8296919: Make system tests that detect memory leaks more robust by using JMemoryBuddy utility #1121

lukostyra · 2023-05-01T14:48:32Z

Verified these changes on all platforms and saw no abnormalities in test execution.

This change is based on a preliminary patch done by @aghaisas , with some minor changes and the addition of one Leak-related web test. EventListenerLeak test was not touched, as it is a separate manual test app instead of a JUnit test.

Progress

Change must be properly reviewed (1 review required, with at least 1 Reviewer)
Change must not contain extraneous whitespace
Commit message must refer to an issue

Issue

JDK-8296919: Make system tests that detect memory leaks more robust by using JMemoryBuddy utility

Reviewers

Andy Goryachev (@andy-goryachev-oracle - Committer)
Kevin Rushforth (@kevinrushforth - Reviewer)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jfx.git pull/1121/head:pull/1121
$ git checkout pull/1121

Update a local copy of the PR:
$ git checkout pull/1121
$ git pull https://git.openjdk.org/jfx.git pull/1121/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 1121

View PR using the GUI difftool:
$ git pr show -t 1121

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jfx/pull/1121.diff

Webrev

Link to Webrev Comment

@aghaisas

Based on preliminary patch by @aghaisas

bridgekeeper · 2023-05-01T14:48:51Z

👋 Welcome back lkostyra! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

mlbridge · 2023-05-01T15:00:23Z

Webrevs

kevinrushforth · 2023-05-01T15:34:35Z

EventListenerLeak test was not touched, as it is a separate manual test app instead of a JUnit test.

In addition to the manual test, the javafx.web module also has an automated EventListenerLeakTest that could benefit from using JMemoryBuddy, but it would not be a trivial change, so it might be better to file a follow-on bug for that.

andy-goryachev-oracle

LG, with minor suggestions. What do you think?

build.gradle

modules/javafx.web/src/test/java/test/javafx/scene/web/LeakTest.java

andy-goryachev-oracle · 2023-05-01T16:07:35Z

tests/system/src/test/java/test/javafx/embed/swing/SwingNodeDnDMemoryLeakTest.java

-        attemptGCSwingNode();
-        assertEquals(TOTAL_SWINGNODE, getCleanedUpSwingNodeCount());
+
+        for (WeakReference<SwingNode> ref : weakRefArrSN) {


I wonder if we should add a utility method to JMemoryBuddy to operate on an array/collection, so we don't have to repeat this code again and again?

andy-goryachev-oracle · 2023-05-01T16:09:52Z

tests/system/src/test/java/test/javafx/embed/swing/SwingNodeMemoryLeakTest.java

-        assertEquals(TOTAL_SWINGNODE, getCleanedUpJLabelCount());
+        SwingUtilities.invokeAndWait(() -> Util.sleep(500));
+
+        for (WeakReference<SwingNode> ref : weakRefArrSN) {


same comment as before about a new utility method

andy-goryachev-oracle · 2023-05-01T16:11:23Z

tests/system/src/test/java/test/javafx/scene/control/TabPaneHeaderLeakTest.java

-        Assert.assertNull("Couldn't collect TextField", textFieldWeakRef.get());
+
+        JMemoryBuddy.assertCollectable(tabWeakRef);
+        JMemoryBuddy.assertCollectable(textFieldWeakRef);


perhaps the utility method operating on an array should accept varargs, like so:

assertCollectable(WeakRef a, WeakRef ... b)

(to differentiate it from a single arg method, and to make sure there are 2+ args)

* JMemoryBuddy now has helpers for arrays of WeakReference, ArrayList of WeakReference and vararg WeakReferences * Updated copyright years in some files which were missed in previous changes

lukostyra · 2023-05-02T10:26:53Z

EventListenerLeak test was not touched, as it is a separate manual test app instead of a JUnit test.

In addition to the manual test, the javafx.web module also has an automated EventListenerLeakTest that could benefit from using JMemoryBuddy, but it would not be a trivial change, so it might be better to file a follow-on bug for that.

