Switch specification doesn't match implementation.

[lrhn]

Implementations of switch currently throw an exception if a case (except the last one) reaches the end of the case body, i.e., where someone might mistakenly expect a fallthrough.

The specification doesn't say that this should happen, but instead just lets execution continue after the switch.

There doesn't seem to be consensus on which behavior is correct.

[iposva-google]

Gilad, please clarify.

---

cc @kasperl.
Set owner to @gbracha.

[gbracha]

I expect we'll change the spec shortly; the open question has been "to what?".

---

Added Accepted label.

[gbracha]

Issue #2300 has been merged into this issue.

[gbracha]

Issue #2047 has been merged into this issue.

[DartBot]

This comment was originally written by kajmagn...@gmail.com

---

+1 for not requiring "break;"

Change the spec to what? What about this:

case 1: ... code ...
case 2: ... code ...
case 3: ... code ...
case 4,
case 5,
case 6: ... code ...

Since 4 & 5 are followed by ',' and are empty, they fall through to case 6:.

(And perhaps ``case X:'' followed by nothing would be an error)

Kind regards, KajMagnus

[gbracha]

Fixe din 0.09 draft

---

Added Done label.