Chat log from the meeting on 2019-05-21

From OpenSimulator

Revision as of 22:40, 30 May 2019 by Sheera Khan (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

[10:59] Kayaker Magic: I'm trying out a Roth body today.
[11:00] Gavin.Hird Where is Roth available from for testing?
[11:00] Leighton.Marjoram how is it
[11:02] Leighton.Marjoram the legs at the hips look a little funky
[11:02] Kayaker Magic: You can get a Roth at RothandRuth island here in OSGrid.
[11:02] Gavin.Hird ok, but is there an unrigged version somewhere too?
[11:03] Kayaker Magic: I was dissapointed to learn that Roth (and Ruth) are not replacements for the system mesh avatar but just rigged mesh you wear, then make your system mesh avatar invisible.
[11:04] Kayaker Magic: Isn't there a cost to load the old system mesh avatar just to paint it transparent?
[11:04] Ubit Umarov: replacements would had to be installed on viewers
[11:04] Ubit Umarov: yeah..LL ugly hack kayaker
[11:04] Gavin.Hird to remove the system avatar would be a pretty drastic rewrite of the viewer
[11:04] Bill.Blight system avatar with no clothing and transparent has a display weight/complexity of around 1000 so not much cost at all ..
[11:05] Gavin.Hird plus it would break all old content
[11:05] Sandy Beachcomber: and making it trans doesnt always work across grids
[11:05] Gavin.Hird remember these new avatars does not have the over 100 morph targets that the system avatar has
[11:05] Kayaker Magic: Well in the mesh upload there are options for "avatar", I foolishly assumed that meant the viewer had code to support loading different avatars.
[11:05] Bill.Blight it works if they have their asset server working correctly and it gets the HG textures right
[11:05] Gavin.Hird it is only for info that
[11:06] Gavin.Hird does not have any practical importance in OS or SL
[11:07] Ubit Umarov: but that should had been the right path to avoid all the load current mesh ones have
[11:07] Kayaker Magic: So when I wear Roth, I'm carrying twice as many vertexes around with me.
[11:07] Gavin.Hird probably 4x as many as the system avatar
[11:07] Ubit Umarov: don't think that hard.. basicly higher resolution meshes basicly
[11:07] Bill.Blight Annnnnd you see why the discussion of overly heavy mesh avatars takes place
[11:07] Gavin.Hird the system avatar is only like 6k
[11:07] Ubit Umarov: that coul be optional acording to machine gpu
[11:08] Sheera Khan: And some wear additional Tattoo-layers and so on - each with the same amount of vertices
[11:08] Bill.Blight yep
[11:08] Ubit Umarov: all layers are textures
[11:08] Gavin.Hird many of the mesh avatars have several layers of mesh
[11:08] Bill.Blight some people are wearing as many as 3 of the mesh avatar just for tattoos and clothing/makeup layers
[11:08] Gavin.Hird exactly Bill
[11:09] Ubit Umarov: yeah mesh clothes also, etc
[11:09] Kayaker Magic: Yeah, I considered making a 4-layer shirt to simulate the tattoo/under/shirt/jacket of the system avatar, but that would be 4x the vertexes of a single Ruth shirt.
[11:09] Bill.Blight Most people don't care how it affects everyone around them and do it anyway
[11:09] Ubit Umarov: that could be just textures
[11:09] Sheera Khan: at least here you _can_ remove unnessessary vertices out of your body
[11:09] Ubit Umarov: and with materials also on basic avatar...
[11:10] Ubit Umarov: that would improve look with not much more load
[11:10] Bill.Blight It's an exclusive club of "Let's lag the sim out, make everyone suffer then blame Ubit"
[11:11] Bill.Blight that was a joke, but more true than I like it
[11:11] Ubit Umarov: well ll decided to go by improments based on incremental hacks
[11:11] Ubit Umarov: understandable.. but bad
[11:12] Ubit Umarov: ...lets add meshes... just replacing the sculpt texture uuid by a mesh uuid
[11:12] Ubit Umarov: .. lets let ppl wear meshs also..
[11:13] Ubit Umarov: etc
[11:13] Kayaker Magic: How come I can't set shiny textures on the mesh shirt I am wearing?
[11:13] Sheera Khan: make them even have an armature and call them AniWhats?