I actually missed that. I looked briefly into it and I agree it is a larger change, I will file a follow-up bug for it after this PR is merged.

andy-goryachev-oracle · 2023-05-02T15:49:41Z

modules/javafx.base/src/test/java/test/util/memory/JMemoryBuddy.java

+        }
+    }
+
+    public static <T> void assertCollectable(ArrayList<WeakReference<T>> weakReferences) {


Could we change the argument to Collection<...>?

andy-goryachev-oracle · 2023-05-02T15:50:17Z

modules/javafx.web/src/test/java/test/javafx/scene/web/LeakTest.java

-
-            Thread.sleep(SLEEP_TIME);
+        for (WeakReference ref : willGC) {
+            JMemoryBuddy.assertCollectable(ref);


should we use newly added methods here?

Done + others.

andy-goryachev-oracle · 2023-05-02T15:50:35Z

modules/javafx.web/src/test/java/test/javafx/scene/web/LeakTest.java

-
-            Thread.sleep(SLEEP_TIME);
+        for (WeakReference ref : willGC) {
+            JMemoryBuddy.assertCollectable(ref);


ditto, new methods?

andy-goryachev-oracle · 2023-05-02T15:50:45Z

modules/javafx.web/src/test/java/test/javafx/scene/web/LeakTest.java

-
-            Thread.sleep(SLEEP_TIME);
+        for (WeakReference ref : willGC) {
+            JMemoryBuddy.assertCollectable(ref);


* Changed ArrayList<> to Collection<> in one of assertCollectable-s * Updated LeakTest to use new assertCollectable helpers

andy-goryachev-oracle · 2023-05-03T18:15:34Z

The changes look good!
Unfortunately, the project won't build in Eclipse due to

The type test.util.memory.JMemoryBuddy is not accessible LeakTest.java /web/src/test/java/test/javafx/scene/web

To fix that, could you please modify line 45 of modules/web/.classpath file to this (er, better include the whole file):

<?xml version="1.0" encoding="UTF-8"?>
<classpath>
	<classpathentry kind="src" path="src/main/java"/>
	<classpathentry kind="src" path="src/main/native/Source/WebCore/bindings/java/dom3/java"/>
	<classpathentry kind="src" output="testbin" path="src/shims/java">
		<attributes>
			<attribute name="test" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="src" output="testbin" path="src/test/java">
		<attributes>
			<attribute name="test" value="true"/>
			<attribute name="optional" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="src" path="src/main/resources">
		<attributes>
			<attribute name="optional" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="src" output="testbin" path="src/test/resources">
		<attributes>
			<attribute name="test" value="true"/>
			<attribute name="optional" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry combineaccessrules="false" kind="src" path="/media">
		<attributes>
			<attribute name="module" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry combineaccessrules="false" kind="src" path="/controls">
		<attributes>
			<attribute name="module" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry combineaccessrules="false" kind="src" path="/graphics">
		<attributes>
			<attribute name="module" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry combineaccessrules="false" kind="src" path="/base">
		<attributes>
			<attribute name="module" value="true"/>
			<attribute name="add-exports" value="javafx.base/com.sun.javafx=javafx.web:javafx.base/test.util.memory=javafx.web"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="con" path="org.eclipse.jdt.junit.JUNIT_CONTAINER/5">
		<attributes>
			<attribute name="test" value="true"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="con" path="org.eclipse.jdt.launching.JRE_CONTAINER">
		<attributes>
			<attribute name="module" value="true"/>
			<attribute name="add-reads" value="javafx.web=java.management"/>
		</attributes>
	</classpathentry>
	<classpathentry kind="output" path="bin"/>
</classpath>

andy-goryachev-oracle

looks good, one .classpath file change is requested.

lukostyra · 2023-05-04T08:17:15Z

Done, thanks for checking!

kevinrushforth

The changes generally look good, although I left a comment about sleeping on the EDT.

modules/javafx.base/src/test/java/test/util/memory/JMemoryBuddy.java

kevinrushforth · 2023-05-04T12:05:45Z

tests/system/src/test/java/test/javafx/embed/swing/SwingNodeMemoryLeakTest.java


-        attemptGCJLabel();
-        assertEquals(TOTAL_SWINGNODE, getCleanedUpJLabelCount());
+        SwingUtilities.invokeAndWait(() -> Util.sleep(500));


Sleeping on either the JavaFX Application thread or the Swing EDT (as in this case) is not generally desirable. What is the sleep trying to achieve? Is there a better way, possibly sleeping on the test thread and then doing a SwingUtilities.invokeAndWait (maybe with a no-op runnable? not sure)?.

