This makes the code a bit more consistent and lets us avoid some
hacky and probably non-performant things:
- abusing CSS classes in the draw/drag datum functions (classed `.target`)
(is this thing target? just check d.properties)
- regexing the id for `-nope$`
(is this thing a nope target? just check d.properties)
- using context.hasEntity to get a the real entity
(is this thing a real osmEntity? just check d.properties)
- fixes code like the restriction editor which uses fake ids for split ways
Drag start is responsible for switching into drag mode, classing
stuff as `active` and kicking off a bunch of other things.
If the drag move happens immediately after this, and includes the
target from the initial active drag, it can cause weird snapping
from the dragnode to its own parent way. (Happened if the user did
a very fast drag from the node along the parent way just next to it)
There was an issue where the lines did not redraw their targets right away
when entering drag node, which could make it possible for a quick drag node
to try to connect to its parent line. With the chooseEdge exclusion it
would not connect to the parent nearby, but in another weird part of the line.
The goal here is that the code that draws the targets should know better
what parts of the lines/vertices are targetable, rather than just
relying on CSS to ignore the pointer events on the whole line.
e.g. when drawing a line, it's ok for it to loop back and connect
to itself, just not on a segment or vertex adjacent to the active
node.
This allows the label collision boxes for the vertices to update
during these modes. Otherwise the labels.js `isInterestingVertex()`
function doesn't consider them interesting enough to update.
Some highlights
- `getSiblingAndChildVertices` are expensive, so they're saved and called less frequently
- draw touch targets for all the visible vertices
- remove redundant css classes and `setClass` function
This is more work to further isolate the layers that entities draw to.
It makes it easier to debug what is going on, and can eventually lead to
deferred drawing, if each draw function is in its own place and not dependant
on anything else.
I've started to replace the vertex-hover with an explicit layer for touch
targets.
Also had to change a lot of the svg tests, which are really brittle.
Things would happen like - the surface would be created, it would kick of a
deferred redraw, which would notice that the zoom was 0 and call
editOff, which would remove the osm layers that were just created and
that the tests were trying to draw to. These tests need proper zoom and
projection otherwise nothing works.
(re: #4271, #3636)
- better classification of "interesting" vertices
(include tagged, selected, or child of selected)
- now we can draw labels on selected lines again (revert #3636)
because the labels will avoid the vertices
- if debugging is on, draw a collision box for the mouse
