Chat log from the meeting on 2016-05-31

From OpenSimulator

Jump to: navigation, search

[11:15] Zani.Zanadu may i ask who is running the show ? I been off sick for a few months
[11:15] George Equus: Usually Andrew Hellershank
[11:15] Robert Adams: we don't have Dan Banner here who is often the MC
[11:15] Robert Adams: but that said, it's usually people asking questions
[11:16] Seth.Nygard Dan is usually working during this meeting
[11:16] Gavin.Hird This mantis: needs some attention
[11:16] Zani.Zanadu So who are our movers and shakers ?
[11:16] Seth.Nygard So how much work is being done in 0.9 to fix some of the many issues and cleanup the code after the 5500+ commits were dropped in it
[11:17] Gavin.Hird of the things broken in 0.9
[11:18] Seth.Nygard At this stage as a grid owner and tester I can do only VERY limited testing on 0.9 because of the many incompatibilities with 0.8.2. HG is somewhat broken with it leaving more people to stop travelling between grids due to poor experiences
[11:19] Robert Adams: ya... I don't like that 0.9 is so broken... I'm going to be spending some time digging to the merge...
[11:19] Robert Adams: I'm going to first go through all the changes manually rather than the scatter shot of going into individual bugs
[11:19] Zani.Zanadu is 0.9 so broken as to make it unviable to fix it ?
[11:20] Andrew.Hellershanks Hello, everyone. Sorry I'm late. I was in the middle of wiring up a circuit.
[11:20] Gavin.Hird it is not THAT broken
[11:20] Seth.Nygard I've spent a lot of time in the code and it is a real shame that the way the merges were done lost/broke the chronology and evolution from 0.8.2. It is very hard to make sense of what was actually changed and why.
[11:20] vegaslon plutonian: the answers for most of the breakage are out in the open it just takes people time to read them
[11:20] Zani.Zanadu sorry gavin it just sounds so bad ( i been away for 4 months )
[11:20] Gavin.Hird I agree with you on that Seth
[11:20] Robert Adams: I don't think it's unfixable... just needs some TLC
[11:21] Billy.Bradshaw Problem is no one knows the extent things have changed due to the size of the merge
[11:21] Gavin.Hird I tried to apply all the changes one by one to get from 0.8.2 to 0.9 and had to give up
[11:21] Robert Adams: and a bunch of the core developers have been off doing RL and things
[11:21] Gavin.Hird it was impossible to make sense of some of it
[11:21] Andrew.Hellershanks The changelog for many of the changes don't really help explain the changes.
[11:22] Robert Adams: I agree Seth, the merge was done not the way I'd have done it... but now have to dig out of where we are
[11:22] Billy.Bradshaw Behavior changes seem to be a worry, don't know it has changed until you use something
[11:22] Gavin.Hird and many just says merge and it is not aparent where the participants in the merge originates from
[11:22] Seth.Nygard I have often wondered if everyone would not be much better served by rolling 0.9 back to 0.8.3 and then applying the changes in a much more controlled and managed manner. That help a great deal with testing and actually validating what needs to be done. I see a lot of regressions in 0.9 that remove the optimizations Justin did and no real reason for them other then that is not what Avination did.
[11:23] Wolf.Freeman i hope they will NOT CHANGE the Sit-Positions again^^
[11:23] Seth.Nygard No matter what is done is a LOT of work
[11:23] Andrew.Hellershanks Seth, Nothing stopping you from doing that. ;)
[11:23] Robert Adams: I'm not spending time re-doing what has been done... I plan on going forward and fixing 0.9
[11:23] Boba Solo: I agree Seth
[11:24] Wolf.Freeman #- OK ! -#
[11:24] Seth.Nygard And to some degree that is exactly what I have been doing for my own purposes. That however does little to help the community as a whole and makes it almost impossible to report issues due to version differences
[11:24] Andrew.Hellershanks Seth, if you are aware of regressions you should file a mantis about it so it can be checked.
[11:24] Zani.Zanadu seems very splintered :(
[11:24] Boba Solo: it seems like a huge undertaking fixing such a huge project in this condition!
