista.Lopes @nile.ics.uci.edu:9000 whispers: 0.8.1.something [11:30] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ok [11:30] BlueWall.Slade @gateway.bluewallvirtual.net:8002: where do you specifiy that? [11:31] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002 still has the TP issue when going from var to regular where the viewer locks up or crashes [11:31] Crista.Lopes @nile.ics.uci.edu:9000: so, the first 3 numbers are opensim version; any additional ones are the module's [11:31] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ok [11:31] BlueWall.Slade @gateway.bluewallvirtual.net:8002: I rememebr something about the 3rd digit using * [11:31] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: Addin("Diva.MyModule", "0.8.1.1")] [11:31] Justin Clark-Casey: what about things like 0.7.6.2 ? [11:32] Justin Clark-Casey: which is an opensim version number [11:32] Crista.Lopes @nile.ics.uci.edu:9000: do we have that justincc? [11:32] BlueWall.Slade @gateway.bluewallvirtual.net:8002: some magic about the comile version [11:32] Crista.Lopes @nile.ics.uci.edu:9000: aren't our numbers just 3 digits? [11:32] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall that is for .net version numbers [11:32] Crista.Lopes @nile.ics.uci.edu:9000: not mono addins [11:32] Crista.Lopes @nile.ics.uci.edu:9000: these are 2 different number series [11:32] Justin Clark-Casey: no, current release is 0.8.0.3, to signal the third bug fix release of 0.8 [11:33] Crista.Lopes @nile.ics.uci.edu:9000: really? 4 digits? [11:33] Aine.Caoimhe @refugegrid.com:8002 nods [11:33] Nebadon Izumi: mono has a similar practice [11:33] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Yes, there can be as many as four in some systems of version control [11:33] Nebadon Izumi: with versioning [11:34] BlueWall.Slade @gateway.bluewallvirtual.net:8002: I will look for the docs I read about it, then send an email. [11:34] Crista.Lopes @nile.ics.uci.edu:9000: well, we should just decide how many digits we want for OpenSim versioning. Then the addin modules can take beyond that [11:34] Marcus.Llewellyn @grid.verymad.net:8002: We've sorta come to treat the first minor number as our major. [11:34] Crista.Lopes @nile.ics.uci.edu:9000: if we want 4 for OS, then my module's version should be [assembly: Addin("DivaMy.Module", "0.8.1.0.1")] [11:35] Crista.Lopes @nile.ics.uci.edu:9000: we need to have a way ot informing consumers of these modules about which OS version the module was compiled against [11:35] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If this versioning system is put in place for modules it will affect 3rd party modules (eg. ossearch and osprofile) requiring minor(?) changes to the modules? [11:36] Crista.Lopes @nile.ics.uci.edu:9000: yes [11:36] Crista.Lopes @nile.ics.uci.edu:9000: very minor change [11:36] Justin Clark-Casey: crista: opensim versioning would be a max of 4 digits [11:36] Crista.Lopes @nile.ics.uci.edu:9000: ok, then [11:36] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: ok. I can put those changes in place when I know what needs to be done [11:36] Crista.Lopes @nile.ics.uci.edu:9000: my modules will use 5 digits, where the last one is for the module itself [11:37] Mic Bowman: so the opensim version plus module version [11:37] Mic Bowman: how is that checked/enforced? [11:37] Mic Bowman: the versioning [11:37] Crista.Lopes @nile.ics.uci.edu:9000: yes, and it's not [11:37] Mic Bowman: on load? [11:37] Crista.Lopes @nile.ics.uci.edu:9000: well [11:37] Crista.Lopes @nile.ics.uci.edu:9000: it's on load yes [11:37] Crista.Lopes @nile.ics.uci.edu:9000: but this check is not the complete story [11:37] Mic Bowman: so if i have a dependency on 0.8.3 opensim module [11:38] Mic Bowman: and the actual version is 0.8.1 [11:38] Mic Bowman: what happens? [11:38] Crista.Lopes @nile.ics.uci.edu:9000: it will fail to load that module, I think [11:38] Crista.Lopes @nile.ics.uci.edu:9000: but as I said these checks are not sufficient [11:38] Mic Bowman: what kind of message? [11:39] Mic Bowman: and is there a way to check at install time? [11:39] Crista.Lopes @nile.ics.uci.edu:9000: I don't remember but I can tell you in a bit [11:39] Justin Clark-Casey: I think all opensim version numbrs will have to be 4 digits for consistency. So it would technically be 0.8.3.0 [11:39] Mic Bowman: just thining about friendly messages [11:39] Crista.Lopes @nile.ics.uci.edu:9000: I'm right now creating diva modules in this manner [11:39] BlueWall.Slade @gateway.bluewallvirtual.net:8002: