Mesh/Archive/2012-10-15

From Second Life Wiki
Jump to navigation Jump to search

Agenda

  1. Why does a pathfinding character's LI go up to 15 by default? Is the formula posted anywhere? (In my example, the PF character is a simple cube primitive.) Ashasekayi Ra 10:57, 1 October 2012 (PDT)
  2. If I shift + copy the same 15 LI pathfinding character, the LI of the copy is 1. Is this expected behavior? (In my example, the PF character is a simple cube primitive.) Ashasekayi Ra 10:58, 1 October 2012 (PDT)
  3. @Ashasekay - see https://jira.secondlife.com/browse/PATHBUG-194 --MartinRJ Fayray 12:11, 1 October 2012 (PDT)
  4. Is anyone else having issues similar to sh-3055 WRT uploading meshes?
  5. We've found it's possible to patch the viewer to accept mesh rigged against attachment points with spaces in the name. The patch is targeted specifically at uploading mesh rigged to the problematic bones. Could this be flagged for inclusion into the official viewer? It will give character creators some 23 more bones to rig and animate, and will work with the current codebase (it's just a patch to the upload). See here for details: http://blog.machinimatrix.org/topic/help-needed-to-test-viewer-patch/ --Magus Freston 10:45am 15 Oct
  6. Is there a reason why scripted multiple seating no longer works on mesh objects? I have tried this with two different scripted systems and someone else has also reported a similar issue (SVC-8210 ); while there is a workaround to link in a normal prim as root, this is not ideal since 1) a lot of content is now broken 2) there are potential LI changes as a result of this addition. Is this something the Mesh team can look in to? and if not, who can help fix this problem? Aki Shichiroji 00:02, 14 October 2012 (PDT)

Meeting Notes

Transcript for Monday October 15, 2012

[12:04] Bazz Fazz (vir.linden) howdy all
[12:04] Nyx Linden greetings all
[12:04] ѕиіρі ⁴ˉ³ˉº (sniper.siemens) Hi nyx
[12:05] xstorm Radek NYX i have come to pick on you
[12:05] Yuzuru (yuzuru.jewell) Hello Nyx.
[12:05] ÄlveKatt (davido.chrome) Hello.
[12:05] Nyx Linden today's agenda is here, feel free to add topics: http://wiki.secondlife.com/wiki/Content_Creation/Mesh_Import_User_Group
[12:06] Nyx Linden shall we get started?
[12:06] Nyx Linden first topic: "Why does a pathfinding character's LI go up to 15 by default? Is the formula posted anywhere? (In my example, the PF character is a simple cube primitive.) Ashasekayi Ra 10:57, 1 October 2012 (PDT)"
[12:07] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) heya prep
[12:07] Prep Linden yoyo folks
[12:08] Asha (ashasekayi.ra) /me wonders why it was so hard to get in here when it isn't that crowded. :/
[12:08] Asha (ashasekayi.ra) Heya Prep
[12:08] Aki Shichiroji Hi Prep
[12:08] ÄlveKatt (davido.chrome) Check SW corner.
[12:10] Nyx Linden when you set an object to be a pathfinding character, its physics weight increases to 15, due to the increased load on the physics system
[12:11] xstorm Radek scripts on
[12:11] Nyx Linden I'll request that the documentation gets updated appropriately, I don't think the docs have caught up to that change just yet
[12:11] xstorm Radek oooppsss
[12:12] Nyx Linden next topic (related): "If I shift + copy the same 15 LI pathfinding character, the LI of the copy is 1. Is this expected behavior? (In my example, the PF character is a simple cube primitive.) Ashasekayi Ra 10:58, 1 October 2012 (PDT)"
[12:13] Nyx Linden related: see the bug posted in topic 3 :)
[12:14] Nyx Linden "@Ashasekay - see https://jira.secondlife.com/browse/PATHBUG-194 --MartinRJ Fayray 12:11, 1 October 2012 (PDT)"
[12:14] Asha (ashasekayi.ra) Yes, thanks. :)
[12:14] Nyx Linden If there aren't questions/comments around this, I'll move on to topic 4 (making record time!)
[12:14] Asha (ashasekayi.ra) No questions from me
[12:14] Nyx Linden next topic: "Is anyone else having issues similar to sh-3055 WRT uploading meshes?"