[11:24] Gavin.Hird This one Andrew needs attention:
[11:24] Robert Adams: and, since our test cases are people's grids, it makes regression testing hard :)
[11:24] Boba Solo: reverting it seems logical
[11:25] Gavin.Hird there is a semi-fix, but needs more work
[11:25] Seth.Nygard The problem is with 0.9 there is no documentation of expected behaviour. Major issues reported in Mantis go over 6 months before they get addressed. Everyone I speak to is VERY concerned with where things are going
[11:26] Seth.Nygard Somehow as a community we need to find a way that everyone can work together to get things fixed. Simply filing a Mantis at this stage many people view as a waste of time which is unfortunate
[11:27] Andrew.Hellershanks I keep saying that if you don't file a mantis about a problem there is much less change of it being fixed as no one else may know about the problem.
[11:28] Gavin.Hird Let's be honest with ourself in answering this question: "is there any (organized) progression on 0.9 at all?"
[11:28] Andrew.Hellershanks I know it can be frustrating that an issue may not be dealt with in a timely fashion.
[11:28] Boba Solo: I believe Justin had it correct in his way of the approach, and everyone that operates any simulator realizes it's Borked at this fork
[11:28] Robert Adams: the commercial grids have programmers hacking on their systems but not on core OS... we could use some more volunteer developersion
[11:28] Andrew.Hellershanks Gavin, it is progressing. One might ask in which direction it is moving. ;)
[11:29] Seth.Nygard From what I can see 0.9 is not an evolution from 0.8.2. While it has some things fixed others have taken a step backward.
[11:29] Boba Solo: yupp
[11:29] Seth.Nygard What I do see is more focus on making is work just like Avination and not the greater community in general.
[11:29] Wolf.Freeman i think 09 went more steps forward than 08
[11:30] Gavin.Hird I tend to agree with Seth that a revert to 0.8.2 would be better, and a planned development forward from there
[11:30] Gavin.Hird there are some very good things in 0.9
[11:30] Gavin.Hird others are not good at all
[11:30] Boba Solo: Justins last RC3 is the beauty
[11:30] Seth.Nygard HG travel from 0.8.2 to 0.9 is MUCH worse and that has had a big negative impact on many users
[11:30] Boba Solo: I wish I hadn't went to .9 myself it corruted everything I have worked so hard at for 8 years
[11:30] Seth.Nygard Yes, 0.9 certainly adds some good things so I am not saying to throw it away
[11:31] Gavin.Hird when I say 0.8.2 it is because this so called 0.8.3 only existed as an OSgrid build for a few moments
[11:31] Seth.Nygard But people need to be able to properly test and report on things and that requires some degree of managed chaos at least
[11:31] Zani.Zanadu I think there needs to be more community adhesion towards improving whichever version the majority think best. Instead of us all feeling like we live on islands with no Postal service
[11:32] Boba Solo: well opensimulator had an unoffical 8.3 rc3 for a week from Justin
[11:32] Boba Solo: Well said Zani thx
[11:32] Gavin.Hird which was the one I referred to
[11:32] Boba Solo: I thought you had said 8.2
[11:32] Seth.Nygard the master dev had a 0.8.3 for about a 1 month period of time before the major Avination merges were dumped in. That is what I am still running in most of my grid.
[11:33] Gavin.Hird revert to the 0.8.2 as a starting point
[11:33] Seth.Nygard It is very close to 0.8.2 stable
[11:33] Gavin.Hird even 0.8.2 have issues that needs to be looked at
[11:33] Billy.Bradshaw I believe it is too late to revert, but there is a heavy reliance on people to provide both the resources and commitment to begin serious testing.
