Java 12-17API additions #10106

zbynek · 2025-03-12T01:37:30Z

Fixes #9872
Fixes #10091
Fixes #9991

niloc132

Nice work on the draft - some quick feedback

.../test-super/com/google/gwt/dev/jjs/super/com/google/gwt/emultest/java12/lang/StringTest.java

user/test/com/google/gwt/emultest/java12/util/stream/CollectorsTest.java

user/test/com/google/gwt/emultest/java12/lang/StringTest.java

user/test/com/google/gwt/emultest/EmulJava12Suite.java

.../com/google/gwt/dev/jjs/super/com/google/gwt/emultest/java12/util/stream/CollectorsTest.java

user/super/com/google/gwt/emul/java/util/stream/Collectors.java

.../com/google/gwt/dev/jjs/super/com/google/gwt/emultest/java12/util/stream/CollectorsTest.java

niloc132 · 2025-03-12T15:53:51Z

.../test-super/com/google/gwt/dev/jjs/super/com/google/gwt/emultest/java12/lang/StringTest.java

+    assertEquals(3, hideFromCompiler("foo").transform(String::length));
+  }
+
+  public void testIndent() {


please also include a set of these tests that does not do a hideFromCompiler, so we can make sure that if the compiler is able to constant-fold these method calls, that it does so correctly.

I did not implement any tests without hideFromCompiler yet since I wasn't sure which methods need them. It seems that tests either hide all inputs from compiler (like the ones from https://github.com/gwtproject/gwt/pull/9975/files#diff-79013db0b72ddfb9522327d93e6c5804f44c160025590e971ddfc7fd4eb7c045R26 ) or hide nothing (https://github.com/zbynek/gwt/blob/f0c1486a46018aec9de93458136cfc161f569e81/user/test/com/google/gwt/emultest/java/lang/StringTest.java#L542 ). Also I'm not sure if it wouldn't be better to test the constant-folding optimization and verify that these are actually folded rather than make assertions where we don't know if we're testing folding or emulation.

Using hideFromCompiler ensures that the emulated sources provided can be called and function the way that the test expects. Passing a constant and allowing constant folding means that the operation will be performed by the compiler itself, using the JRE's own implementation, which does two things for us: we know that constant folding doesn't somehow break with this call, and that the JRE implementation doesn't behave in a way that fails the test.

This isn't a guarantee - the constant folding mechanism is pretty specific (I've been meaning to file a bug about that) and won't always kick in, but it does let us get a little more coverage.

user/super/com/google/gwt/emul/java/util/stream/Stream.java

zbynek · 2025-05-01T20:30:39Z

Full build https://github.com/zbynek/gwt/actions/runs/14282864925

user/super/com/google/gwt/emul/java/util/stream/LongStream.java

user/test/com/google/gwt/emultest/java17/util/MapEntryTest.java

user/test/com/google/gwt/emultest/java17/lang/StringTest.java

...m/google/gwt/emultest/super/com/google/gwt/emultest/java17/util/stream/DoubleStreamTest.java

Co-authored-by: Colin Alworth <[email protected]>

zbynek · 2025-07-22T21:48:08Z

https://github.com/zbynek/gwt/actions/runs/16452377759 is 🟢 , @niloc132 thanks for the feedback

gwt.iml

user/super/com/google/gwt/emul/java/lang/String.java

user/super/com/google/gwt/emul/java/util/BitSet.java

user/super/com/google/gwt/emul/java/util/stream/Collectors.java

user/super/com/google/gwt/emul/java/util/BitSet.java

...test-super/com/google/gwt/emultest/super/com/google/gwt/emultest/java17/lang/StringTest.java

user/super/com/google/gwt/emul/java/lang/String.java

niloc132 · 2025-09-02T18:55:00Z

user/super/com/google/gwt/emul/java/lang/String.java

@@ -488,10 +491,6 @@ public String intern() {
    return checkNotNull(this);
  }

-  public boolean isEmpty() {
-    return length() == 0;


far-off future nit: this is going to end up being inlined as something like
<expr>.length==0, and promoting this to the supertype (which, i think i suggested) will mean we keep getting that. But, it would be smaller to just implement as <expr>=="" in the cases where we know we have a java.lang.String, since java String is final+immutable, and js string can be ==/=== compared. I think we could also live with !<expr>.length, but that's still longer regardless of == or === usage.

niloc132 · 2025-09-02T18:55:20Z

user/super/com/google/gwt/emul/java/lang/String.java

@@ -811,6 +814,131 @@ private int getTrailingWhitespaceLength() {
    return length;
  }

+  public String indent(int spaces) {
+    if (spaces  == 0) {


checkstyle ignores supersource, for better or worse.

Suggested change

if (spaces == 0) {

if (spaces == 0) {

EDIT: actually this is wrong, javadoc specifies that line endings are rewritten

tools/api-checker/config/gwt212_213userapi.conf

dev/core/src/com/google/gwt/dev/javac/testing/impl/JavaResourceBase.java

niloc132 · 2025-09-03T15:37:40Z

user/super/com/google/gwt/emul/java/lang/String.java

+
+  public String stripIndent() {
+    if (isEmpty()) {
+      return "";


technically this could also be return this;, since the "isEmpty" check is "are you the empty string". Might result in making the code bigger though, but probably will be smaller - String is final so all methods are made static and this is an argument (so will be rewritten to argument a or something) - one char instead of two.