[12:15] Darien Caldwell yes :p
[12:15] Prep Linden https://jira.secondlife.com/browse/SH-3055
[12:15] Darien Caldwell i've actually been working with Nicky Dasmijn on a debug viewer to try to pinpoint the issue. But so far, we haven't found anything conclusive.
[12:16] Nyx Linden the bug as reported is a general symptom of a number of possible underlying issues, the logs attached probably have more detail
[12:16] Darien Caldwell the requests that fail seem identical to the requests that succeed.
[12:16] Darien Caldwell we know that much
[12:16] Darien Caldwell but not being able to see the server side make debug very difficult
[12:17] Nyx Linden and I see both example assets as well as log files, so hopefully we have the information needed to hunt down the issue.
[12:18] Darien Caldwell my pet theory it's the CURL is still my leading theory
[12:18] Darien Caldwell given that 3.2.4 works flawlessly, the issue has to be on the client side
[12:19] Nyx Linden curl can be fickle, we've been doing what we can lately to identify and hunt down issues around that library, I haven't been able to dig deep enough to know if that's the cause though
[12:20] Darien Caldwell /me nods
[12:20] Darien Caldwell it's not something I look forward to digging into
[12:20] Darien Caldwell trying everything else first
[12:20] Nyx Linden how prevalent is this error condition?
[12:20] Darien Caldwell it's fickle, thats a good word
[12:20] Nyx Linden and are the logs reporting the same error every time?
[12:20] Darien Caldwell well, that's the problem, there is no error
[12:21] Darien Caldwell either you press Calculate, or Upload, and it just waits, forever
[12:21] Darien Caldwell or sometimes it works
[12:21] Nyx Linden the log has a status code, and lacks additional information - does it give the same report back each time
[12:21] Nyx Linden ?
[12:21] ÄlveKatt (davido.chrome) Packet loss?
[12:22] Darien Caldwell no it doesn't show any failure at all. it sends the fee request for instance, and waits for a reply back,
[12:22] Darien Caldwell but nothing ever comes
[12:22] Darien Caldwell or maybe it does
[12:22] Darien Caldwell but it never acts like it gets it
[12:22] Darien Caldwell thats where it's hard to debug
[12:23] Darien Caldwell we don't know if the server gets our request, or not
[12:23] Darien Caldwell or if the response is sent, or not
[12:23] Nyx Linden ok
[12:23] Nyx Linden and retrying the request sometimes works, sometimes does not?
[12:23] Darien Caldwell right
[12:23] Nyx Linden are there specific versions that work better/worse?
[12:23] Darien Caldwell viewer 3.2.4 always works
[12:24] Darien Caldwell the problems started with 3.2.5, which was when the multi-threaded CURL was added
[12:24] Darien Caldwell at that time a 408 error was being returned, but some other curl fixes has been done since that elimited that issue
[12:24] Darien Caldwell now it's just nothing
[12:25] Nyx Linden ok I'll let our devs who were working in curl know
[12:25] Darien Caldwell ok
[12:25] Nyx Linden next topic: "We've found it's possible to patch the viewer to accept mesh rigged against attachment points with spaces in the name. The patch is targeted specifically at uploading mesh rigged to the problematic bones. Could this be flagged for inclusion into the official viewer? It will give character creators some 23 more bones to rig and animate, and will work with the current codebase (it's just a patch to the upload). See here for details: http://blog.machinimatrix.org/topic/help-needed-to-test-viewer-patch/ --Magus Freston 10:45am 15 Oct"
[12:26] Darien Caldwell Since Magus isn't here, I"ll take this one too. I've been working with him on it
[12:26] Liborator Hello Oz
[12:26] Darien Caldwell A limitation of the attachment points in the LL character is that many of them have names with spaces, like “Left Pec”.
[12:27] Darien Caldwell Collada 1.4 doesn’t handle bone names with spaces as space is used to delimit bone names.
[12:27] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) the creator of the pathc would need to make a ptach, put it into a bitbucket repo applied to a cloned version of viewer-dev, and file a jira stating what it does, along with the patch and a link tot he bitbucket repo
[12:27] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) thats currently the main way for LL to take in user patches
[12:28] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) for large pathces, sending it to code review wouldnt hert either
[12:28] Darien Caldwell it's pretty small
[12:29] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) yeah, i generally dont send them to code review myself
[12:29] Nyx Linden I'd have to look at exactly how we use that joint map to make sure it was safe/complete, but would be nice if we can add more bones for rigging with such a simple patch :)
[12:30] Darien Caldwell yeah, it seems to work pretty well, and the best thing is, unpatched viewers can see the results as is.