[11:34] Seth.Nygard serious testing can only be done when things actually work together, and 0.9 last I checked has a lot more problems with HG visitors than any version of 0.8.2 or 0.8.3
[11:34] Boba Solo: Seth is a real Genius in regards to opensimulator as is may others but perhaps the greatest of minds could think tank this position
[11:35] Gavin.Hird I think that radical changes needs better justification than this is how it was done on Avination
[11:35] Seth.Nygard I know there is no ideal solution, reverting is less than ideal, but trying to test 5500+ commits all at once is also unrealistic
[11:35] Simulator Version v0.5 ruft: OpenSim Dev bcee4e3: 2016-05-23 19:33:36 +0200 (Unix/Mono)
[11:37] Seth.Nygard Does Avination work that perfectly that it even CAN be used as the gold standard for OpenSim?
[11:37] Gavin.Hird for me, 0.9 suddenly became stable when a fix for thread naming was submitted
[11:37] Gavin.Hird It doesn't Seth
[11:37] Zani.Zanadu agreed Gavin, Melanies comment " i removed this as it didnt work for avination " -- this i found very strange to remove/break code because of a personal grid.
[11:37] Gavin.Hird but the worst thing about 0.9 is the total lack of doumentation for expected behavior
[11:38] Gavin.Hird it makes it very hard to test and debug
[11:38] Seth.Nygard That is certainly the biggest issues. How does onyone know what is a bug vs expected behaviour. Too many things changed all at once
[11:38] Zani.Zanadu may i also ask, are we following SL  ? or not ?
[11:38] Billy.Bradshaw As an aside, the developers I work with have expressed concern as to the value of any governance models that may or may not exist. Debugging Opensim and developing patches is expensive.
[11:38] Seth.Nygard A change with no mention of expected behaviour could be considered a bug in itself
[11:39] Seth.Nygard People testing OpenSim put a great deal of time and effort into the project, perhaps even more than the devs in some cases. This is a major investment of time/effort for everyone
[11:40] Gavin.Hird some of the testing takes a long time as issues are only surfacing under certain loads or usage patterns
[11:41] Seth.Nygard I know several creators that have stopped doing anything because of so much uncertainly with things like sit positons and permissions.
[11:41] Billy.Bradshaw It seems in some instances that patches provided via the mantis system have to be promoted before being noticed.
[11:41] Billy.Bradshaw Patches have a shelf life, and core can move rendering patch useless.
[11:41] Wolf.Freeman true Gavin
[11:41] Gavin.Hird I am one of them
[11:42] Gavin.Hird so I started using time on the viewer
[11:42] Robert Adams: LL had it easy -- whatever the code did was 'the standard'
[11:42] Billy.Bradshaw ++Robert
[11:42] Seth.Nygard yes, so what can we as a group consider to be the OpenSim standard?
[11:42] Seth.Nygard that may be the best starting point in all this
[11:43] Gavin.Hird Melanie seems to be of the opinion that everything that works in SL is the standard
[11:43] Andrew.Hellershanks Billy, It is usually not a case of a report not being noticed. The issue is who is available that has the knowledge of OpenSim and the time in which to address the issue.
[11:43] Gavin.Hird which is true up to a certain point, but I see SL starting to do many strrange things
[11:43] Gavin.Hird such as avatar complexity
[11:44] Seth.Nygard The mess that the 0.9 code base is in currently makes that problem much more significant. In many cases there is only 1 or 2 people now that know what was done or why
[11:44] Gavin.Hird which is like treating the symptom and not the patient
[11:45] Robert Adams: I've always thought that OS was SL plus... that is, if it works in SL, it works in OS but there are additions to OS that don't exist in SL
[11:45] Seth.Nygard A lot of the changes were done to add ubODE which is currently every bit as unsupported as ODE itself with Ubit away for so long
[11:45] Robert Adams: that and keeping up with the viewers since that is OS's only option at the moment
[11:45] Zani.Zanadu i thought OS was or had the ability to knock the socks off SL
[11:45] Zani.Zanadu as we were not confined to the same things as SL
[11:46] Gavin.Hird Seth some of the restructuring was good
[11:46] Gavin.Hird but
[11:46] Robert Adams: the ability to modify the simulator (new modules, etc) is what makes OS much better than SL
[11:46] Seth.Nygard I agree that some was good, and necessary
[11:46] Seth.Nygard there are many things in 0.9 I like
[11:46] Gavin.Hird the very significant changes in Scene and entity transfer has borken 0.9
[11:46] Gavin.Hird it is very hard to understand the changes and why
[11:47] Seth.Nygard the problem is that it is currently unmanageable to some degree due to how the merges were handled
[11:47] Gavin.Hird agreed
[11:48] Gavin.Hird I tried to patch up the two I mentioned and never succeded from the commit log
[11:48] Billy.Bradshaw With 0.9 we have in effect started again, making any comparison to earlier versions would be far to complex imo
[11:49] Seth.Nygard The bigger question is now with the mess there is how can we all work together to get it fixed for everyone's benefit?