[11:13] Gavin.Hird ask your viewer
[11:14] Ubit Umarov: yeah that was the last one... lets use "avatar animation engine on some objects also"
[11:14] Ubit Umarov: not a proper implementation.. just another hack
[11:14] Sheera Khan: I'm still happy you did that, @Ubit
[11:15] Ubit Umarov: well my part was easy.. the viewer side hack was the hard part :)
[11:15] Kayaker Magic: I'm happy about animesh, although I have had no time to test it yet... :(
[11:15] Kayaker Magic: Ari, is your body Ruth?
[11:16] Arielle.Popstar no, athena
[11:16] Sheera Khan: I've got Ruth 2
[11:16] Gavin.Hird I have added more restrictive mesh upload code to the viewer, and thrown at it just about any mesh I have gotten hold of. It refuses to upload a lot of crappy mesh it allowed before, but all well-formed and well-rigged mesh gets through.
[11:17] Kayaker Magic: nice.
[11:17] Gavin.Hird I have not tested ruth / roth yet, so hence I need to find both rigged and unrigged versions
[11:17] Bill.Blight Gavin, you know Opensimulator is no place for GOOD mesh
[11:17] Gavin.Hird I know :-((
[11:18] Andrew.Hellershanks Excellent, Gavin. If that change gets adopted by a lot of (or all of the) other viewers it will help save grids from continuing to get bad mesh.
[11:18] Sandy Beachcomber: sounds good to me
[11:18] Kayaker Magic: Is there a path for you to submit changes back to the FireStorm project?
[11:18] Bill.Blight FS already had some code for that, but people cried about it, so they added a debug setting to disable it
[11:18] Arielle.Popstar Opensimulator is not a place were we have a wealth of professional graphics artists trying to make a living
[11:18] Andrew.Hellershanks I think I can already hear the complaints from some users about why they can no longer upload a mesh object that used to be able to upload.
[11:18] Sandy Beachcomber: unless peiple resort to old viewer to upload when new fails
[11:19] Gavin.Hird FS has a JIRA announced on their website
[11:19] Sheera Khan: there's a git repository where you can even get the Blender files for Ruth and Roth - I just don't have the address atm :-(
[11:19] Ubit Umarov: fs still refuses a lot even "removing" the new checks
[11:19] Gavin.Hird I'll find my way to them for testing
[11:19] Leighton.Marjoram I think they have a group on MeWe with the links to the repository let me check
[11:20] Kayaker Magic:
[11:20] Leighton.Marjoram
[11:20] Gavin.Hird speaking of the viewer, there will be a (final) version available today before starting on the 64-bit Win builds
[11:20] Ubit Umarov: so the options don't really unlock all uploads ( good but it does refuse 2 much )
[11:20] Andrew.Hellershanks hm... I only found this one ->
[11:20] Gavin.Hird it has a few updates, including a long standing issue with rendering system water
[11:21] Ubit Umarov: in my last upload tests i had to use older singu
[11:21] Kayaker Magic: That is the old one, they moved it when one of the devs (Ingen) quit.
[11:21] Leighton.Marjoram Thats the original repository but Shin has left the project Andrew
[11:21] Gavin.Hird a render uptade to RLV and update of MKRLV to version which is also the latest version from Marine Kelly
[11:21] Arielle.Popstar there is a couple of repositories at least
[11:22] Gavin.Hird also all DARWIN code has been removed from the Windows repository
[11:22] Arielle.Popstar whats darwin code?
[11:23] Sandy Beachcomber: on the crap mesh issue - should the opensim or robust be made to refuse it?
[11:23] Gavin.Hird Darwin is the name of the macOS kernel, so it is used in build files to denounce macOS specific code
[11:23] Arielle.Popstar oh ok
[11:23] Ubit Umarov: opensim only does vey basic tests on uploads
[11:23] Kayaker Magic: I say yes, reject bad mesh, but display good documentation about WHY
[11:23] Sandy Beachcomber: should it do more
[11:23] Ubit Umarov: basicly checks if all parts are there
[11:23] Sandy Beachcomber: it seems a crucial aspect
[11:24] Arielle.Popstar not if you dont want to have less people running it
[11:24] Bill.Blight Seems to me more people with crap mesh is just as bad as less people running it, both ruin the user experience
[11:25] Gavin.Hird it will complain about missing physics model (if not supplied or generated), about errors in material assignment, rigging, and mesh that cannot be verified
[11:25] Bill.Blight Quality over Quantity
[11:25] Andrew.Hellershanks Kayaker, you know that people won't read the docs to learn why. :)
[11:25] Kayaker Magic: RTFM....
