| • Esanders | |
| Apr 26 2019, 1:48 PM |
| F28885500: image.png | |
| May 1 2019, 1:09 PM |
In ve.ce.Surface.prototype.annotationsAtModelSelection we move the cursor one to the right, and use getNodeAndOffset to find the inner annotation, then traverse the DOM to find all surrounding annotations.
The technique fails when the annotation is one character wide, and is generally quite hacky.
We can easily gather annotation models just by looking at the linear data, so there should be a way to do this for the view.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T221967 Can't reliably gather list of annotation views from a DM offset | |||
| Open | None | T228006 Inserting special character after link links that character, too | |||
| Open | None | T228396 'Change label' and label preview in link context don't work after keyboard navigating on 1 or 2 character links | |||
| Resolved | ppelberg | T227241 'Change label' and label preview in link context don't work after clicking on 1 or 2 character links |
A side effect of this is if you cursor onto a link from the right, the preview doesn't render:
I recall poking around for this in the past, and we've been fairly good at avoiding the model knowing about its views. Which is sensible design, but inconvenient at times like this.
That's correct, but the pattern for nodes is view.getNodeFromOffset( modelOffset ) which doesn't violate that rule, so we should have something similar for annotations.
@dchan Perhaps we need a new data structure to keep track of annotations and their offsets?