4096 Jump Issue and new OpenSim branches of TPVs

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

4096 Jump Issue and new OpenSim branches of TPVs

aiaustin
I wonder if the 4096 jump limit is something that anyone can address now there OpenSim speciic branches of viewers, such as Firestorm.. or if its an issue that even in such an OS specific branch still cannot it be addressed?
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Michael Emory Cerquoni
I am not any expert on this topic AI, but I suspect to really fix this problem it might require changes on both the viewer and the simulator, but I can not say with any certainty that this is true, I do however suspect that to truly fix this problem it will break old methods, and I think this is probably why no real fix has been put forward yet, hopefully I am wrong about this, noting would make me happier than to be wrong about this!


On Sun, Apr 28, 2013 at 2:10 PM, Ai Austin <[hidden email]> wrote:
I wonder if the 4096 jump limit is something that anyone can address now there OpenSim speciic branches of viewers, such as Firestorm.. or if its an issue that even in such an OS specific branch still cannot it be addressed?
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users



--
Michael Emory Cerquoni

_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Dahlia Trimble
In reply to this post by aiaustin
There was a time when libomv based viewers could teleport beyond 4096 and still function correctly, however  since then someone added some code to OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While this effectively prevented the viewing failures experienced with LL-based viewers, it also prevented anyone from being able to fix the problem from the viewer side. Given that libomv-based clients did not have the problem suggests that it's likely entirely withing the LL codebase.


On Sun, Apr 28, 2013 at 11:10 AM, Ai Austin <[hidden email]> wrote:
I wonder if the 4096 jump limit is something that anyone can address now there OpenSim speciic branches of viewers, such as Firestorm.. or if its an issue that even in such an OS specific branch still cannot it be addressed?
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users


_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Diva Canto
The server-side behavior is configurable -- there's a config var that can be switched so that the simulator doesn't check.
And yes, the issue is entirely due to the LL code base.
Several people have tried, and failed, to fix it. I don't know what the issue is, but it seems to touch on some fundamental design decisions that pervade the code base.

On 4/28/2013 12:59 PM, Dahlia Trimble wrote:
There was a time when libomv based viewers could teleport beyond 4096 and still function correctly, however  since then someone added some code to OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While this effectively prevented the viewing failures experienced with LL-based viewers, it also prevented anyone from being able to fix the problem from the viewer side. Given that libomv-based clients did not have the problem suggests that it's likely entirely withing the LL codebase.


On Sun, Apr 28, 2013 at 11:10 AM, Ai Austin <[hidden email]> wrote:
I wonder if the 4096 jump limit is something that anyone can address now there OpenSim speciic branches of viewers, such as Firestorm.. or if its an issue that even in such an OS specific branch still cannot it be addressed?
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users



_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users


_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

aiaustin
In reply to this post by aiaustin
Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.

I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than ON by default?


On 28 Apr 2013, at 20:59, Dahlia wrote:
> There was a time when libomv based viewers could teleport beyond 4096 and
> still function correctly, however  since then someone added some code to
> OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
> this effectively prevented the viewing failures experienced with LL-based
> viewers, it also prevented anyone from being able to fix the problem from
> the viewer side. Given that libomv-based clients did not have the problem
> suggests that it's likely entirely withing the LL codebase.
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Teravus Ovares
When I reported this on the linden Jira years ago..   I suspected it was probably the viewer failing to compare numbers in scientific notation with numbers that are not in scientific notation.  Unfortunately, it was closed and not fixed..  so here we are :).   One thing to note is the 4096 number seems to be when the meter count flips into scientific notation.  
 
Best Regards

Teravus


On Mon, Apr 29, 2013 at 3:38 AM, Ai Austin <[hidden email]> wrote:
Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.

I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than ON by default?


On 28 Apr 2013, at 20:59, Dahlia wrote:
> There was a time when libomv based viewers could teleport beyond 4096 and
> still function correctly, however  since then someone added some code to
> OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
> this effectively prevented the viewing failures experienced with LL-based
> viewers, it also prevented anyone from being able to fix the problem from
> the viewer side. Given that libomv-based clients did not have the problem
> suggests that it's likely entirely withing the LL codebase.
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users


_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Michael Emory Cerquoni
I recall it being something like that too where it is something like a 16bit integer and when the distance was over 4096 it overflowed and was truncated.


