On Mon, 19 Dec 2022 19:23:25 GMT, Naoto Sato <na...@openjdk.org> wrote:
> Moving the built-in implementation of `Console` from `java.io` package into
> `jdk.internal.io` package. It now implements `JdkConsole` interface and is
> accessed through `ProxyingConsole`.
src/java.base/share/classes/jdk/internal/io/JdkConsoleImpl.java line 365:
> 363: // For convenience
> 364: public static JdkConsole getJdkConsole(Charset cs) {
> 365: return INSTANCE != null ? INSTANCE : new
> JdkConsoleImpl().console(true, cs);
Would this singleton/caching be necessary? Or could we expect that the caller
of this method would cache it (like the `java.io.Console` actually does right
now)?
In theory, this current implementation in this method can potentially return an
incorrect (cached) instance on a second call if the `Charset` that's passed is
different from the one that was used by the cached instance that is returned.
But, since we know/expect this method to be called only by the
`java.io.Console` and with the same `Charset` instance, I think this is OK. I
do think that if we remove the caching/singleton logic here, this class could
be a bit more simpler.
-------------
PR: https://git.openjdk.org/jdk/pull/11729
