Bugzilla

Here's what seems to be going on, based on the accessibility events Orca gets.

Old way, before the regression:
1. User deleted the selected message.
2. The parent container fired children-changed:removed and removed the deleted item. I assume that associated accessible object was being destroyed.
3. A selection-change and/or focus-change event was fired for the next item in the list that became selected as a consequence of the deletion. Orca presented the newly-selected/focused item.

New way, after the regression:
1. User deletes the selected message.
2. The accessible object associated with the selected message fires a name-changed event changing the name to match the name of the next item in the list that became selected as a consequence of the deletion. In other words, it appears to me that the accessible object is no longer being destroyed like before; it's being repurposed.
3. A selection-change event is fired on this renamed item. But Orca sees that the accessible object firing the selection change is the same accessible object that was selected prior to the deletion. In other words, the selection did not change and this event seems like an error or event spam and Orca ignores it.

While I personally think repurposing accessible objects in this fashion is not ideal, I cannot exactly call it a bug. So I just changed Orca in its main/unstable branch to present the name change. Prior to this change, Orca had been ignoring most name changes from Thunderbird due to user complaints of chattiness. But Orca no longer can do so. That's fine, and it fixes the user experience regression reported in related (duplicate?) issue 1866121.

HOWEVER, making the change I made in Orca has side effects, some of which will be easy to handle, others of which may not. As a result of this, I do not intend to backport the changes into stable versions of Orca. At least not any time soon. We don't yet know what else in Thunderbird might be firing name-changed events that Orca should not present, but now is. 

Jamie: Do you have any opinions on this?

Here's what seems to be going on, based on the accessibility events Orca gets.

Old way, before the regression:
1. User deleted the selected message.
2. The parent container fired children-changed:removed and removed the deleted item. I assume that associated accessible object was being destroyed.
3. A selection-change and/or focus-change event was fired for the next item in the list that became selected as a consequence of the deletion. Orca presented the newly-selected/focused item.

New way, after the regression:
1. User deletes the selected message.
2. The accessible object associated with the selected message fires a name-changed event changing the name to match the name of the next item in the list that became selected as a consequence of the deletion. In other words, it appears to me that the accessible object is no longer being destroyed like before; it's being repurposed.
3. A selection-change event is fired on this renamed item. But Orca sees that the accessible object firing the selection change is the same accessible object that was selected prior to the deletion. In other words, the selection did not change and this event seems like an error or event spam and Orca ignores it.

While I personally think repurposing accessible objects in this fashion is not ideal, I cannot exactly call it a bug. So I just changed Orca in its main/unstable branch to present the name change. Prior to this change, Orca had been ignoring most name changes from Thunderbird due to user complaints of chattiness. But Orca no longer can do so. That's fine, and it fixes the user experience regression reported in related (duplicate?) bug 1866121.

HOWEVER, making the change I made in Orca has side effects, some of which will be easy to handle, others of which may not. As a result of this, I do not intend to backport the changes into stable versions of Orca. At least not any time soon. We don't yet know what else in Thunderbird might be firing name-changed events that Orca should not present, but now is. 

Jamie: Do you have any opinions on this?