[11:25] Arielle.Popstar even if that sort of mesh is already inworld Gavin?
[11:25] Ubit Umarov: what complains about physics model?
[11:25] Sandy Beachcomber: but if os can stop it surely that is a bonus?
[11:25] Gavin.Hird on upload
[11:26] Arielle.Popstar what does that accomplish?
[11:26] Ubit Umarov: physics model is optional ( not convex hull always present )
[11:26] Gavin.Hird it actually does some verification of in world meshes too, so I have seen a few examples on rezz where you nonly get the triangle
[11:26] Bill.Blight viewer still gives the error , "Missing convex shape" or whatever , once in a while
[11:27] Gavin.Hird yes, if a convex model cannot be generated
[11:27] Ubit Umarov: all meshes must have a basic convex hull for physics
[11:27] Gavin.Hird usually seen on grouped mshes
[11:27] Bill.Blight yep
[11:27] Arielle.Popstar will it bypass checks on regions running meshmeriser where there is a default physics map?
[11:27] Ubit Umarov: they do it even without setting anything on physics
[11:27] Gavin.Hird where each member in the group is rigged
[11:27] Ubit Umarov: opensim does refuse meshes without it...
[11:28] Ubit Umarov: ( needed for type convex )
[11:28] Gavin.Hird then you have to upload each rigged ember separately and recombine in-world
[11:28] Kayaker Magic: By documentation, I meant when the viewer refuses to upload your mesh it tells you why so you have a better chance of fixing it.
[11:28] Bill.Blight Why should opensim or FS even consider an unsupported configuration such as using the wrong meshermizer?
[11:28] Arielle.Popstar because opensim is a platform
[11:28] Arielle.Popstar you wasnt that sort of security, code it yourself
[11:28] Arielle.Popstar on your grid
[11:28] Bill.Blight All platforms have a right and wrong way to run them
[11:29] Ubit Umarov: i really don't care about ppl that don't RTFM and if read, thing knew better that what it says
[11:29] Sandy Beachcomber: i would if i could Ari
[11:29] Arielle.Popstar and then you wonder why people arent updating to newer versions of opensim
[11:30] Ubit Umarov: that irrational behavour is just to ignore...
[11:30] Ubit Umarov: or behavior ? :)
[11:31] Bill.Blight That goes both ways, if you want to run an unsupported configuration, don't expect anybody to help you fix it or to even bother listening when you have issues
[11:31] Arielle.Popstar do whatever bgut dont complain when people arent updating to more restrictive viewers and sims
[11:32] Bill.Blight I don't think anybody is complaining that others are not updating, others that are not updated are complaining they don't have the new toys ..
[11:32] Kayaker Magic: Moving on to another topic: What is the feasibility of adding an osTestAvailability(string funcname); that would tell you if you have the thread level and user permissions to run a particular OSSL function?
[11:32] Arielle.Popstar there is plenty of quotes from you Bill that people were so slow to update to .9
[11:32] Gavin.Hird if grids or individuals won't update, it is absolutely their choice
[11:32] Arielle.Popstar it is
[11:32] Kayaker Magic: *threat level
[11:33] Bill.Blight Yes , there are, but I don't care if they do or , do not, but don't want to listen to them complain about "incompatibility" either
[11:33] Ubit Umarov: not that usefull kay :)
[11:33] Kayaker Magic: Right now the only way to find out is to try and get an ANNOYING RED DEBUG WINDOW.
[11:33] Ubit Umarov: the script using that, would still had the tested function, so would fail to compile anyways
[11:34] Ubit Umarov: ( NO not adding conditional compiling !! )
[11:34] Kayaker Magic: I see two kinds of users: Those who change the ini file of their grid to run OSSL functions, and ordinary users who can never use the OSSL functions because most of them fail for most ordinary users.
