mirror of
https://github.com/opensim/opensim.git
synced 2026-08-11 20:05:33 +08:00
Add regression TestDetachScriptedAttachmentToInventory()
This currently only does a relatively crude check for a ScriptState node in the serialized xml
This commit is contained in:
@@ -339,7 +339,7 @@ namespace OpenSim.Region.CoreModules.Avatar.Attachments
|
||||
ShowAttachInUserInventory(sp, attachmentPt, newAttachmentItemID, group);
|
||||
}
|
||||
|
||||
public ISceneEntity RezSingleAttachmentFromInventory(IScenePresence sp, UUID itemID, uint AttachmentPt)
|
||||
public SceneObjectGroup RezSingleAttachmentFromInventory(IScenePresence sp, UUID itemID, uint AttachmentPt)
|
||||
{
|
||||
if (!Enabled)
|
||||
return null;
|
||||
@@ -527,6 +527,7 @@ namespace OpenSim.Region.CoreModules.Avatar.Attachments
|
||||
/// </remarks>
|
||||
/// <param name="sp"></param>
|
||||
/// <param name="grp"></param>
|
||||
/// <param name="saveAllScripted"></param>
|
||||
private void UpdateKnownItem(IScenePresence sp, SceneObjectGroup grp, bool saveAllScripted)
|
||||
{
|
||||
// Saving attachments for NPCs messes them up for the real owner!
|
||||
@@ -720,18 +721,23 @@ namespace OpenSim.Region.CoreModules.Avatar.Attachments
|
||||
null, assetID, Vector3.Zero, Vector3.Zero, UUID.Zero, (byte)1, true,
|
||||
false, false, sp.UUID, true);
|
||||
|
||||
// m_log.DebugFormat(
|
||||
// "[ATTACHMENTS MODULE]: Retrieved single object {0} for attachment to {1} on point {2}",
|
||||
// objatt.Name, remoteClient.Name, AttachmentPt);
|
||||
|
||||
if (objatt != null)
|
||||
{
|
||||
// m_log.DebugFormat(
|
||||
// "[ATTACHMENTS MODULE]: Rezzed single object {0} for attachment to {1} on point {2} in {3}",
|
||||
// objatt.Name, sp.Name, attachmentPt, m_scene.Name);
|
||||
|
||||
// HasGroupChanged is being set from within RezObject. Ideally it would be set by the caller.
|
||||
objatt.HasGroupChanged = false;
|
||||
bool tainted = false;
|
||||
if (attachmentPt != 0 && attachmentPt != objatt.AttachmentPoint)
|
||||
tainted = true;
|
||||
|
||||
// FIXME: Detect whether it's really likely for AttachObject to throw an exception in the normal
|
||||
// course of events. If not, then it's probably not worth trying to recover the situation
|
||||
// since this is more likely to trigger further exceptions and confuse later debugging. If
|
||||
// exceptions can be thrown in expected error conditions (not NREs) then make this consistent
|
||||
// since other normal error conditions will simply return false instead.
|
||||
// This will throw if the attachment fails
|
||||
try
|
||||
{
|
||||
|
||||
Reference in New Issue
Block a user