Bug 1809492 Comment 26 Edit History

Note: The actual edited comment in the bug view page will always show the original commenter’s name and original timestamp.

Original comment by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame`? (I guess not)
Revision 1 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame`? (I guess not)
Revision 2 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame`? (This part sounds safe I would say)
Revision 3 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame*`? (This part sounds safe I would say)
Revision 4 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects? Are we sure that we find the correct `XULTreeElement`?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame*`? (This part sounds safe I would say)
Revision 5 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects? Are we sure that this function finds the correct `XULTreeElement`?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame*`? (This part sounds safe I would say)
Revision 6 by
on 
There is some mechanism in place to try to guarantee that `mTreeBody` outlives the `XULTreeElement`: `nsTreeBodyFrame::DestroyFrom()` will call  `mTree->BodyDestroyed(mTopRowIndex);` if `mTree` is set for that `nsTreeBodyFrame`. That will in turn set `mTreeBody` to `nullptr` for that `XULTreeElement`. This raises some questions:

- Are we sure that `mTree` is always correctly set for any `nsTreeBodyFrame` whose pointer is part of the layout of a `XULTreeElement`? As far as I can tell, `mTree` is set only when we call into `nsTreeBodyFrame::GetBaseElement()`. Do we always call into that function for these `nsTreeBodyFrame` objects? Are we sure that this function finds the correct `XULTreeElement`?
- Are we sure that after `mTreeBody` has been set to `nullptr` following a call to `XULTreeElement::BodyDestroyed`, the next call to `XULTreeElement::GetTreeBodyFrame()` cannot set it back to the same `nsTreeBodyFrame*`?

Back to Bug 1809492 Comment 26