I'm not 100% confident in that analysis, so maybe don't change this just from my guesswork, just trying to think through what JS we'll get.

niloc132 · 2025-09-03T15:42:29Z

...test-super/com/google/gwt/emultest/super/com/google/gwt/emultest/java17/lang/StringTest.java

+    assertEquals("x\ny\n", hideFromCompiler("  x\n  y").indent(-2));
+    assertEquals("x\ny\n", hideFromCompiler("  x\r\n  y").indent(-2));
+    assertEquals("x\ny\n", hideFromCompiler("  x\r  y").indent(-2));
+  }


From https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/lang/String.html#indent(int), we should have a few more tests:

validate that \t is removed as if it was a single space

validate that indent(0) changes line endings to \n - this will fail

generally validate that \r\n is rewritten to \n (edit: this is covered for positive and negative already I see)

validate that negative indent with fewer spaces just removes what spaces are available - and that indent with too many spaces leaves extra spaces.

validate that the empty string is handled correctly, and other strings with no line break - I think this will fail

validate trailing/leading linebreak in the string

niloc132 · 2025-09-03T15:42:48Z

user/super/com/google/gwt/emul/java/lang/String.java

+    } else {
+      indentedLines = lines().map(
+          line -> line.substring(Math.min(-spaces, line.getLeadingWhitespaceLength())));
+    }


For a little more code, this could be a little faster in the spaces == 0 case, just assign indentedLines = lines() with no calls to getLeadingWhitespaceLength(), bounds checks, or min() tests for each line. I think we would rather the code you have now though, as that seems a little silly to do, but just pointing it out.
https://docs.oracle.com/en/java/javase/17/docs/api/java.base/java/lang/String.html#indent(int)

It would be good to have tests that validate this too, added a note below for this.

Co-authored-by: Colin Alworth <[email protected]>

zbynek marked this pull request as draft March 12, 2025 01:37

zbynek changed the title ~~wip~~ Java 12-17API additions Mar 12, 2025

zbynek force-pushed the java17-emul branch 2 times, most recently from 42c7d0a to ec22afe Compare March 12, 2025 08:42

niloc132 reviewed Mar 12, 2025

View reviewed changes

zbynek force-pushed the java17-emul branch 3 times, most recently from 077dfea to 55d86e0 Compare March 13, 2025 01:39

natros reviewed Mar 13, 2025

View reviewed changes

user/super/com/google/gwt/emul/java/util/stream/Stream.java Show resolved Hide resolved

zbynek force-pushed the java17-emul branch 4 times, most recently from 033deb4 to b6adb5d Compare March 18, 2025 06:02

zbynek force-pushed the java17-emul branch from b6adb5d to d0e547d Compare March 29, 2025 08:20

zbynek mentioned this pull request Mar 29, 2025

Small patch, adding CharSequence::isEmpty #10097

Closed

zbynek force-pushed the java17-emul branch 6 times, most recently from 0257e71 to b8996b5 Compare April 2, 2025 21:53

zbynek marked this pull request as ready for review April 2, 2025 22:38

zbynek force-pushed the java17-emul branch 3 times, most recently from ec661eb to a8e31dc Compare April 5, 2025 14:31

zbynek added the Category-JRE label Apr 27, 2025

zbynek mentioned this pull request May 7, 2025

Add BigDecimal.TWO #10129

Open

niloc132 reviewed Jul 20, 2025

View reviewed changes

zbynek force-pushed the java17-emul branch from ead4e41 to c5b445c Compare July 21, 2025 22:09

zbynek and others added 3 commits July 22, 2025 02:50

Java 12 emulation

195c8dd

Apply suggestions from code review

b04e627

Co-authored-by: Colin Alworth <[email protected]>

More emulation

dcf9006

zbynek force-pushed the java17-emul branch from a743415 to acb0298 Compare July 22, 2025 00:51

Review feedback

1ef880d

zbynek force-pushed the java17-emul branch from acb0298 to 1ef880d Compare July 22, 2025 18:21

niloc132 reviewed Jul 23, 2025

View reviewed changes

zbynek added 5 commits July 29, 2025 21:16

Remove iml file

4549a2c

Review feedback

a5611f0

Merge branch 'main' into java17-emul

beab47e

Mention override as missing

ceb2069

Undo unused refactoring

f71b697

niloc132 reviewed Sep 3, 2025

View reviewed changes

zbynek and others added 2 commits September 3, 2025 20:11

Explain moved method

8dc1856

Co-authored-by: Colin Alworth <[email protected]>

Remove useless override in test code

0023483

Java 12-17API additions #10106

Are you sure you want to change the base?

Java 12-17API additions #10106

Uh oh!

Conversation

zbynek commented Mar 12, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

niloc132 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

niloc132 Mar 12, 2025

Choose a reason for hiding this comment

Uh oh!

zbynek May 9, 2025

Choose a reason for hiding this comment

Uh oh!

niloc132 Jul 30, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

zbynek commented May 1, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

zbynek commented Jul 22, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

niloc132 Sep 2, 2025

Choose a reason for hiding this comment

Uh oh!

niloc132 Sep 2, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

niloc132 Sep 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

niloc132 Sep 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

niloc132 Sep 3, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

zbynek commented Mar 12, 2025 •

edited

Loading

niloc132 Sep 3, 2025 •

edited

Loading

niloc132 Sep 3, 2025 •

edited

Loading