Age | Commit message (Collapse) | Author |
|
Submitting on behalf of Thickbrick Sleaford.
One of the LLSelectNode constructors has a leak where it does "new LLPermisions()" twice, thus leaking the address of the first object created.
This constructor is called every time you interact (click, hover, select, possibly other) with an object, once for each prim in the object. Since sizeof(LLPermissions) is 92 bytes, this can be a significant amount after a while.
I think this might explain VWR-18528 (leaking LLpemissions instances), at least partially.
This was fixed in snowglobe 1.x as part of SNOW-267.
|
|
|
|
back-out the back-out for this branch. yay.
|
|
Backing out this merge that I pushed (prematurely) to the wrong place.
|
|
|
|
|
|
Done:
- 1. Dropped the obsolete "MultipleAttachments" setting.
- 2. Added an "Add" item to the following attachment-related context menus:
* My Appearance (ex-My Outfits) context menu.
* Edit Outfit -> Add More context menu.
* Object in-world context menu.
* Inventory context menu.
* Object inspector gear menu.
- 3. Modified "Attach To / Attach To HUD" to perform the "add" instead of "replace" action.
TODO:
- Ability to attach multiple objects at once from the Add More panel (bulk attach).
- Make sure there's no memleak when you click Wear/Attach in the in-world object context menu
and the callback isn't invoked (because e.g. avatar fails to get close enough to the object).
Issues:
0. I'm not sure whether LLAgentWearables::userAttachMultipleAttachments()
should replace attachments or add them. Assumed the former.
1. I couldn't verify that adding objects from the object inspector menu works
because I either could wear an object or see its inspector, not both.
2. > 1. Right-click on an object in your inventory and select "Wear".
> VERIFY: Attaches the object and replaces whatever's there; asks for
> confirmation before replacing an existing object.
I think this is impossible to implement because we don't know in advance
what point the object will be attached to, so we can't display a confirmation dialog.
Reviewed by Seraph at https://codereview.productengine.com/secondlife/r/843/
--HG--
branch : product-engine
|
|
This #ifdef is now keyed to MultipleAttachments debug setting
|
|
|
|
https://hg.lindenlab.com/seraphlinden/seraph-viewer-for-server-1.40/
|
|
|
|
|
|
|
|
Superficial cleanup to change all typedef std::vector<LLUUID> to use a common typedef uuid_vec_t instead.
|
|
Superficial cleanup to replace all instances of "gAgentAvatar" with "gAgentAvatarp".
|
|
Superficial cleanup to replace all instances of "LLVOAvatarSelf *avatarp = gAgent.getAvatarObject" with "gAgentAvatar".
|
|
Superficial cleanup so that all instances of gAgent.getAvatarObject() use "LLVOAvatarSelf *avatarp = gAgent.getAvatarObject".
|
|
|
|
First check-in; only compiles, nothing more.
|
|
|
|
EXT-4013 : Functionality loss: Inspect object (aka see a list of creators / timestamps for all prims in a linkset)
Simple fix to only use Inspect floater if it's actually active (i.e. visible).
|
|
EXT-4013 : Functionality loss: Inspect object (aka see a list of creators / timestamps for all prims in a linkset)
Very simple fix that addresses some brokenness with the inspect multi-object floater.
Nature of changes is that when we re-enabled the floater (i.e. uncommented the associated code) per last-minute request, a couple of blocks of code had been left commented out.
|
|
|
|
settings
Review #95
This change undoes some prior change that seems to not allow showing
media data for multiple selection. There was all of this code put in
to support displaying tentative state for media, but code was added to
basically override it. Perhaps the reason was tentative state items
were not ignored on apply...this change does that.
|
|
Checker: UNINIT_CTOR
Function: LLSelectMgr::LLSelectMgr()
File: /indra/newview/llselectmgr.cpp
|
|
Checker: UNINIT_CTOR
Function: LLSelectNode::LLSelectNode(const LLSelectNode&)
File: /indra/newview/llselectmgr.cpp
|
|
--HG--
branch : avatar-pipeline
|
|
--HG--
branch : avatar-pipeline
|
|
|
|
Renamed gPie* and menu_pie_* to just gMenu* and menu_* because they are no
longer pie menus. Created separate other-attachment menus.
Reviewed with Richard.
|
|
Object inventory was grabbing edit menu handling during refresh.
Put back Angela's code so that Right-click > Remove > Delete
is now disabled if you have a single face selected. Fixed
LLSelectMgr::contains() for SELECT_ALL_TES case.
Reviewed with Richard.
|
|
May cause right-click delete to fail - testing that
Backed out changeset: 1f8e1e73f9f4
|
|
|
|
|
|
|
|
|
|
Also replaced many duplicate calls to LLViewerCamera::getInstance() with local pointer.
Reviewed with Ambroff
|
|
|
|
Fixed many other includes related to this file.
Cleaned out llfloatertestlistview to be an empty test floater.
|
|
--HG--
branch : avatar-pipeline
|
|
Media focus is still tracked separately from LLSelectMgr (which was a change made during a refactor a while ago), but LLViewerMediaFocus now keeps LLSelectMgr updated with the current media focus when it changes.
Added a special case for media focus to LLSelectMgr::deselectAllIfTooFar(), since it was making the focus ring not show up when media focus was on distant media.
|
|
Using activePanel hack to update selection when it changes.
--HG--
branch : avatar-pipeline
|
|
|
|
|
|
EXT-1810 Right clicking own avatar displayed semi-transparent menus
EXT-1927 Hook up Push-to-talk in Preferences > Advanced
reviewed by richard
|
|
|
|
|
|
floater, and anywhere secondlife:///app/group URLs appear. Reviewed with Leyla.
|
|
svn+ssh://svn.lindenlab.com/svn/linden/branches/avatar-pipeline/avatar-pipeline-multiattachments-3 into svn+ssh://svn.lindenlab.com/svn/linden/branches/viewer/viewer-2.0.0-3
For DEV-32867 : Avatar Pipeline Project - Multi-Attachments
This is a viewer-side-only, infrastructure-only merge for multi-attachments. This should not lead to any new functionality; it simply provides backwards compatibility for when multi-attachments are enabled in viewer2.1, so that viewer2.0 users can see viewer2.1 avatars correctly.
Reviewed by: Vir.
|
|
LLViewerMediaFocus now tracks two separate objects: the currently focused media object, and the media object that's currently being hovered over. It no longer stores smart pointers to either the LLViewerObject or the LLViewerMediaImpl -- it now looks up both by UUID every time they're needed, and fails gracefully if either goes away. This will prevent it from keeping objects from being deleted.
The poorly-understood "mouseOverFlag" has been expunged.
LLViewerMediaFocus no longer uses LLSelectMgr at all. The object to focus on is explicitly passed between LLViewerMediaFocus and LLPanelMediaHUD instead of going indirectly through the selection manager.
LLViewerMediaFocus also no longer interacts with the pick from LLToolPie -- the data it needs from the pick (the object and normal vector) is passed explicitly.
LLViewerMediaFocus::setCameraZoom and LLViewerMediaFocus::getBBoxAspectRatio now have no dependencies on the LLViewerMediaFocus object -- all the data they need is passed in when they're called by the LLPanelMediaHUD. I made them static member functions, but they could be moved to LLPanelMediaHUD or even made into file-scoped static functions. The only reason I didn't do either of those is that it seems like they belong with the LLViewerMediaFocus code as opposed to the HUD.
|