MantisBT - Issues
Updated: 44 min 13 sec ago
- A region module with a ChatFromWorldEvent handler was not picking up messages from llRegionSay() script events in an object.<br /> - Other scripted objects listening on the appropriate channel were reacting to those messages.<br /> - The same ChatFromWorldEvent handler was picking up messages from scripted llSay() events.<br /> <br /> llSay(), llShout() and llWhisper() in LSL_Api.cs all seem to trigger a ChatFromWorldEvent via the World.SimChat() call, which is absent from the llRegionSay() method.<br /> <br /> Copying the following line to from llSay() to llRegionSay() yielded the behaviour I was expecting (i.e. the event handler picking up the message):<br /> <br /> World.SimChat(Utils.StringToBytes(text), ChatTypeEnum.Region, channelID, m_host.ParentGroup.RootPart.AbsolutePosition, m_host.Name, m_host.UUID, false);
Since the change from abort to co-op script terminations I have been seeing a lot of these errors. They seem to be triggered by scripts inside attached objects and when the avatar logs out, there is a flood of these errors on the console. Doesn't happen every time and I'm not sure if its totally limited to scripted attachments.
llSetPrimitiveParams() is ignored after llSetTextureAnim(FALSE)
Destination simulator console when HG teleporting reports [LLCLIENTVIEW] exception (Additional information). Only a few objects within the destination scene are visible to the visiting avatar.<br /> <br /> Commit 41f2f3132bdcbfb8020c7fd6e5f3b7e48c75b1cf 2014-09-05
Hi,<br /> this is my 1st post/interaction with the OS community, so please be gentle... ;-)<br /> related to posts (i think):<br /> <a href="http://opensimulator.org/mantis/view.php?id=6877">http://opensimulator.org/mantis/view.php?id=6877</a> [<a href="http://opensimulator.org/mantis/view.php?id=6877" target="_blank">^</a>]<br /> <a href="http://opensimulator.org/mantis/view.php?id=3274">http://opensimulator.org/mantis/view.php?id=3274</a> [<a href="http://opensimulator.org/mantis/view.php?id=3274" target="_blank">^</a>]<br /> <br /> When using the CtrlAltStudio viewer, and the Oculus Rift, going into RiftLook mode causes the behaviour quoted in the summary- ie when walking forward, if you look down at the floor your avatar jumps.<br /> <br /> this doesn't happen in 1st person mode / non RiftLook mode.<br /> <br /> I've been in contact with the developer of the viewer, and he can reproduce the problem. However it doesn't happen in SL under the same circumstances.<br /> <br /> I asked the developer of the viewer to investigate further, and this is his most recent response:<br /> <br /> "I instrumented the code in places where it initiates a jump and none of them were called, i.e., the viewer’s not misbehaving. What this means is that it’s an OpenSimulator problem and would need to be fixed in its code. "<br /> <br /> I have no reason not to believe him.<br /> <br /> How do we move forward with this?<br /> Thanks in advance,<br /> dan
One of my residents uses a in-region-only "sit position" teleport system, where you place them around the region, set their location in the Description field and reset them through their menu, so that they link up with each other and become visible on each other's menus. To use them, you left click on it, select your destination from the menu, then left click on it again to go there.<br /> <br /> The worked perfectly on the code I last built on 7/22. They are broken on HEAD. I was testing with two of them in a region with no neighbors, with one in the NW corner of the sim and one in the NE corner. When they work, they land you on the destination pad. When they fail, they throw you much too far. For example, when I am in the NE and try to go to the NW pad, I either get slammed against the west border or dropped way out past the west border - where there is not actually a region. The same going the other way - either slammed against the east border or put way out past it where there is no region.<br /> <br /> Using 'git bisect' (my first time using it - it's really useful! and quick!) it determined that this is the culprit:<br /> <br /> [ste@sa5 opensim]$ git bisect bad<br /> a02dae566ce510e7f43c6324996e70442cba3f46 is the first bad commit<br /> commit a02dae566ce510e7f43c6324996e70442cba3f46<br /> Author: Justin Clark-Casey (justincc) <<a href="mailto:email@example.com">firstname.lastname@example.org</a>><br /> Date: Sat Aug 30 00:26:18 2014 +0100<br /> <br /> Fix issue where moving a seated avatar would not adjust their subsequent stand position.<br /> <br /> Addresses <a href="http://opensimulator.org/mantis/view.php?id=7299">http://opensimulator.org/mantis/view.php?id=7299</a> [<a href="http://opensimulator.org/mantis/view.php?id=7299" target="_blank">^</a>]<br /> <br /> :040000 040000 3742d1531521620d1e4964b19927f331f236b76f 3b98963fbe685dc673bc7fa08f12e43b0c6f7be3 M OpenSim<br /> [ste@sa5 opensim]$
These are the IAR file patches I was talking with justincc about on friday night.<br /> <br /> 1)Adds the --perm flag for IAR file creation. Works same way as saving oar files. I assume permmission checks are not as complex for inventory because the permissions of a prims contents etc are folded to the current permmisions of the inventory item when the item is taken into the users inventory. I also included checking the modify permission in the switch, I notice oar files do not check that.<br /> <br /> 2) Simple change to the exisiting event callback for saving IAR files so that the refernce to the console command is a UUID and not a Guid.<br /> <br /> 3) Add new callback event for completion of IAR loads so that region modules can subscibe for notification when the task is complete. I also included some extra code to count how many items were loaded and skipped (due to the perm flag from patch 1), this is then passed to the completed event. I felt it was important to notify the user of any items filtered out, then while i was adding that i included the inventory items loaded. The current code only states how many assets and not items.