[11:49] Seth.Nygard Can we establish what the expected behaviours are supposed to be?
[11:49] Gavin.Hird one single change in i.e scene presence could imply 70-80 other changed files all over the place and it was impossible to unwind it
[11:49] Robert Adams: It is an opportuntity to actually define the HG protocols... it's kind of a mashup and has always been... the merge showed many of the weak spots
[11:50] Seth.Nygard ++Robert
[11:50] Zani.Zanadu won't there come a time when the mash ups are just a mess up and no way back ?
[11:50] Seth.Nygard HG in its current state I consider a prototype or proof of concept. It works despite itself in some cases
[11:51] Gavin.Hird there is a Prize for that
[11:51] Seth.Nygard HG has become a cornerstone feature for OpenSim and there is an expectation now that it works. MANY people rely on it daily
[11:51] George Equus: A starting point might be to stop ALL additions, new features, new technologies (such as ubODE) and focus entirely on structuring current problems and try to fix those first.
[11:52] George Equus: One at a time, not several,
[11:52] Seth.Nygard I would agree. Fixes are critical and not the shiny new features
[11:52] Robert Adams: for sure... HG is probably the most important feature to get working
[11:52] Seth.Nygard In a controlled and managed manner
[11:52] Billy.Bradshaw Well if you look at the recent commit to fix landmark, folder should be set using m_Scene.InventoryService.GetFolder, yet what looks like a hack... if folder null then folder = m_Scene.InventoryService.GetFolderForType(remoteClient.AgentId, (FolderType)invType);
[11:53] Zani.Zanadu May i ask one thing, who decides what goes and gets cut ? I mean if you fix something can someone just remove it if it doesnt please them ( or their grid )
[11:53] Billy.Bradshaw kind of a blue print for HG
[11:53] George Equus: I see no point in building on something that is fundamentally messy. Cleaning up the mess first then add one thing at a time, let that sit and test it.
[11:53] Seth.Nygard I for one would rather put my time into fixing 0.9 than my own fork from 0.8.3. But I also need something that is manageable and testable across various situations.
[11:53] Seth.Nygard Code cleanup, while no fun benefits everyone in the long run.
[11:54] George Equus: I am no coder but seem rational to me to make this very structured
[11:54] Gavin.Hird stablizing 0.9 and then fixing it is probably a better approach
[11:54] Seth.Nygard Who does decide what is "correct" is perhaps a key question
[11:54] George Equus: Yes, I think so too
[11:54] Boba Solo: Agree
[11:55] Zani.Zanadu if much work is done, and one person can just mess it all up ,, we need to know
[11:55] George Equus: correct" is a functional feature
[11:55] Boba Solo: Agree
[11:56] George Equus: Make HG work, focus on that single issue, nothing else eg.
[11:56] Billy.Bradshaw The only contentious issues with fixes is when then relate to behavior specifically, all other issues should win on the basis of logic imo
[11:56] Seth.Nygard Keep in mind that with HG there is an expectaion of users that things work the same everywhere they go.
[11:56] George Equus: when that is in order, move on to next major issue
[11:56] Gavin.Hird True Seth, but we have grids such as OSgrid running Hypergrid protocol version 1
[11:57] Andrew Hellershanks: Seth, that is a problem when different grids are running different versions of OpenSim.
