[JEP 325 Switch Expression] switch labeled block completion issue?

[Jan Lahoda]

Hi Manoj,

Please see:
https://bugs.openjdk.java.net/browse/JDK-8212982

Which is marked as fixed in build 21.

Jan

On 28.11.2018 05:19, Manoj Palat wrote:
> Hi Joe, all,
> The following program compiles without errors.
>
> package se2;
>
> import java.io.IOException;
> public class X {
> public static int twice(int i) throws IOException {
> int k = switch (i) {
> case 1 -> {
> System.out.println("hello");
> // break 1;
>    }
> case 2 -> {
> System.out.println("world");
> // break 2;
>    }
>    default ->  10;
> };
> return k;
> }
> public static void main(String[] args) {
> try {
> System.out.println(X.twice(1));
> } catch (IOException e) {
> //do nothing;
> }
> }
> }
>
> $ javac12 --enable-preview -source 12 se2/X.java
> Note: se2\X.java uses preview language features.
> Note: Recompile with -Xlint:preview for details.
>
> $ java12 --enable-preview se2/X
> hello
> world
> 10
>
> Other details:
> $ uname -a
> CYGWIN_NT-6.1 XXXXX 2.9.0(0.318/5/3) 2017-09-12 10:18 x86_64 Cygwin
>
> The output of the program is given above – essentially a fall-through akin
> to a switch statement’s expected behavior.
> However, From JLS 12 15.29.1,* “Given a switch expression, if the switch
> block consists of switch labeled rules, then it is a compile-time error if
> any switch labeled block can complete normally. If, on the other hand, the
> switch block consists of switch labeled statement groups, then it is a
> compile-time error if either the last statement in the switch block can
> complete normally, or the switch block includes one or more switch labels
> at the end.” *
>
>  From the above I expected a compilation error to be shown since the case 1
> and case 2 arms can complete normally without a break expression.
>
> Is this an error in Javac or is it a wrong interpretation of "complete
> normally" in the spec on my part?
>   (Once the break expressions are uncommented, the switch expression returns
> the appropriate  values which is of course the correct behavior.) I am
> using ea 20 for this.
>
> Regards,
> Manoj Palat
> Eclipse Java Dev,
> IBM.
>
> PS: I had sent this mail yesterday from my IBM official id, but I don;t see
> that mail in the mailing list yet - hence resending from personal id.
>
> From: "Joseph D. Darcy" <joe.darcy at oracle.com>
> To: jdk-dev <jdk-dev at openjdk.java.net>
> Date: 11/27/2018 06:29 AM
> Subject: Reminder: submit CSRs for JDK 12 with sufficient time ahead of
> rampdown 1 milestone
> Sent by: "jdk-dev" <jdk-dev-bounces at openjdk.java.net>
>
>
>
> Hello,
>
> As a reminder, per the JDK 12 schedule
> (
> https://urldefense.proofpoint.com/v2/url?u=http-3A__openjdk.java.net_projects_jdk_12_&d=DwICaQ&c=jf_iaSHvJObTbx-siA1ZOg&r=A51zLdywEjS50U7u2UMWvsDIrUGEJ5IDXskL5MxIEjA&m=NBCxwvEUF_6bxl1FYxFdnHSL7MKL5Ya8ynDOXXPMdM4&s=nQVLCHjGW7vanaEOQjwFI3s_yk30KByjLdaV09RMXlA&e=)
> rampdown phase 1 of the
> release is coming up in a few weeks on Dec. 13, 2018.
>
> If you are planning to get a change into JDK 12 before the rampdown 1
> milestone and the change requires a CSR, the CSR needs to be an the
> Approved state before the change is pushed. The nominal review time of
> the CSR is one week. Note this is the time to review, not necessarily
> the time to approve. As with code reviews, the CSR process can provide
> feedback which has to be acted upon before a request is approved.
>
> Please plan accordingly; cheers,
>
> -Joe
> CSR Lead
>