Avatar fails to sit sometimes on prim when you do ground sit (ctrl-alt-s in singularity) Instead its going to sit on the prim its moveing out of camera view to ground level / terrain. This happens not always. sometimes the avatar sit on the prim.<br /> <br /> Opensim 0.7.6 release.
if llSetTextureAnim or llSetLinkTextureAnim are called with mode == FALSE (0), the animation should be removed. Instead, it was just being modified. I have provided a patch which handles the removing of animated textures better.
On my own add-on regions using the standard add-on region distribution from osgrid.org for recent versions of 0.8.0 dev master I have avatars which always appeared as a cloud (using Firestorm 4.6.1 if that is relevant).<br /> <br /> The outfit worn was not loading properly also and no alternative outfit could be worn as a replacement as the system thought the avatar had no outfit set. It reported:<br /> <br /> "Could not put on outfit. The outfit folder contains<br /> no clothing, body parts, or attachments"<br /> <br /> I found that replacing the basic body parts then let the avatar rez.. but then the clothing and attachments that should have been worn were also added and they were not detachable. <br /> <br /> I found that teleporting to OSGrid Wright Plaza (for example) allowed the avatar to rez, and logging in there initially was okay each time.<br /> <br /> I tried turning off the default xbakes setup for OSGrid add on regions, but on or off made no difference.<br /> <br /> I am posting this issue to see if anyone else can report similar problems or start to pin down what might be happening.
# mmap(...PROT_NONE...) failed<br /> Stacktrace:<br /> <br /> <br /> Native stacktrace:<br /> <br /> mono() [0x4916ba]<br /> /lib/x86_64-linux-gnu/libpthread.so.0(+0xfcb0) [0x7feb3ffc4cb0]<br /> /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x35) [0x7feb3fc2b425]<br /> /lib/x86_64-linux-gnu/libc.so.6(abort+0x17b) [0x7feb3fc2eb8b]<br /> mono() [0x5d90b9]<br /> mono() [0x5daace]<br /> mono() [0x5ea8d0]<br /> mono() [0x5d0e75]<br /> mono() [0x5d1065]<br /> mono() [0x5d1486]<br /> mono() [0x5d150e]<br /> mono() [0x4e6568]<br /> mono() [0x5bc832]<br /> mono() [0x5dcc1d]<br /> /lib/x86_64-linux-gnu/libpthread.so.0(+0x7e9a) [0x7feb3ffbce9a]<br /> /lib/x86_64-linux-gnu/libc.so.6(clone+0x6d) [0x7feb3fce93fd]<br /> <br /> Debug info from gdb:<br /> <br /> <br /> =================================================================<br /> Got a SIGABRT while executing native code. This usually indicates<br /> a fatal error in the mono runtime or one of the native libraries<br /> used by your application.<br /> =================================================================<br /> <br /> Aborted (core dumped)
In order to easily interact with in-world environment, it would be interesting having the possibility to send keystrokes from a client to the server, for example enabling a sort of "keybard mode".<br /> Mouse clicking (or equivalent shortcut keys) is very limited when users need to interact with more objects at the same time. For example, playing a virtual musical instrument or controlling a control panel in-world, where more objects might need to be clicked at the same time, could be achieved mapping keystrokes to in-world events.<br /> Of course this can be done through http, building an app ad-hoc that connects with an object using a connection url, but I think a direct way to send keystrokes or codes in general would simplify the process.<br /> I may be unaware of techniques already existing, though.
I created a small hud with llGetVel() in order to get the speed of movement of my avatar. It works very well.<br /> Then I installed my avatar on a single vehicle and there I found that my HUD was not working. llGetVel() simply returns 0 all the time. So it does not work when the avatar is on a vehicle.<br /> I have been on SL for my test and normal HUD with a vehicle and it works fine.<br /> I therefore concluded that there is a bug about it in OpenSim.
I tried using llGetObjectDetails worn as HUD and i found it will NOT work 100%.<br /> <br /> I try with:<br /> OBJECT_POS is ok <br /> OBJECT_ROT remains 0 <br /> OBJECT_VELOCITY remains 0<br /> <br /> I tested these 3 parameters in SL and all works fine.<br /> So there is also a bug with llGetObjectDetails in OpenSim.
If an NPC is sitting when a osNpcMoveToTarget() is issued for that NPC, the script will throw an error: <br /> <br /> "Runtime error:<br /> (0): Object reference not set to an instance of an object"
Attempting to compile a current OpenSim + DSG fails with 16 errors.<br /> <br /> OpenSim git hash 3a643e24560a8200dd99a855d1cc7ed1d3b4b9c5 : [r/25116]<br /> DSG git hash cbdc0713a53f1b9933201e8a106e22762864d31b<br /> <br /> The last DSG commit was November 13, 2013<br /> The first change to OpenSim that would have caused a fatal compile error was on Nov 29, 2013 with additional changes in April 2014.<br /> <br /> My attached patch is ugly and untested.<br /> <br /> There are 4 changes in the patch:<br /> 1. The "HandleUndeliverableMessage" DSGMessageTransferModule.cs copied to from HandleUndeliveredMessage - this approach is probably all wrong, and it really needs just a rename of "HandleUndeliveredMessage" below it and change from private to public. I left the existing "HandleUndeliveredMessage" alone because I don't know what effect renaming and changing it from private to public that would have elsewhere.<br /> <br /> 2. Scene.RegionInfo.GetOtherSetting() was renamed to GetSetting in April - so a bunch of these needed to be fixed in DSG's RegionSync.cs<br /> <br /> 3. Scene.PresenceInfo.Updated flag/boolean was removed in late Nov 2013.<br /> <br /> 4. SendAppearanceToAllOtherAgents() was renamed to SendAppearanceToAllOtherClients() at some point.