[12:30] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) oh hi Oz
[12:30] Geenzian Scientist (geenz.spad) more bones for rigging is good!
[12:30] Darien Caldwell it's only needed to help the uploader assign
[12:30] Asha (ashasekayi.ra) It worked really well for me.
[12:30] ÄlveKatt (davido.chrome) Sounds scary, like a workaround that makes it problematic to put in a real fix later because it will break content that uses the workaround.
[12:30] Asha (ashasekayi.ra) The test viewer with the patch added that is
[12:31] Nyx Linden makes sense. Would we need to update any published avatar files to make it easier to rig against those bones (correcting names)?
[12:31] Darien Caldwell perhaps. the names just need to have underscores instead of spaces
[12:32] ÄlveKatt (davido.chrome) Weren't those bones defining clickability areas?
[12:32] Asha (ashasekayi.ra) At this time, you either have to make the armature yourself with the list of attachment points or use Avastar. So, it won't cause issues with anything out there anyway.
[12:32] Nyx Linden ok thanks for the report. and the patch, tankmaster is correct on the patch submission process - we do have those mechanisms in place to make sure contributions are properly reviewed & credited
[12:33] Darien Caldwell okay, I"ll relay that info to Magus
[12:33] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) and if they take it in, you get a free laser and t-shirt! :P
[12:33] Asha (ashasekayi.ra) :D
[12:33] Darien Caldwell lol
[12:33] ÄlveKatt (davido.chrome) Bzooon
[12:33] Iva (ivanova.shostakovich) And a full pack of bears.
[12:34] Nyx Linden next topic: "For Oct 15 meeting: Is there a reason why scripted multiple seating no longer works on mesh objects? I have tried this with two different scripted systems and someone else has also reported a similar issue (SVC-8210 ); while there is a workaround to link in a normal prim as root, this is not ideal since 1) a lot of content is now broken 2) there are potential LI changes as a result of this addition. Is this something the Mesh team can look in to? and if not, who can help fix this problem? Aki Shichiroji 00:02, 14 October 2012 (PDT)"
[12:34] Aki Shichiroji In the space of time when i posted that last night and now, i've found a solution
[12:35] Aki Shichiroji at least for my own experience of the issue
[12:35] Aki Shichiroji i do not know if there are similar issues not related to sim-edge sculpts or nearby large sculpts in general.
[12:36] Nyx Linden what fixed the issue for your case?
[12:36] Aki Shichiroji I will consult with the others who noted they were having issues and see if they had similar oversized sculpt proximity
[12:36] Aki Shichiroji removing a sim edger
[12:36] Aki Shichiroji even though it was set to phantom
[12:36] Darien Caldwell its the old issue of being in the bounding box of a megaprim. plays havoc with sit targets
[12:37] Aki Shichiroji and with my products, i try not to have many pieces to link together
[12:37] Aki Shichiroji so... i end up not having enough spaces for sit targets
[12:37] Aki Shichiroji ultimately i'm going to have to link in some dummy prims and sit targets
[12:37] Aki Shichiroji which is not ideal
[12:38] Aki Shichiroji but if the problem has been around for a while... do we have much choice but to use workarounds?
[12:39] Aki Shichiroji i guess i'm wondering whether this has gone from 'bug' to 'expected behavior' territory.
[12:39] Nyx Linden did you see Maestro's comment (most recent) on the JIRA?
[12:40] Aki Shichiroji ah, no i hadn't
[12:40] Nyx Linden it was within the past 24 hours, so no worries there. for those following along at home: "Hi Cory, I think I see the problem. Since you only have 1 prim for your object, there can only be 1 sit target. This means that any avatar after the first does not have a sit target set, and is relying on the raycast method to find a sit position.
[12:40] Aki Shichiroji he suggests that the geometry not being three-dimensional being the problem...
[12:41] Aki Shichiroji but my issue occurs with fully enclosed, non-plane objects.
[12:41] Iva (ivanova.shostakovich) Can a sit target be changed and then sat upon by a second avatar?