how about the AddinDependency? [11:40] BlueWall.Slade @gateway.bluewallvirtual.net:8002: [assembly: AddinDependency("OpenSim", "0.5")] [11:40] Crista.Lopes @nile.ics.uci.edu:9000: that's the thing that we're talking about [11:41] BlueWall.Slade @gateway.bluewallvirtual.net:8002: seems that it should change each time we change an interface [11:41] Crista.Lopes @nile.ics.uci.edu:9000: for exampoe, my modules now have this [11:41] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: AddinDependency("OpenSim.Region.Framework", "0.8.1")] [11:41] BlueWall.Slade @gateway.bluewallvirtual.net:8002: OK [11:41] BlueWall.Slade @gateway.bluewallvirtual.net:8002: [assembly: Addin("SimBus", "0.1")] [11:41] Crista.Lopes @nile.ics.uci.edu:9000: justin should we switch to 4 digits all over? [11:41] BlueWall.Slade @gateway.bluewallvirtual.net:8002: So, that one is my module's version [11:41] Justin Clark-Casey: crista: in those addin dpeendency annotations? [11:42] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall, I am going to number my modules so that they include opensim's version [11:42] Crista.Lopes @nile.ics.uci.edu:9000: I think you should do the same [11:42] Crista.Lopes @nile.ics.uci.edu:9000: otherwise people whon't know [11:42] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: The fourth one could be optional. [11:44] Justin Clark-Casey: It would be easier just to have 4 digit v numbers always, if other stuff can handle it [11:44] Justin Clark-Casey: special cases can be a pain [11:44] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002 nods [11:44] Crista.Lopes @nile.ics.uci.edu:9000: ok, there are two numbers series, and they are independent [11:44] Crista.Lopes @nile.ics.uci.edu:9000: one is the .net versioning numbers [11:45] Crista.Lopes @nile.ics.uci.edu:9000: they seem to have 3+some special digit that they manage automatically [11:45] Crista.Lopes @nile.ics.uci.edu:9000: these are our declarations like this [11:45] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: AssemblyVersion("0.8.1.*")] [11:45] Crista.Lopes @nile.ics.uci.edu:9000: another completely different numbers series is mono addins version numbers [11:45] Seth.Nygard @refugegrid.com:8002: If OpenSim is already using all 4 numbers for its version, then a fifth digit is needed to support Christa's module versioning scheme. Release.Major.Minor.BugFixLevel +ModuleVersion [11:46] Crista.Lopes @nile.ics.uci.edu:9000: they are independent, and we could use completely different numbers [11:46] Crista.Lopes @nile.ics.uci.edu:9000: but I think that's a bad idea [11:46] Crista.Lopes @nile.ics.uci.edu:9000: I think we should use the same numbers [11:46] Crista.Lopes @nile.ics.uci.edu:9000: oh [11:46] Crista.Lopes @nile.ics.uci.edu:9000: and there's a 3rd number series [11:46] Crista.Lopes @nile.ics.uci.edu:9000: the version number we tag our releases with [11:46] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I wouldn't mix two sets of version numbers in one [11:47] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: The modules I maintain have their own version numbers that are x.y [11:47] Crista.Lopes @nile.ics.uci.edu:9000: thats in VersionInfo.cs [11:47] Crista.Lopes @nile.ics.uci.edu:9000: Andrew you can do that [11:48] Crista.Lopes @nile.ics.uci.edu:9000: however, people who use your modules will not know against which version of opensim your module was compiled [11:48] Crista.Lopes @nile.ics.uci.edu:9000: you need to tell them by some other means [11:48] Crista.Lopes @nile.ics.uci.edu:9000: as such I am going to include opensim's version in my own module's versions, because I don't want people to ask me questions. You can do whatever you want for yours! [11:49] Mic Bowman: are you including that in your module's version? [11:49] Crista.Lopes @nile.ics.uci.edu:9000: yes [11:49] Mic Bowman: or in the opensim dependency you puth there [11:49] Mic Bowman: ok [11:49] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: Addin("Diva.Modules", "0.8.1.0.1")] [11:49] Mic Bowman: and is there a way to write some other script that pulls the opensim version number? [11:49] Mic Bowman: to test it on install? [11:50] Crista.Lopes @nile.ics.uci.edu:9000: not with mautil. no [11:50] Mic Bowman: my modules have two statements [11:50] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: Addin("Diva.Modules", "0.8.1.0.1")] [assembly: AddinDependency("OpenSim.Region.Framework", "0.8.1")] [11:50] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If you want to mix