[11:57] Gavin.Hird when the release is on version 2
[11:57] Gavin.Hird that alone makes a lot of issues
[11:57] Seth.Nygard A lot of the HG methods are due ot the viewer limitations. I am not however convinced that can not change and benefit everyone. Perhaps this is our opportunity to do that.
[11:57] Seth.Nygard Yes, and until 0.9 works better with HG there is a major problem with that
[11:58] Gavin.Hird To blame viewer limitation is a bit of a stretch
[11:58] Boba Solo: it's a tree top issue----I had the osgrid's safari club visit ---they could HG but then staying or moving or escape DOA was a huge issue for me ....I was embarrassed
[11:58] Robert Adams: inter-version compatability is a BIG problem... sometimes cleaning up or fixing needs moving ahead... that's not easy to make happen
[11:58] Gavin.Hird nobody have described how and what they want the viewer to do
[11:58] Seth.Nygard I am not blaming viewer limitation, but have repeatedly heard that used as a reason for many decisions
[11:59] Boba Solo: agree
[11:59] Seth.Nygard I think it is possible to work WITH the viewer devs cooperatively
[11:59] Gavin.Hird it could be but none the less true
[11:59] Gavin.Hird well of course Seth :-)
[11:59] George Equus: Isn't the viewer second to the main os code here? No point in developing a viewer with so many unknowns around
[11:59] Gavin.Hird how else would a viewer get developed?
[12:00] Billy.Bradshaw ++George
[12:00] Boba Solo: I believe if wasnt miss reading but Justin was working or wanted to work with CtrlAltStudio viewer
[12:00] Seth.Nygard They are symbiotic, one does not exist without the other
[12:00] Gavin.Hird the current viewers are about 25k files in the source tree and very complicated
[12:01] Seth.Nygard yes, and add to that they rely heavily on SL code libraries and constantly change
[12:01] Robert Adams: we do need some viewer help (for HG and grids in general) and for any new features (like varregions)... but, in general, we should work with them to find problems and work them out
[12:01] George Equus: Say, just for the sake of discussion here, that Firestorms latest viewer (or any viewer that work well today) is used and all fixes are tested with that viewer.
[12:01] Gavin.Hird we are in the process of separating the viewer development into an OS version and one SL version
[12:01] Gavin.Hird because it is almost impossible to maintain a unified version any more
[12:01] Robert Adams: and the viewer developers are focusing on their users and SL
[12:02] Gavin.Hird Firestorm has said they want to focus on SL
[12:02] Seth.Nygard I am concerned that we are going to see OpenSim suppport in the viewer degraded as SL remove support for some older protocols and methods that we still use
[12:02] Gavin.Hird they have about 450k users on SL and maybe 10k on OS
[12:02] George Equus: Not sure but think heard that current FS is the Last to be developed for OS?
[12:03] George Equus: no more adaptations for OS
[12:03] Seth.Nygard We need to be working with the viewer devs to try to ensure ongoing support for OpenSim, that only is possible with cooperation from both sides
[12:03] George Equus: focus 100% on SL
[12:03] Zani.Zanadu any viewer developer would take us far more seriously if we showed a united front,
[12:03] Gavin.Hird As I have mentioned before the code changes to support SL in the viewers right now are very substantial
[12:03] Seth.Nygard If that is true then we perhaps need a true OpenSim viewer, or at least find one solid TPV that we can work with
[12:04] George Equus: My perception too
[12:04] Gavin.Hird they have more or less ripped the guts out of it and replaced it with something else and UDP support will be turned off completely shortly
[12:04] Seth.Nygard A new viewer is a major project and one that requires a very high skill level in multiple areas
[12:05] George Equus: Why I propose to choose a current, well working viewer to be used as main viewer during code cleanup
[12:05] Billy.Bradshaw Would be good to have a viewer developers pay a visit and explain at least the extent of changes. free cookies or something
[12:05] Seth.Nygard working with the TPVs is our best choice
[12:05] Gavin.Hird just to rewrite the current functionality into something more sensible is in the order of 10 manyears
[12:05] George Equus: then, later build on that for a new OS focused viewer
[12:05] Seth.Nygard offer them cake even if cookies are not enough, just get them to the table
[12:05] Gavin.Hird for Kokua we are now baselining it for opensim
[12:05] George Equus: I realize the cahllange here...