[11:35] Gavin.Hird Thanks for the Ruth and Roth links Andrew
[11:35] Andrew.Hellershanks Gavin, yw
[11:36] Kayaker Magic: I'd like to write scripts for ordinary users that use OSSL, but for the most part I cannot. A test function would make it possible for my scripts to at least tell users why.
[11:36] Bill.Blight That is another RTFM condition Kay
[11:36] Arielle.Popstar that would be handy
[11:36] Ubit Umarov: as is the script would not compile anyway kayaker
[11:36] Arielle.Popstar such a pita not knowing why not
[11:36] Sandy Beachcomber: i think so too
[11:37] Sandy Beachcomber: kayaker writes scripts for users in different grids#
[11:37] Andrew.Hellershanks It would be possible to have such a function. I wonder if it would be a problem for a grid to have that feature but any user intending ill can find out what functions are enabled/disabled by trial and error.
[11:37] Ubit Umarov: it had to be conditional compile...
[11:37] Arielle.Popstar but that might make it easier for users so that is not the sort of user experience dev's like
[11:37] Bill.Blight As ubit says, the script if using any non-enabled OSSL functions in the open, it would fail to compile, so it would never make it to the testing function ..
[11:37] Bill.Blight or at very minimum fail to run
[11:37] Sandy Beachcomber: it would on some grids
[11:37] Sandy Beachcomber: surely?
[11:38] Ubit Umarov: ie like that impossible to read C++ things
[11:38] Bill.Blight If you try and use a non-enabled ossl function, the script stops at that point
[11:38] Kayaker Magic: Lots of OSSL functions compile, then produce an ANNOYING RED DEBUG WINDOW at run time.
[11:38] Sandy Beachcomber: not good for the end user on a different grid
[11:38] Arielle.Popstar nope
[11:38] Andrew.Hellershanks That is why you would a separate script to test if something was available before you try to use that function in a different script.
[11:38] Ubit Umarov: not if you use new ossl things on old regions
[11:38] Bill.Blight was just getting to that, some functions will work if in a sub-function the script will compile until that function is run ..
[11:39] Sheera Khan: Would a stub telling "Function osXyz() is unsupported due to configuration" help?
[11:39] Kayaker Magic: It would help.
[11:39] Sheera Khan: so all functions would be defined and working, some just would display that message without doing the stuff
[11:39] Bill.Blight it is just a stop gap
[11:40] Gavin.Hird Cant you test upfront in the script if the required OSSL is available, and take a different codepath if not?
[11:40] Bill.Blight even if it tells them it is disabled, they would still have to go enable it, so just as easy to check first
[11:40] Gavin.Hird I seem to remember a TP script that does that
[11:40] Kayaker Magic: No, I want a way for a script to determine if a function works, then quietly change its behavior.
[11:40] Bill.Blight so it does not actually gain anything
[11:40] Selby.Evans A script with an OSSL function could give a notice about threat level limitations and a link to further info:
[11:40] Sheera Khan: it would at least give the user an understandable hint on why it fails and what could be done to prevent it
[11:41] Ubit Umarov: well the try catch of xmr would possible help on that
[11:41] Sandy Beachcomber: the end user may not have access to ini files
[11:41] Sandy Beachcomber: i see where k is coming from
[11:41] Arielle.Popstar thought Freaky tech posted one time about being able to enable ossl functions on the fly without having to go through a restart after enabeling
[11:41] Ubit Umarov: but im not sure Y will have it :)
[11:41] Andrew.Hellershanks Would it be better to just send a message to the script owner if a function is used when the script owner doesn't have perms instead of throwing it in to a debug window with red text that winds up being seen by anyone in the area?
[11:41] Sandy Beachcomber: the script does a work around for no avail funcs
[11:41] Sheera Khan: at the time being the script would fail and the user wouldn't even know why and where to look
[11:42] Bill.Blight well, they would get a red debug telling them that function is not enabled
[11:42] Sandy Beachcomber: thats a poor user experience in my book
[11:43] Bill.Blight sadly people don't read those any more than they read the manual
[11:43] Andrew.Hellershanks Lots of other people see those. I've seen red text when in an area where someone has a bad script.
[11:43] Arielle.Popstar sadly they are not very explanatory
[11:43] Ubit Umarov: yeah that global red ink is a pain
[11:43] Kayaker Magic: If my script could test before crashing with ANGRY RED TEXT, then my script could document why: Display nice error messages to the end user.