On Mon, Apr 29, 2013 at 2:33 PM, Teravus Ovares <[hidden email]> wrote:
When I reported this on the linden Jira years ago..   I suspected it was probably the viewer failing to compare numbers in scientific notation with numbers that are not in scientific notation.  Unfortunately, it was closed and not fixed..  so here we are :).   One thing to note is the 4096 number seems to be when the meter count flips into scientific notation.  
 
Best Regards

Teravus


On Mon, Apr 29, 2013 at 3:38 AM, Ai Austin <[hidden email]> wrote:
Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.

I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than ON by default?


On 28 Apr 2013, at 20:59, Dahlia wrote:
> There was a time when libomv based viewers could teleport beyond 4096 and
> still function correctly, however  since then someone added some code to
> OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
> this effectively prevented the viewing failures experienced with LL-based
> viewers, it also prevented anyone from being able to fix the problem from
> the viewer side. Given that libomv-based clients did not have the problem
> suggests that it's likely entirely withing the LL codebase.
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users


_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users



--
Michael Emory Cerquoni

_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

justincc
In reply to this post by aiaustin
So that Dahlia knows who to blame, I was the one who created this check.  I was very tired of accidentally teleporting
to very remote regions and screwing up my session and I don't think it's a good user experience.

As other people have said, this restriction is configurable.  One could also have a regular expression whitelist to let
selected viewers to attempt the teleport.

I don't favour changing the default until there are a significant proportion of viewers in use that could handle 4096+
teleports.  As we've seen, we're not even sure what the problem is in the LL viewer codebase and TPVs.

On 29/04/13 09:38, Ai Austin wrote:

> Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.
>
> I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than ON by default?
>
>
> On 28 Apr 2013, at 20:59, Dahlia wrote:
>> There was a time when libomv based viewers could teleport beyond 4096 and
>> still function correctly, however  since then someone added some code to
>> OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
>> this effectively prevented the viewing failures experienced with LL-based
>> viewers, it also prevented anyone from being able to fix the problem from
>> the viewer side. Given that libomv-based clients did not have the problem
>> suggests that it's likely entirely withing the LL codebase.
> _______________________________________________
> Opensim-users mailing list
> [hidden email]
> https://lists.berlios.de/mailman/listinfo/opensim-users
>


--
Justin Clark-Casey (justincc)
OSVW Consulting
http://justincc.org
http://twitter.com/justincc
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

drwhiet@spacefriends.de
Where is the setting to disable it ?
Could not find it yet ?  

Thanks in advance,
Wordfromthe Wise

-----Ursprüngliche Nachricht-----
Von: [hidden email]
[mailto:[hidden email]] Im Auftrag von Justin
Clark-Casey
Gesendet: Montag, 29. April 2013 23:37
An: [hidden email]
Betreff: Re: [Opensim-users] 4096 Jump Issue and new OpenSim branches of
TPVs

So that Dahlia knows who to blame, I was the one who created this check.  I
was very tired of accidentally teleporting to very remote regions and
screwing up my session and I don't think it's a good user experience.

As other people have said, this restriction is configurable.  One could also
have a regular expression whitelist to let selected viewers to attempt the
teleport.

I don't favour changing the default until there are a significant proportion
of viewers in use that could handle 4096+ teleports.  As we've seen, we're
not even sure what the problem is in the LL viewer codebase and TPVs.

On 29/04/13 09:38, Ai Austin wrote:
> Thanks Dahlia. The explicit limit was added by Diva to try to make lie
easier for users/avatars getting stuck in limbo, but can be turned on and
off with an OpenSim.ini parameter, default being that it is on. So it can
clearly be removed. Now there is a distinct branch for some viewers like
Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to
improve the OpenSim experience, some critical issues like zthe rally panful
and awkward to explain to new users 4096 jump issues just might now be on te
cards to be fixed.
>
> I am sure if this were technically possible in commonly used LL viewers
with an Openim specific branch then the OS server side default setting for
the 4096 jump restriction could simply be set as OFF by default rather than
ON by default?

>
>
> On 28 Apr 2013, at 20:59, Dahlia wrote:
>> There was a time when libomv based viewers could teleport beyond 4096
>> and still function correctly, however  since then someone added some
>> code to OpenSimulator to prevent *any* viewer from teleporting beyond
>> 4096. While this effectively prevented the viewing failures
>> experienced with LL-based viewers, it also prevented anyone from
>> being able to fix the problem from the viewer side. Given that
>> libomv-based clients did not have the problem suggests that it's likely
entirely withing the LL codebase.
> _______________________________________________
> Opensim-users mailing list
> [hidden email]
> https://lists.berlios.de/mailman/listinfo/opensim-users
>


--
Justin Clark-Casey (justincc)
OSVW Consulting
http://justincc.org
http://twitter.com/justincc
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users

_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

R.Gunther
In reply to this post by justincc
Dahlia, the 4096 teleport block is good. dont want to mis it. i hate
relogging because a bad teleport.
And i dont ecpect that limit is going to disapear soon. so its really a
must have piece of code.