[12:41] Nyx Linden ok, so there are some issues with the raycast operation we have to do for multi-sit prims. Could we get some example meshes?
[12:42] Aki Shichiroji I can send you a copy of the item in-world? or did you mean via .dae?
[12:42] Nyx Linden either would work
[12:43] Aki Shichiroji oh sec
[12:43] Aki Shichiroji that's the wrong one
[12:43] Nyx Linden .dae would be best if you have an object that you can post to the jira
[12:43] Nyx Linden (obviously nothing you're hoping to sell)
[12:43] Aki Shichiroji it's a product that's on the market already. can it be submitted privately?
[12:43] Aki Shichiroji ah
[12:43] Aki Shichiroji yeah, it's been out for a while now.
[12:45] Nyx Linden I can probably pass the in-world version to whomever ends up working on the bug, a .dae emailed would be useful as well, but not necessary.
[12:45] Aki Shichiroji ok
[12:46] Nyx Linden that wraps up the listed topics so far. Anyone else have questions or issues this week?
[12:47] Iva (ivanova.shostakovich) Are the minutes going to be posted for all the most recent meetings?
[12:47] Nyx Linden yes, apologies for getting behind on that. I hope to catch up soon
[12:47] Iva (ivanova.shostakovich) Ok, thanks. :)
[12:47] Liborator Can you tell me why phantom does not work now
[12:48] Gramma Fiddlesticks apologize if this was alreayd asked, but I got here late. May I ask if there is any update on the status of the deformer?
[12:48] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) qarl is still being a pain....
[12:48] Nyx Linden Liborator - could you elaborate? I have not personally seen a reported bug around phantom, but I haven't watched all the submitted bugs recently
[12:48] Gramma Fiddlesticks Ok!
[12:49] Nyx Linden I do not personally have an update on the deformer project
[12:49] Gramma Fiddlesticks thanks Tank.
[12:49] Gramma Fiddlesticks thanks Nyx
[12:49] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) oh, nyx... server side baking update?
[12:49] Liborator well there used to be personnel phantom that stoped people pushing you
[12:50] Darien Caldwell ah the viewer hack
[12:50] Liborator I know it was on phonix
[12:50] Darien Caldwell the one where your avatar was phantom but you couldn't move?
[12:50] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) i dont think LL ever had that phantom mode, thats something the Emerald devs made which other TVs incorperated later on
[12:50] Liborator yes Darian
[12:50] Darien Caldwell yeah it was a hack. possible the new havok broke it
[12:51] Liborator ok thank you
[12:51] Aki Shichiroji /me still relies upon the good ol' sittin' on a prim-trick.
[12:51] Nyx Linden Ah, I was not personally familiar with that one, thanks all for helping to answer it :)
[12:51] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) it basicly just stoped the regon from updating your avatar aperence location, but the avatar bounding box would still move
[12:51] Nyx Linden as for server-side texture baking, still plugging along at it :)
[12:51] Gramma Fiddlesticks sit on a prim works every time
[12:52] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) any chance on seeing some of the work by the end of the year? :D
[12:52] Liborator smiles
[12:52] Darien Caldwell is that why Nyx is on a little blue triangle?
[12:52] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) its not a small simple project, i know...
[12:53] Nyx Linden its a complex project with many moving pieces, we'll let you know when there are updates, and I will definitely be asking for beta testers here when we're ready for feedback
[12:54] Iva (ivanova.shostakovich) Bye all, as a fellow resident interested in better creations, thanks for being here.
[12:54] Nyx Linden if there's nothing else for now, we'll wrap up for this week. Thanks everyone for coming out!
[12:55] Lares Carter Thanks for the meeting. :)
[12:55] Aki Shichiroji Thanks for the meeting!
[12:55] ÄlveKatt (davido.chrome) Thank you.
[12:55] Phoenix-FirestormViewer Rocks (tankmaster.finesmith) thx for your time and info, nyx
[12:55] Darien Caldwell no problem. Thanks Linden Minions for your work :)
[12:55] Liborator thanyou
[12:55] Darien Caldwell *giggle*
[12:55] Asha (ashasekayi.ra) Thanks for the meeting, Nyx. :)
[12:55] Liborator have a good day
Retrieved from "https://wiki.secondlife.com/w/index.php?title=Mesh/Archive/2012-10-15&oldid=1173816"