[11:43] Sandy Beachcomber: exactly!
[11:44] Sandy Beachcomber: or work around it
[11:44] Andrew.Hellershanks How about a simpler message such as "Sorry, You do not have the permissions required to use the function osBlah".
[11:44] Sheera Khan: and other visitors wouldn't be bothered with that debug message they don't have anything to do with
[11:44] Andrew.Hellershanks Exactly, Sheera.
[11:45] Arielle.Popstar if it says exactly what ossl function(s) need to be enabled great
[11:45] Bill.Blight but I think that is the viewer not having multiple debug channels, and just blasts it out
[11:45] Sheera Khan: I'ld vote for at least getting rid of that red spam ...
[11:45] Andrew.Hellershanks Getting spammed by red text has occasionally been helpful as I report it to the object owner (when I can track down the owner of the offending object).
[11:45] Andrew.Hellershanks Arielle, IIRC, it does say which function failed.
[11:46] Selby.Evans Most people I work with can't figure out why the script gives an error message they don't understand.
[11:46] Sandy Beachcomber: i find it useful if i wrote the script
[11:46] Andrew.Hellershanks nods
[11:46] Arielle.Popstar does function = the ossl line?
[11:46] Sandy Beachcomber: but i too would like to to write scripts that gracefully avoid the red
[11:46] Andrew.Hellershanks Yes, the Allow_os lines.
[11:47] Sheera Khan: nothing against sending the debug message to the owner of the object... but why send it out to everyone?
[11:47] Arielle Popstar: didnt used to do that
[11:47] Gavin.Hird It is actually not sent out to everyone as such
[11:47] Arielle Popstar: its a new thing in some .9 version
[11:47] Andrew Hellershanks: It is sent to the debug channel which everyone sees.
[11:48] Selby.Evans If I give out a script with an OSSL function, I give the recipient a link to my article
[11:48] Arielle Popstar: i dont remember seeing global ones in .8.2
[11:48] Selby.Evans (If I remember)
[11:48] Andrew Hellershanks: Its difficult to remember that far back but I think it has been that way for a very long time.
[11:49] Kayaker Magic: Or maybe people are finally trying to use those functions?
[11:49] Arielle Popstar: or is there a viewer setting to turn it off for global ones?
[11:49] Kayaker Magic: And discovering that most cannot use them.
[11:49] Selby.Evans It gives a red error message in 8.x
[11:49] Andrew Hellershanks: I think it depends a little on the viewer being used as to how it responds when those red text debug messages are spit out.
[11:49] Sheera.Khan listening to a (debug) channel should be able to filter messages according to owner or other users
[11:49] Gavin.Hird there probably is Arielle
[11:49] Gavin.Hird I'll have to test it
[11:52] Kayaker Magic: Bill: Tell me how your Robust ban by MAC works. New INI settings? A database?
[11:52] Bill Blight: ini settings
[11:52] Arielle Popstar: i been noticing that when i log into Lbsa or other plaza that the FS Quick Preference Hover Height does not work. But it will work after i teleport elsewhere
[11:52] Kayaker Magic: You need to change the INI and restart to add one more user to the ban list?
[11:52] Andrew Hellershanks: I noticed the settings today while I was upgrading a grid.
[11:54] Bill Blight: yes kay
[11:54] Arielle Popstar: same goes for HG jump to a new grid, it and ao's and huds don't work until after a jump to a different region within the same grid
[11:55] Bill Blight: I have not tried the config set , to see if it works on the fly, but robust starts up so fast that it is not really an issue
[11:55] Kayaker Magic: there are terrible race conditions with scripts in HUDs while HG jumping.
[11:56] Gavin.Hird Things you want to work on HG transfers needs to be in My Suitcase
[11:56] Kayaker Magic: Sometimes the script arrives uncompiled, sometimes it arrives "not found in the database".
[11:56] Arielle Popstar: osgrid doesnt use suitcase
[11:56] Arielle Popstar: but even so, it still works on a different region after a tp
[11:56] Gavin.Hird does not wrongly use my suitcase
[11:57] Kayaker Magic: Similar problems with HUDs can happen just crossing a region border. the script arrives in a bad state.