On 2013-04-29 23:36, Justin Clark-Casey wrote:

> So that Dahlia knows who to blame, I was the one who created this
> check.  I was very tired of accidentally teleporting to very remote
> regions and screwing up my session and I don't think it's a good user
> experience.
>
> As other people have said, this restriction is configurable.  One
> could also have a regular expression whitelist to let selected viewers
> to attempt the teleport.
>
> I don't favour changing the default until there are a significant
> proportion of viewers in use that could handle 4096+ teleports. As
> we've seen, we're not even sure what the problem is in the LL viewer
> codebase and TPVs.
>
> On 29/04/13 09:38, Ai Austin wrote:
>> Thanks Dahlia. The explicit limit was added by Diva to try to make
>> lie easier for users/avatars getting stuck in limbo, but can be
>> turned on and off with an OpenSim.ini parameter, default being that
>> it is on. So it can clearly be removed. Now there is a distinct
>> branch for some viewers like Firestorm, and that there are already
>> improvements in Firestorm 4.4.0(OS) to improve the OpenSim
>> experience, some critical issues like zthe rally panful and awkward
>> to explain to new users 4096 jump issues just might now be on te
>> cards to be fixed.
>>
>> I am sure if this were technically possible in commonly used LL
>> viewers with an Openim specific branch then the OS server side
>> default setting for the 4096 jump restriction could simply be set as
>> OFF by default rather than ON by default?
>>
>>
>> On 28 Apr 2013, at 20:59, Dahlia wrote:
>>> There was a time when libomv based viewers could teleport beyond
>>> 4096 and
>>> still function correctly, however  since then someone added some
>>> code to
>>> OpenSimulator to prevent *any* viewer from teleporting beyond 4096.
>>> While
>>> this effectively prevented the viewing failures experienced with
>>> LL-based
>>> viewers, it also prevented anyone from being able to fix the problem
>>> from
>>> the viewer side. Given that libomv-based clients did not have the
>>> problem
>>> suggests that it's likely entirely withing the LL codebase.
>> _______________________________________________
>> Opensim-users mailing list
>> [hidden email]
>> https://lists.berlios.de/mailman/listinfo/opensim-users
>>
>
>

_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

Dahlia Trimble
In reply to this post by justincc
I wasn't looking for people to blame, but now that you mention it..... *evil grin*

I am a little concerned that a bit of a catch-22 situation has come about with the current solution.  Viewer developers are unable to test in existing grids due to the default setting, and nobody will want to change the setting as it might disrupt the user experience. Any viewer testers will need to overcome the hurdle of learning that the setting exists and setting up a test grid. In the mean time users of alternative viewers are left out in the cold.

 What would you consider "significant"? Personally I'd consider Radegast usage significant and it is not based on LL code. While it's based on libomv and I have had success teleporting beyond 4096 with other libomv based clients, I've not been able to test it in Radegast to date due to the existing default.

Perhaps a protocol extension might be an alternative? Some way that a viewer could signal the simulator that it's capable of distant teleports and the sim could bypass the restriction for such viewers?


On Mon, Apr 29, 2013 at 2:36 PM, Justin Clark-Casey <[hidden email]> wrote:
So that Dahlia knows who to blame, I was the one who created this check.  I was very tired of accidentally teleporting to very remote regions and screwing up my session and I don't think it's a good user experience.

As other people have said, this restriction is configurable.  One could also have a regular expression whitelist to let selected viewers to attempt the teleport.

I don't favour changing the default until there are a significant proportion of viewers in use that could handle 4096+ teleports.  As we've seen, we're not even sure what the problem is in the LL viewer codebase and TPVs.


On 29/04/13 09:38, Ai Austin wrote:
Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.

I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than ON by default?


