JDK 16 RFR of JDK-8235496 : "Start of release updates for JDK 16" and related work

[Joe Darcy]

Before pushing, I'll also update the copyright year.

On 5/28/2020 4:10 PM, Jonathan Gibbons wrote:
> Reading the javac parts ...
>
>
> http://cr.openjdk.java.net/~darcy/8235496.4/src/jdk.compiler/share/classes/com/sun/tools/javac/code/Source.java.sdiff.html 
>
>
> The tbd in "15, tbd" should be updated sometime.


Sure; I'll all a note for "text blocks", as done in SourceVersion.


>
> Lines 161-174: why not a switch?


Hmm. Not sure; the method may date back to the pre-enum days and it 
wasn't updated to use a switch. I'm convert it over.


>
> http://cr.openjdk.java.net/~darcy/8235496.4/test/langtools/tools/javac/api/T6395981.java.sdiff.html 
>
>
> Long line


Is there a way to get line wrap for those jtreg lines?


>
> http://cr.openjdk.java.net/~darcy/8235496.4/test/langtools/tools/javac/lib/JavacTestingAbstractProcessor.java.sdiff.html 
>
>
> Line 145: why is this line not updated?


Looks like I missed it from last time and it didn't get picked up with 
my grep/sed for "RELEASE_15"; good catch.


>
> Line 114, 125: Why the annotations on abstract classes? Do they have 
> any effect?


The SupportedSourceVersion annotation is not @Inherited so it wouldn't 
have a direct effect, but IIRC there are cases where subclasses query 
the parent's annotation to determine that information. Regardless, I 
think it is easier to maintain the SupportedSourceVersion annotation on 
all the visitor type even if not strictly necessary on some of them.


>
> http://cr.openjdk.java.net/~darcy/8235496.4/test/langtools/tools/javac/versions/Versions.java.sdiff.html 
>
>
> Line 301: by analogy with preceding methods, should it include New16.java


There isn't non-preview new-in-16 code, so no New16.java just yet :-)

(In an unrelated edit, I was considering only have a single javac 
invocation for all the files expected to compile successfully for a 
given source level, which should cut down on the test run time a bit, 
but that can be done under a separate bug.)


>
> Line 296: Should it have a call of expectedFail or a comment?

Sure; I'll replicate the comment.

Thanks,

-Joe