0003029: attach(key )attached) event unrecognized

(0008924)
melanie (administrator)
2009-01-21 14:53

Because attachments have no persistent state in opensim, the handling is different.

(0008925)
mohax (reporter)
2009-01-21 14:55

Does this mean there are no plans to implement the attach() event as handles in Second Life currently? If so, just need to know so we can work around.

(0008926)
melanie (administrator)
2009-01-21 14:59

There are no immediate plans for compatibility. Scripts in OpenSim will have to account for the fact that script state and data is lost and scripts restart from the beginning on each attach.

(0008927)
mohax (reporter)
2009-01-21 15:01

I wonder if just an on_rez() event with more primitive llGetAttached(ATTACH_RHAND) might cover it. Will test.

(0008928)
mohax (reporter)
2009-01-21 15:11

The following work around tested successfully in ReactionGrid. Primitive, but gets the job done at least for event handling that normally would have happened when first wearing/attaching an item.

default
{
    state_entry()
    {
        llSay(0, "Work around for missing attach() event handler");
    }

    on_rez(integer _param)
    {
        if (llGetAttached() == ATTACH_RHAND)
        {
            llSay(0, "Attached to right hand.");
        }
    }
}

(0008953)
mohax (reporter)
2009-01-22 11:12

I can only assume the reason attach(key attached) event handling is being omitted is because the 'attached' key is part of the attachment state. Still couldn't we call the attach() handler without that property? Most just use it to indicate an attachment has been worn or removed. It seems ignoring that requirement would be unnecessary, even against the OpenSim goals of SL compatibility.

(0012521)
greg0254 (reporter)
2009-07-19 06:16

Since no one has posted any comments here since January, I thought I would throw my two cents in and say I would love to see the attach event implemented, mainly for reasons of compatibility and code portability. I am just hoping, with all of the changes and refactoring that have taken place since January, that maybe it has become more do-able. If not...well, all thing considered, we're still doing OK.

(0012522)
melanie (administrator)
2009-07-19 06:54

We need ROBUST to do it. attach() goes hand in hand with proper attachment state handling, which we can't have until we have overwritable private assets, which ROBUST provides API and functionality for.Without that, attach() makes no sense.

Issue History
Date Modified	Username	Field	Change
2009-01-21 14:23	mohax	New Issue
2009-01-21 14:23	mohax	SVN Revision	=> 0
2009-01-21 14:23	mohax	Run Mode	=> Grid (Multiple Regions per Sim)
2009-01-21 14:23	mohax	Physics Engine	=> ODE
2009-01-21 14:23	mohax	Environment	=> Mono / Windows
2009-01-21 14:23	mohax	Mono Version	=> None
2009-01-21 14:53	melanie	Note Added: 0008924
2009-01-21 14:55	mohax	Note Added: 0008925
2009-01-21 14:59	melanie	Note Added: 0008926
2009-01-21 15:01	mohax	Note Added: 0008927
2009-01-21 15:11	mohax	Note Added: 0008928
2009-01-22 11:12	mohax	Note Added: 0008953
2009-01-25 13:32	idb	Relationship added	duplicate of 0002837
2009-03-17 09:22	BeatrixNewt	Issue Monitored: BeatrixNewt
2009-07-19 06:16	greg0254	Note Added: 0012521
2009-07-19 06:54	melanie	Note Added: 0012522

Notes
(0008924) melanie (administrator) 2009-01-21 14:53	Because attachments have no persistent state in opensim, the handling is different.

(0008925) mohax (reporter) 2009-01-21 14:55	Does this mean there are no plans to implement the attach() event as handles in Second Life currently? If so, just need to know so we can work around.

(0008926) melanie (administrator) 2009-01-21 14:59	There are no immediate plans for compatibility. Scripts in OpenSim will have to account for the fact that script state and data is lost and scripts restart from the beginning on each attach.

(0008927) mohax (reporter) 2009-01-21 15:01	I wonder if just an on_rez() event with more primitive llGetAttached(ATTACH_RHAND) might cover it. Will test.

(0008928) mohax (reporter) 2009-01-21 15:11	The following work around tested successfully in ReactionGrid. Primitive, but gets the job done at least for event handling that normally would have happened when first wearing/attaching an item. default { state_entry() { llSay(0, "Work around for missing attach() event handler"); } on_rez(integer _param) { if (llGetAttached() == ATTACH_RHAND) { llSay(0, "Attached to right hand."); } } }

(0008953) mohax (reporter) 2009-01-22 11:12	I can only assume the reason attach(key attached) event handling is being omitted is because the 'attached' key is part of the attachment state. Still couldn't we call the attach() handler without that property? Most just use it to indicate an attachment has been worn or removed. It seems ignoring that requirement would be unnecessary, even against the OpenSim goals of SL compatibility.

(0012521) greg0254 (reporter) 2009-07-19 06:16	Since no one has posted any comments here since January, I thought I would throw my two cents in and say I would love to see the attach event implemented, mainly for reasons of compatibility and code portability. I am just hoping, with all of the changes and refactoring that have taken place since January, that maybe it has become more do-able. If not...well, all thing considered, we're still doing OK.

(0012522) melanie (administrator) 2009-07-19 06:54	We need ROBUST to do it. attach() goes hand in hand with proper attachment state handling, which we can't have until we have overwritable private assets, which ROBUST provides API and functionality for.Without that, attach() makes no sense.