[12:05] George Equus: very very time consuming
[12:06] Gavin.Hird then will start to modify it for a better opensim viewer
[12:06] Robert Adams: we don't have enough coders
[12:06] Billy.Bradshaw I believe viewer developers are willing to chat, just some relationships have been strained.
[12:06] Seth.Nygard It is all very time consuming for everyone; devs, testers, reporters, creators alike
[12:06] Zani.Zanadu we need to attract coders and devs back and new ones
[12:06] Robert Adams: drama has gotten in the way... Internet people are so touchy
[12:07] Gavin.Hird it gets even more time consuming when expected behavior isn't documented :-))
[12:07] Billy.Bradshaw ++Robert
[12:07] Seth.Nygard they can be and VR seems to make that even worse in many cases
[12:07] Zani.Zanadu the written word you cannot hear me laugh or see me cry --- so sit on your hands and read a second time
[12:07] Seth.Nygard defining expected behaviour/features I think is critical for anyone to be able to move forward
[12:08] George Equus: YES
[12:08] Gavin.Hird absolutely
[12:08] Boba Solo: agree
[12:08] Robert Adams: my personal project is, over the next few weeks, to do a diff on the big merge and then go through it by hand
[12:08] George Equus: First and foremost stop adding features of any kind
[12:08] Boba Solo: Thank you Robert
[12:09] Gavin.Hird make that next few months Robert
[12:09] Boba Solo: many thanks for all you do Kind Sir
[12:09] Seth.Nygard Having already done some of that exercise that is more than a couple weeks work Robert. It is a mess.
[12:09] Robert Adams: stay tuned for where I'll publish my progress... opensim-dev is a start, but there are some people one G+
[12:09] Robert Adams: are there other social places OS people hang out?
[12:10] Seth.Nygard It seems to be a mix of inworld, G+ and skype for most of my contacts
[12:10] Boba Solo: some USED to hang out on my Grid
[12:10] Zani.Zanadu maybe we need a special one for moving forward dedicated
[12:10] vegaslon plutonian: definitely is alot of people on G+
[12:10] Boba Solo: but it has been a trash heap for long time now
[12:10] Billy.Bradshaw G+ Opensim Technical is one I know
[12:10] Seth.Nygard Some of our parties on Refuge Gird become tech sessions
[12:10] Gavin.Hird G+ is not ideal for documenting and progessing the code
[12:11] Billy.Bradshaw Other feeds appear to focus more on commercial issues
[12:11] Robert Adams: on Skype, I'm 'mistermisterblue'
[12:11] Seth.Nygard I think there are a lot of people that want to see this work properly, many I believe are willing to help but need to see some cooperative progress before they jump in
[12:11] George Equus: Wouldn't be a good central point where all go?
[12:12] Gavin.Hird wikis are impossible to use for maintaining a project like this
[12:12] vegaslon plutonian: not alot of people look at the osgrid wiki since it has not been updated for so long
[12:12] Boba Solo: osgrid is a personal grid now
[12:12] Seth.Nygard That is a problem with all the wikis, they are out of date
[12:12] Boba Solo: many don't look see on osg anymore
[12:13] George Equus: That page is pretty brand new, and up to date
[12:13] Zani.Zanadu A new community hangout seems needed where everyone can be heard
[12:13] Boba Solo: agree
[12:13] Gavin.Hird remember OSGrid is not the same as the opensimulator code
[12:13] George Equus: An effort to bring things up to date in one place
[12:14] Seth.Nygard yes, OSG does not run all the current core modules
[12:14] Boba Solo: most of osgrid is except a select few
[12:14] George Equus: Whatever, my point being, a web page central for all dev
[12:14] Billy.Bradshaw WE need to communicate with people who have the resources to run and test 0.9 grids
[12:14] Boba Solo: agree
[12:14] Seth.Nygard those few still cause issues and complicate testing
[12:14] Zani.Zanadu agree
[12:14] George Equus: not G++, Facebook, whatever
[12:14] Gavin.Hird set up a proper project on Atlassian
[12:14] Zani.Zanadu George++
[12:15] Seth.Nygard While I can run a 0.9 grid it can't be my main grid until things are more cleaned up. That makes it less useful.