On 28 Apr 2013, at 20:59, Dahlia wrote:
There was a time when libomv based viewers could teleport beyond 4096 and
still function correctly, however  since then someone added some code to
OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
this effectively prevented the viewing failures experienced with LL-based
viewers, it also prevented anyone from being able to fix the problem from
the viewer side. Given that libomv-based clients did not have the problem
suggests that it's likely entirely withing the LL codebase.
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users



--
Justin Clark-Casey (justincc)
OSVW Consulting
http://justincc.org
http://twitter.com/justincc

_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users


_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users
Reply | Threaded
Open this post in threaded view
|

Re: 4096 Jump Issue and new OpenSim branches of TPVs

justincc
On 29/04/13 22:54, Dahlia Trimble wrote:

> I wasn't looking for people to blame, but now that you mention it..... *evil grin*
>
> I am a little concerned that a bit of a catch-22 situation has come about with the current solution.  Viewer developers
> are unable to test in existing grids due to the default setting, and nobody will want to change the setting as it might
> disrupt the user experience. Any viewer testers will need to overcome the hurdle of learning that the setting exists and
> setting up a test grid. In the mean time users of alternative viewers are left out in the cold.
>
>   What would you consider "significant"? Personally I'd consider Radegast usage significant and it is not based on LL
> code. While it's based on libomv and I have had success teleporting beyond 4096 with other libomv based clients, I've
> not been able to test it in Radegast to date due to the existing default.
>
> Perhaps a protocol extension might be an alternative? Some way that a viewer could signal the simulator that it's
> capable of distant teleports and the sim could bypass the restriction for such viewers?

That would be good, though I'm not sure there's any existing mechanism for a viewer to signal the simulator as to what
features it supports.  Possibly one could do something along the same line as the SimulatorFeatures cap and make an
uploaded set of key:value pairs available in the AgentData.  This would be a better long term solution than adding
viewer names to a configured list.

>
>
> On Mon, Apr 29, 2013 at 2:36 PM, Justin Clark-Casey <[hidden email] <mailto:[hidden email]>> wrote:
>
>     So that Dahlia knows who to blame, I was the one who created this check.  I was very tired of accidentally
>     teleporting to very remote regions and screwing up my session and I don't think it's a good user experience.
>
>     As other people have said, this restriction is configurable.  One could also have a regular expression whitelist to
>     let selected viewers to attempt the teleport.
>
>     I don't favour changing the default until there are a significant proportion of viewers in use that could handle
>     4096+ teleports.  As we've seen, we're not even sure what the problem is in the LL viewer codebase and TPVs.
>
>
>     On 29/04/13 09:38, Ai Austin wrote:
>
>         Thanks Dahlia. The explicit limit was added by Diva to try to make lie easier for users/avatars getting stuck in
>         limbo, but can be turned on and off with an OpenSim.ini parameter, default being that it is on. So it can
>         clearly be removed. Now there is a distinct branch for some viewers like Firestorm, and that there are already
>         improvements in Firestorm 4.4.0(OS) to improve the OpenSim experience, some critical issues like zthe rally
>         panful and awkward to explain to new users 4096 jump issues just might now be on te cards to be fixed.
>
>         I am sure if this were technically possible in commonly used LL viewers with an Openim specific branch then the
>         OS server side default setting for the 4096 jump restriction could simply be set as OFF by default rather than
>         ON by default?
>
>
>         On 28 Apr 2013, at 20:59, Dahlia wrote:
>
>             There was a time when libomv based viewers could teleport beyond 4096 and
>             still function correctly, however  since then someone added some code to
>             OpenSimulator to prevent *any* viewer from teleporting beyond 4096. While
>             this effectively prevented the viewing failures experienced with LL-based
>             viewers, it also prevented anyone from being able to fix the problem from
>             the viewer side. Given that libomv-based clients did not have the problem
>             suggests that it's likely entirely withing the LL codebase.
>
>         _________________________________________________
>         Opensim-users mailing list
>         [hidden email] <mailto:[hidden email]>
>         https://lists.berlios.de/__mailman/listinfo/opensim-users <https://lists.berlios.de/mailman/listinfo/opensim-users>
>
>
>
>     --
>     Justin Clark-Casey (justincc)
>     OSVW Consulting
>     http://justincc.org
>     http://twitter.com/justincc
>
>     _________________________________________________
>     Opensim-users mailing list
>     [hidden email] <mailto:[hidden email]>
>     https://lists.berlios.de/__mailman/listinfo/opensim-users <https://lists.berlios.de/mailman/listinfo/opensim-users>
>
>
>
>
> _______________________________________________
> Opensim-users mailing list
> [hidden email]
> https://lists.berlios.de/mailman/listinfo/opensim-users
>


--
Justin Clark-Casey (justincc)
OSVW Consulting
http://justincc.org
http://twitter.com/justincc
_______________________________________________
Opensim-users mailing list
[hidden email]
https://lists.berlios.de/mailman/listinfo/opensim-users