two vresions in one you need a way to separate out the OS version from the module version. [11:50] Mic Bowman: yeah [11:50] Crista.Lopes @nile.ics.uci.edu:9000: the first is documentation for consumers of my module [11:50] Crista.Lopes @nile.ics.uci.edu:9000: the second is for mono addins framework to check [11:51] Mic Bowman: got it [11:51] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Then, that assembly version can actualy be anything (fro the benefit of the mautils)? [11:52] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall yes. mautils doesn't check assembly version numbers [11:52] Crista.Lopes @nile.ics.uci.edu:9000: it only checks mono addin version numbers [11:52] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Ok [11:52] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: will someone be documenting the proposed versioning system so we can do a proper review and provide comments, or suggest changes, before implementation? [11:52] Crista.Lopes @nile.ics.uci.edu:9000: so, as I said, these chacks are not sufficient [11:52] Crista.Lopes @nile.ics.uci.edu:9000: 'there are many scenarios in which mautils will suceed, but the running time will fail [11:53] BlueWall.Slade @gateway.bluewallvirtual.net:8002: have you packed the assemblies into a repo yet? [11:53] Crista.Lopes @nile.ics.uci.edu:9000: Andrew, the versioning of your own modules is entirely up to you. I have a prefered way of doing it, that I'm explaining, and I would recommend others to use it. But it's not mandatory in any way shape or form [11:53] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall yes [11:54] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If one module needs to use another there should be a way to query the version number of a module. [11:54] BlueWall.Slade @gateway.bluewallvirtual.net:8002: It makes a basic index.html [11:55] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I'd be interested in seeing the proposal so I'm clear on compile time vs runt ime versioning of modules. [11:55] Justin Clark-Casey: crista: so a way forward is to make the OpenSim version always four digits now? [11:56] Crista.Lopes @nile.ics.uci.edu:9000: I don't know justin, that seems like a hassle [11:56] Justin Clark-Casey: before, the first release of 0.7.6 would just be set as "0.7.6" [11:56] Justin Clark-Casey: and then a security fix release is 0.7.6.1 [11:56] Crista.Lopes @nile.ics.uci.edu:9000: I really don't remember any releases having more than 3 digits, but if you say they do, then let's make that consistent [11:57] Justin Clark-Casey: look on the download page right now :) [11:57] Justin Clark-Casey: 0.8 is on it's third security release, 0.8.0.3 [11:57] Crista.Lopes @nile.ics.uci.edu:9000: ok [11:57] Crista.Lopes @nile.ics.uci.edu:9000: then yes, let's make it consistent [11:57] Crista.Lopes @nile.ics.uci.edu:9000: 4 digits throughout [11:58] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If it will make the code harder to deal with by having both 3 and 4 digit versions, 4 would be the way to go for all. [11:58] Mic Bowman: security releases do not change interfaces [11:58] Crista.Lopes @nile.ics.uci.edu:9000: heheh [11:58] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Now, have you updated a module and have mautils update it? [11:58] Crista.Lopes @nile.ics.uci.edu:9000: so yo'd hope mic! :) [11:58] Mic Bowman: :-) [11:58] Justin Clark-Casey: yes, I would strenuously avoid that. But if one's idea of an interface is the entire surface of opensimulator methods then that might be very hard to avoid [11:58] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: The default value for a mising fourth digit would be 0 anyway [11:58] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ha, those should be major releases [11:58] Justin Clark-Casey: security rleases are as simple as possible precisely so that it's easy to upgrade [11:59] Mic Bowman: my point was just that specifying a dependency on 0.8.* is probably good enough [11:59] Mic Bowman: or more likely 0.8.1.* [11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ls [11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ++ [12:00] Mic Bowman: so for crista's proposal... my module could be 0.8.0.myversion [12:00] Crista.Lopes @nile.ics.uci.edu:9000: I would rather have 3 digits... but I don't know... I'm afraid that interface changes will slip through [12:00] Mic Bowman: implying a dependency on 0.8.0.