[12:15] Zani.Zanadu setting up a new space would also enable a real headcount and allow those lost in the past to return
[12:15] Boba Solo: Agree
[12:16] George Equus: Yes
[12:16] Seth.Nygard There is no one platform everyone of us use except OpenSim itself
[12:16] Boba Solo: yes
[12:17] Billy.Bradshaw I have both Linux and windows grids running 0.9
[12:17] Billy.Bradshaw I don't care if they are damaged
[12:17] Billy.Bradshaw actually using the new windows nano server
[12:17] Boba Solo: and possibly a natural site could be in place for all like OPENSIMULATOR OFFICAL DECK
[12:18] Gavin.Hird I run a straight 0.9 dev on mono with one of the supported databases
[12:18] PcTek CyberStar: the oracle rad database is free.
[12:19] Billy.Bradshaw we like to hack, so always running the latest etc
[12:19] PcTek CyberStar: if you got a delphi programming environment.
[12:19] Seth.Nygard Let's not throw a new DB in the mix at this stage
[12:19] Gavin.Hird it already is in the mix Seth as supported by the release
[12:19] Seth.Nygard no shiny new features until things are fixed and on track
[12:20] Gavin.Hird it is not something I have come up with
[12:20] George Equus: And a web page also have the benefit of being available to everyone on the planet without a need for special arrangements of any kind.
[12:20] Zani.Zanadu i still am unsure who decides on the latest version
[12:20] Boba Solo: agree
[12:20] Zani.Zanadu and if one person adds to core who say YEah or NO
[12:20] Gavin.Hird Melanie does
[12:21] Zani.Zanadu why ?
[12:21] Zani.Zanadu who made her chief ?
[12:21] Seth.Nygard We barely have support for Postgresql even though it is a better fit than MySQL. Adding yet another is not such a good idea IMHO
[12:21] Boba Solo: seems to me like Melanie also
[12:21] Gavin.Hird either by passive aggressive behavior or by active submission for what she wants there
[12:21] Boba Solo: folks have piggy backed with her signing off on it
[12:21] Seth.Nygard May I ask then why Melanie is never seen at these meetings?
[12:22] Zani.Zanadu LOL ok now seriously I mean you're NOT serious
[12:22] Boba Solo: EXACTLY
[12:22] Seth.Nygard It is critical to have the key devs involved
[12:22] Boba Solo: YES
[12:22] Andrew Hellershanks: It depends on the type of changes. Some things just get put in to the code base by someone who has the perms to push changes to the repository. Other changes will get discussed amongst the core developers.
[12:22] PcTek CyberStar: oh
[12:22] PcTek CyberStar: you don't have to add support for interbase... interbase is the delphi opensource free database that is very small and fast and been around for years.
[12:22] Zani.Zanadu so we are all dominated by a grid that adds what they want and removes what they want ( not just since ubode )
[12:23] Seth.Nygard I for one would like to see ALL discussions public so everyone can be involved if they chose to be
[12:23] Billy.Bradshaw With the exception of Robert's stuff, we are now able to fix problems that arise during our usage. but we are prepared to assist with testing if required.
[12:24] Seth.Nygard As I have said, I am willing to help. I would rather my time/effort benefit everyone than just my own needs
[12:25] Seth.Nygard We need a community focus for features and expected behaviours
[12:25] Boba Solo: Agree
[12:26] Zani.Zanadu maybe start with something like a good forum for starters for communication.
[12:26] Gavin.Hird I can help too but it is not possible to make any firm commitment till I know how my health will progress
[12:26] Robert Adams: but where is that community?