[11:57] Arielle Popstar: seems like something wrong with the initial login
[11:57] Gavin.Hird having it in my suitcase increase the likelihood of it working
[11:58] Bill Blight: It is an issue, more so, if you are testing yEngine and jump to an xEngine region or vice versa ...
[11:58] Arielle Popstar: this is x to x
[11:58] Arielle Popstar: no doubt Y will compound the problem
[11:58] Ubit Umarov: do not edit scripts on attachments.. you will loose them..
[11:59] Ubit Umarov: ( on HG )
[11:59] Arielle Popstar: not what i am bringing up
[11:59] Kayaker Magic: For a working HUD script to survive region crossing or HG jump, the script has to be sent and then it has to be compiled. Sometimes the compile message arrives before the text of the script. It never recovers.
[11:59] Arielle Popstar: there was no editing of scripts
[11:59] Arielle Popstar: so it needs to be throttled?
[12:00] Arielle Popstar: to slow it down?
[12:00] Sandy Beachcomber: i remember jim tarber explaining region crossings to me once
[12:01] Bill Blight: Internet weather and processing speed of the region can impact that a lot as well
[12:01] Kayaker Magic: Some kind of syncronization: If you are told to compile a script tht is not there, schedule it for another try.
[12:01] Ubit Umarov: for security all scripts are compiled on tps ( or partially at least )
[12:01] Arielle Popstar: it seems consistent
[12:01] Sandy Beachcomber: i was amazed anything ever crossed a border
[12:01] Kayaker Magic: LOL
[12:01] Ubit Umarov: but there is also a script state sent, with current variables etc..
[12:01] Sandy Beachcomber: tf for big vars
[12:02] Gavin.Hird even SL seems to be out on a limb with region crossings these days
[12:02] Ubit Umarov: all that will may fail if regions are of diferent versions..
[12:02] Arielle Popstar: aside from that the Hover height in quick prefs is a viewer thing, not a script
[12:02] Leighton.Marjoram Nods @Sandy
[12:02] Gavin.Hird will may? :-)
[12:03] Ubit Umarov: h,, X -> Y tps/crossings should work better now... not Y -> X
[12:03] Gavin.Hird is that in the category of a might do loop?
[12:03] Ubit Umarov: my car can cross fine from X to Y, stops on Y to X
[12:03] Ubit Umarov: ( my test car on my test regions )
[12:06] Arielle Popstar: any information on the mailing list Andrew?
[12:07] Kayaker Magic: Hmm, if Andrew were not AFK he might have ended the meeting already....
[12:08] Gavin.Hird it got very quiet
[12:08] Andrew Hellershanks: Anything on the mailing list re: which topic?
[12:08] Arielle Popstar: probably all in IM
[12:08] Sandy Beachcomber: lol
[12:08] Sandy Beachcomber: i was gonna ask about version 10?
[12:08] Gavin.Hird oh, in any case look for a new Windows viewer build at in about an hour
[12:08] Andrew Hellershanks: You've been talking about scripts, border crossing, and hover height.
[12:08] Arielle Popstar: i asked you last week about the nabble part of mailing list not being updated
[12:09] Arielle Popstar: you said you would look intoit
[12:09] Andrew Hellershanks: Arielle, I don't know about a nabble part of a mailing list.
[12:10] Arielle Popstar: are devs no longer communicating through the mailing list?
[12:11] Andrew Hellershanks: Arielle, Sometimes we do and sometimes we use IRC.
[12:11] Arielle Popstar: nothing since Feb 15 on the mail list?
[12:11] Bill Blight: nobody has posted
[12:12] Andrew Hellershanks: Talk to the people who maintain that archive to see if they have a problem.
[12:12] Arielle Popstar: who does the mailing list then?
[12:13] Andrew Hellershanks: Arielle, the official archive of the developer mailing list is here ->
[12:14] Andrew Hellershanks: It is linked to from the Mailing Lists page on the wiki pages at
[12:16] Arielle Popstar: well someone in core is responsible for it
[12:17] Andrew Hellershanks: I will have to ask Melanie if we are still archiving the mailing list on the machine. I don't have access to it.
[12:17] Ubit Umarov: not much emails on mail lists

Personal tools
About This Wiki