* [12:00] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If * is supported that leads to cases of only 3 parts in the version number which then makes on think that the fourth should only be used if it isn't 0 [12:00] Crista.Lopes @nile.ics.uci.edu:9000: yes, that would be my preference too [12:00] Justin Clark-Casey: how does it know that'#s not 0.8.0.13 for instance? [12:00] Crista.Lopes @nile.ics.uci.edu:9000: * is NOT supported in mono addin numbers [12:00] Justin Clark-Casey: er, 0.8.0.3 [12:00] Mic Bowman: so security updates that require interface changes require a minor version number upgrade [12:01] Justin Clark-Casey: that might not be possible [12:01] Justin Clark-Casey: if both 0.7.6 and 0.7.7 exist and they both need updates [12:01] Mic Bowman: poo... you're right [12:01] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Mic, that would push 0.8 to 0.9. Did you mean a bump of micro (third), or the extra fourth? [12:01] Mic Bowman: back patches [12:01] Mic Bowman: third [12:01] Crista.Lopes @nile.ics.uci.edu:9000: justin, teverything that we're talking about will only apply to versions 0.8.1 onwards [12:01] Mic Bowman: sorry [12:01] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: np. [12:01] Justin Clark-Casey: yes, the numbers are just for examples [12:02] Justin Clark-Casey: 0.8.1 and 0.8.2, then say [12:02] Justin Clark-Casey: or rather, 0.8.1.0, 0.8.2.0 [12:02] Crista.Lopes @nile.ics.uci.edu:9000: well, I'd rather have 3 digits for this, but I'm hearing you, and let's go with 4 [12:03] Justin Clark-Casey: there's also a problem that sometimes those updates don't just contain security fixes [12:03] Crista.Lopes @nile.ics.uci.edu:9000: that means that in OpenSim, the addins will be versioned 0.8.1.0 [12:03] Justin Clark-Casey: sometimes they do contain backported functionality [12:03] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If you want to include use of a * then the parsing of the version should handle missing parts. [12:03] Crista.Lopes @nile.ics.uci.edu:9000: Andrew, mono addins deosn't support * [12:03] Justin Clark-Casey: but nothing super major..... [12:04] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: ok. Somone suggested use of .* [12:05] Crista.Lopes @nile.ics.uci.edu:9000: ok, we'll use four digits for the opensim version numbers in mono addins -land [12:06] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I'll bet you some people may forget the fourth sometimes as some (lots of ?) other open source projects only use 3 digits [12:06] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: :) [12:07] Crista.Lopes @nile.ics.uci.edu:9000: well, my concern is that WE will forget it too [12:07] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: yeah. It will probably happen [12:07] Mic Bowman: i wasn't suggesting a regex (at least not for mono dependencies) [12:07] Crista.Lopes @nile.ics.uci.edu:9000: if we go with , that means that everytime justin does a trivial release that bumps up the 4th digit he will need to remember to update the mono addins version numbers [12:07] Mic Bowman: just that the dependcy is on any version like that [12:08] Mic Bowman: and... a separate install script should be able to check for more expressive dependencies [12:08] Mic Bowman: specifying the dependencies in the modules is necessary [12:08] Justin Clark-Casey: crista: I will put it on my big release to do list, which will probably encourage me to automate it in some way [12:09] Mic Bowman: for building an install script that non-techies could install will require a different kind of install time check [12:09] Justin Clark-Casey: tbh, most of it is automated and I really need to push the scripts into core itself, though some of it is specific to the way opensimulator.org is laid out, etc. [12:09] BlueWall.Slade @gateway.bluewallvirtual.net:8002: [assembly: AssemblyVersion("0.8.1.*")] [12:10] Nebadon Izumi: ouch [12:10] Nebadon Izumi: heh [12:10] Nebadon Izumi turns speakers down [12:10] BlueWall.Slade @gateway.bluewallvirtual.net:8002: We ahve those in the tree now. Could we use those, somehow? [12:10] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall those are .net asembly version numbers [12:10] Crista.Lopes @nile.ics.uci.edu:9000: not mono addin version numbers [12:10] Crista.Lopes @nile.ics.uci.edu:9000: DIFFERETE THING [12:10] BlueWall.Slade @gateway.bluewallvirtual.net:8002: right [12:10] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ok, so we won't track those in the adins? [12:10] Crista.Lopes @nile.ics.uci.edu:9000: only informally [12:11] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: That's is why a doc would be helpful to point out the two different sets of numbers and their use [12:11] BlueWall.Slade @gateway.bluewallvirtual.net:8002: that is what I'm saying [12:11] Crista.Lopes @nile.ics.uci.edu:9000: I believe that we need to track them, otherwise it's confusing as hell [12:11] Crista.Lopes @nile.ics.uci.edu:9000: you'd have an opensim release 0.8.1 corresponding to an opensim addin release 0.1 [12:11] Crista.Lopes @nile.ics.uci.edu:9000: chaos! [12:11] Justin Clark-Casey: 0.8.1.0..... [12:11] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: hehe [12:12] Justin Clark-Casey: it's bad enough with ppl out there calling it opensim 8.0... [12:12] Crista.Lopes @nile.ics.uci.edu:9000: currently there is nothing in code that enforces that those 2 numbers ar ethe same [12:12] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Yes, if we are updating the assembly versions now, and we want soemthing similar in the addins, then we might make some scripts to update core modules crom that information? [12:12] Crista.Lopes @nile.ics.uci.edu:9000: in fact there is nothing in the code that enforces that all opensim DLLs have the same assembly version number [12:12] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ok [12:13] Crista.Lopes @nile.ics.uci.edu:9000: maybe we should enforce these things [12:13] Crista.Lopes @nile.ics.uci.edu:9000: let me see if it works [12:13] BlueWall.Slade @gateway.bluewallvirtual.net:8002: we could even go as far as signing the asemblies too :) [12:13] Richardus.Raymaker @hypergrid.org:8002: oh, itsd late. need to go. see you all. [12:13] BlueWall.Slade @gateway.bluewallvirtual.net:8002: see ya RiRa [12:13] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Bye, Richardus [12:13] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Happy new year [12:13] Aine.Caoimhe @refugegrid.com:8002: happy New Year everyone [12:14] Richardus.Raymaker @hypergrid.org:8002: You to. sooner better. enough fireworks noise [12:14] Seth.Nygard @refugegrid.com:8002: Have a great day Richardus, take care [12:14] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Yes, Hapy New Year [12:14] Richardus.Raymaker @hypergrid.org:8002: Happy newyear [12:14] Seth.Nygard @refugegrid.com:8002: Have a great day Aine, take care [12:14] Nebadon Izumi: are we going to jump to the Keynotes? [12:15] Nebadon Izumi: or do we have enough data? [12:15] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I have to go out to do some errands so I can't head over there toay [12:15] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: s/toay/today/ [12:15] Crista.Lopes @nile.ics.uci.edu:9000: Justin, t looks like it's possible to use OPenSim.VersionInfo in the mono addin annotations, but the version number cannot have letters, just numbers [12:16] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Nebadon, why don't we take the rest of the year off? ;) [12:16] Nebadon Izumi: well Mic requested the load test [12:16] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: ok [12:16] Nebadon Izumi: i have already given him some data on my 60 bots [12:16] Nebadon Izumi: they have been logged in at keynotes the entire meeting [12:16] Nebadon Izumi: and i assume justin has been feeding mic data from the meeting [12:16] Justin Clark-Casey: crista: so VersionInfo.VERSION_NUMBER isn't ok? [12:17] Mic Bowman: yup [12:17] Justin Clark-Casey: currently "0.8.1", which would become 0.8.1.0 of course [12:17] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I'd be fine with that [12:17] Justin Clark-Casey: ah, maybe because VersionInfo.Version tacks on the dev, rc1, etc thing? [12:17] Nebadon Izumi: i really need to replace these chairs [12:17] Nebadon Izumi: they suck [12:18] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I need to get going. [12:18] BlueWall.Slade @gateway.bluewallvirtual.net:8002: what's with the hanging animation? [12:18] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: See everyone in the new year. [12:18] BlueWall.Slade @gateway.bluewallvirtual.net:8002: see ya Andrew [12:19] Seth.Nygard @refugegrid.com:8002: Have a great day Andwer, take care [12:19] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: Thanks, Seth [12:19] Seth.Nygard @refugegrid.com:8002: Hapy New Year to everyone [12:19] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Bye Seth [12:19] Seth.Nygard @refugegrid.com:8002: Were we still planning a mini load test today? [12:20] Mic Bowman: i'm gonna head over to the key note [12:20] BlueWall.Slade @gateway.bluewallvirtual.net:8002: me too [12:20] Justin Clark-Casey: yeah, let's see if we can reproduce a previous issuents [11:50] Crista.Lopes @nile.ics.uci.edu:9000: [assembly: Addin("Diva.Modules", "0.8.1.0.1")] [assembly: AddinDependency("OpenSim.Region.Framework", "0.8.1")] [11:50] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If you want to mix two vresions in one you need a way to separate out the OS version from the module version. [11:50] Mic Bowman: yeah [11:50] Crista.Lopes @nile.ics.uci.edu:9000: the first is documentation for consumers of my module [11:50] Crista.Lopes @nile.ics.uci.edu:9000: the second is for mono addins framework to check [11:51] Mic Bowman: got it [11:51] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Then, that assembly version can actualy be anything (fro the benefit of the mautils)? [11:52] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall yes. mautils doesn't check assembly version numbers [11:52] Crista.Lopes @nile.ics.uci.edu:9000: it only checks mono addin version numbers [11:52] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Ok [11:52] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: will someone be documenting the proposed versioning system so we can do a proper review and provide comments, or suggest changes, before implementation? [11:52] Crista.Lopes @nile.ics.uci.edu:9000: so, as I said, these chacks are not sufficient [11:52] Crista.Lopes @nile.ics.uci.edu:9000: 'there are many scenarios in which mautils will suceed, but the running time will fail [11:53] BlueWall.Slade @gateway.bluewallvirtual.net:8002: have you packed the assemblies into a repo yet? [11:53] Crista.Lopes @nile.ics.uci.edu:9000: Andrew, the versioning of your own modules is entirely up to you. I have a prefered way of doing it, that I'm explaining, and I would recommend others to use it. But it's not mandatory in any way shape or form [11:53] Crista.Lopes @nile.ics.uci.edu:9000: BlueWall yes [11:54] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If one module needs to use another there should be a way to query the version number of a module. [11:54] BlueWall.Slade @gateway.bluewallvirtual.net:8002: It makes a basic index.html [11:55] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: I'd be interested in seeing the proposal so I'm clear on compile time vs runt ime versioning of modules. [11:55] Justin Clark-Casey: crista: so a way forward is to make the OpenSim version always four digits now? [11:56] Crista.Lopes @nile.ics.uci.edu:9000: I don't know justin, that seems like a hassle [11:56] Justin Clark-Casey: before, the first release of 0.7.6 would just be set as "0.7.6" [11:56] Justin Clark-Casey: and then a security fix release is 0.7.6.1 [11:56] Crista.Lopes @nile.ics.uci.edu:9000: I really don't remember any releases having more than 3 digits, but if you say they do, then let's make that consistent [11:57] Justin Clark-Casey: look on the download page right now :) [11:57] Justin Clark-Casey: 0.8 is on it's third security release, 0.8.0.3 [11:57] Crista.Lopes @nile.ics.uci.edu:9000: ok [11:57] Crista.Lopes @nile.ics.uci.edu:9000: then yes, let's make it consistent [11:57] Crista.Lopes @nile.ics.uci.edu:9000: 4 digits throughout [11:58] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: If it will make the code harder to deal with by having both 3 and 4 digit versions, 4 would be the way to go for all. [11:58] Mic Bowman: security releases do not change interfaces [11:58] Crista.Lopes @nile.ics.uci.edu:9000: heheh [11:58] BlueWall.Slade @gateway.bluewallvirtual.net:8002: Now, have you updated a module and have mautils update it? [11:58] Crista.Lopes @nile.ics.uci.edu:9000: so yo'd hope mic! :) [11:58] Mic Bowman: :-) [11:58] Justin Clark-Casey: yes, I would strenuously avoid that. But if one's idea of an interface is the entire surface of opensimulator methods then that might be very hard to avoid [11:58] Andrew.Hellershanks @login.greatcanadiangrid.ca:8002: The default value for a mising fourth digit would be 0 anyway [11:58] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ha, those should be major releases [11:58] Justin Clark-Casey: security rleases are as simple as possible precisely so that it's easy to upgrade [11:59] Mic Bowman: my point was just that specifying a dependency on 0.8.* is probably good enough [11:59] Mic Bowman: or more likely 0.8.1.* [11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ls [11:59] BlueWall.Slade @gateway.bluewallvirtual.net:8002: ++