[12:26] Sheera.Khan At least the meetings here are documented ;-)
[12:26] Robert Adams: in the past, it was often on the IRC channels
[12:26] Andrew Hellershanks: I'll send an email to the other core developers and see if I can get some discussion going amongst them so we can determine where we are, who is still working on code development, and what we need to do to move forward.
[12:26] Seth.Nygard People need to know where things are going. If they have to change then that has to be explained and justified to the community as a whole
[12:26] George Equus: One individual being able to add/remove code by him/her self unsupervised seem like a good receipt for a royal mess...
[12:26] Zani.Zanadu :) sounds like a good start andrew :)
[12:26] Zani.Zanadu George++
[12:27] Seth.Nygard +^1024 George
[12:27] PcTek CyberStar: should it be a vote of all devs ?
[12:27] George Equus: Maybe this is where the current code suffer from... initially
[12:27] PcTek CyberStar: and how will it be secured
[12:27] Seth.Nygard There has to be some degree of democracy I think
[12:27] Andrew Hellershanks: Seth, that is the nature of open source. People get the ability to push code changes once they have shown they can be trusted with the ability to push code. If someone pushes something others don't like someone else in the project can start a discussion about the change or revert the changes.
[12:28] Seth.Nygard I know the nature of OpenSource very well. There are however many very successful projects that manage it well.
[12:28] Gavin.Hird that is the nature of the beast Andrew
[12:28] Andrew Hellershanks: I need to log off. Been over an hour and this viewer is maxing one of my CPU cores and pushing its temperature up higher than it should be.
[12:28] Seth.Nygard Cooperation is mandatory for everyone for it to work
[12:29] George Equus: and a clear leader
[12:29] Zani.Zanadu seems time the community picked the baby back out of the bathwater
[12:29] Seth.Nygard A leader, but never with absolute power
[12:29] PcTek CyberStar: andrew run it with -cooperative 50
[12:29] George Equus: A "main coordinator" if you will
[12:29] Andrew Hellershanks: how do I do that?
[12:30] Billy.Bradshaw Opensim has dna from other opensim forks, developers need fun not regimented, kills innovation
[12:30] PcTek CyberStar: like... coolvlviewer.exe -cooperative 50
[12:30] PcTek CyberStar: or like... singularity.exe -cooperative 50
[12:30] Andrew Hellershanks: The viewer doesn't usually push my CPU this high. It is just doing it now.
[12:30] Seth.Nygard It is possible to have fun and still work with guidelines
[12:30] Andrew Hellershanks: Is that a documented command line option for viewers?
[12:30] PcTek CyberStar: yes
[12:30] Seth.Nygard otherwise it is a mess and breaks down
[12:31] Andrew Hellershanks: See you all next week.
[12:31] Billy.Bradshaw well the methods currently in place seem to be failing, or is that not what people are saying
[12:32] PcTek CyberStar: yes but is there any evidence of sabotage? any evidence of malicious change?
[12:32] PcTek CyberStar: nope, just attempts to make it better.
[12:32] Seth.Nygard I would hope there is no malicious intent from anyone
[12:32] Zani.Zanadu misguided maybe
[12:33] PcTek CyberStar: and the versioning system is there to keep back up copies of copies...
[12:33] Seth.Nygard But when you see changes for one specific grid it raises questions from many people, that is what custom forks are for - the specific needs
[12:33] PcTek CyberStar: maybe assigning power to a group or committee, is just a powershift ?
[12:34] Zani.Zanadu i admit concern when i read " i removed because it broke my grid "! ( no thought for others grids at all )
[12:34] Gavin.Hird :-))
[12:35] Gavin.Hird that's the spirit!
[12:35] Seth.Nygard I think is it important for changes to detail not only what was done but why and the expected behaviour
[12:35] PcTek CyberStar: good idea
[12:35] Simulator Version v0.5 ruft: OpenSim Dev bcee4e3: 2016-05-23 19:33:36 +0200 (Unix/Mono)
[12:36] Seth.Nygard I know then I make a change why I did it and what I expect to happen as a result

Personal tools
About This Wiki