BR line-break placeholders are already bridge-compatible

What happened

The next sentinel-strategy question sounded bigger than it actually was:

• do zero-width sentinels stay at all?
• can some cases move to ?

The honest first thing to prove was not “delete FEFF everywhere.”

It was:

• does the slate-dom-v2 bridge already tolerate -style line-break placeholders?

What the proof showed

Yes.

For line-break placeholders, the bridge already works with a zero-width wrapper that contains instead of depending on a sentinel text child.

That is because the bridge logic is marker-shaped, not FEFF-shaped:

• path lookup climbs from the child DOM node to a bound ancestor
• DOM-point resolution descends to the first leaf node under the bound node
• zero-width offset normalization keys off the zero-width marker, not the sentinel character itself

What this means

The larger strategy seam is not “can the bridge tolerate ?”

It can.

The real remaining seam is:

• which DOM node the renderer binds for each zero-width case
• which cases still need FEFF-like sentinel text because of IME/native selection pressure

That is a renderer and input-method policy question, not a DOM bridge tolerance question.

Reusable rule

Before trying to replace zero-width sentinels globally:

• prove whether the bridge already tolerates the alternative DOM shape
• if it does, move the decision upward to renderer and IME policy

Do not waste time “teaching the bridge ” if the bridge already speaks it.