I can't believe I missed that!
I wonder if we could use CountDownLatch instead here and use Util.await(CountDownLatch) or something like that.

This is something that came from the original patch - I'm no Swing expert so I originally figured I should leave it here. My only guess was that it (among other Thread.sleep()-s in this patch) is trying to let the other threads run for a bit and chew through the resources to make sure they're collectable.

Upon closer inspection it looks like JMemoryBuddy is already doing those sleeps, so I'll remove them* - tests should work fine without those.

* EDIT: them == sleeps done right before JMemoryBuddy calls

JMemoryBuddy performs sleeps to check if objects are collectable by itself, there is no reason to do those.

lukostyra · 2023-05-05T10:32:21Z

I removed the sleeps, ran the test multiple times on all platforms and didn't notice any abnormalities.

kevinrushforth

I saw a one-time failure that suggests waiting for the EDT to run something, and possibly a sleep, might still be needed.

kevinrushforth · 2023-05-05T12:22:03Z

tests/system/src/test/java/test/javafx/embed/swing/SwingNodeMemoryLeakTest.java

@@ -77,8 +77,6 @@ public void testSwingNodeMemoryLeak() throws InterruptedException, InvocationTar
            testSwingNodeObjectsInStage();
        });

-        SwingUtilities.invokeAndWait(() -> Util.sleep(500));
-
        JMemoryBuddy.assertCollectable(weakRefArrSN);


On my Mac (MacBook Pro x64 running macOS 13.2.1) I saw a one-time test failure here. So there might be a need to wait for the EDT to at least do something, possibly followed by a sleep (on the test thread, not on the EDT).

Interesting, I will add a runAndWait() noop and a test thread sleep then.

I think you mean SwingUtilities::invokeAndWait, right?

Btw, I also got a similar one-time failure in SwingNodeDnDMemoryLeakTest, so perhaps the same could be added there?

Yes, I meant invokeAndWait. A force of habit from using our Util class...

I also added it to SwingNodeDnDMemoryLeakTest just to be sure. After a few runs it works fine on my end (but also, it did earlier so I can't really fully test it).

On some platforms Swing memory leak tests occasionally failed, this might help to mitigate this problem.

kevinrushforth

Looks good.

openjdk · 2023-05-05T14:02:12Z

@lukostyra This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8296919: Make system tests that detect memory leaks more robust by using JMemoryBuddy utility

Reviewed-by: angorya, kcr

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 7 new commits pushed to the master branch:

27ee7e3: 8297071: Provide gradle "TEST_ONLY" flag to completely suppress building the sdk and shims
c50ce60: 8307208: Add GridPane constructor that accepts hGap and vGap values
0a1fd6e: 8299335: Monkey Tester Application
50bafc2: 8302816: Refactor sorting-related classes
97405de: 8307076: gradle test should always run tests
1975165: 8296590: StraightLineTest fails always on Linux and sometimes on other platforms
0b36052: 8306590: Add Windows/macOS system files to .gitignore

Please see this link for an up-to-date comparison between the source branch of this pull request and the master branch.
As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

As you do not have Committer status in this project an existing Committer must agree to sponsor your change. Possible candidates are the reviewers of this PR (@andy-goryachev-oracle, @kevinrushforth) but any other Committer may sponsor as well.

➡️ To flag this PR as ready for integration with the above commit message, type /integrate in a new comment. (Afterwards, your sponsor types /sponsor in a new comment to perform the integration).

lukostyra · 2023-05-05T15:13:14Z

@andy-goryachev-oracle I'll wait for your last look at it before I integrate it

andy-goryachev-oracle

minor javadoc changes, I'll re-approve if you want to make them.

andy-goryachev-oracle · 2023-05-05T16:13:39Z

modules/javafx.base/src/test/java/test/util/memory/JMemoryBuddy.java

@@ -91,6 +92,38 @@ public static void assertCollectable(WeakReference weakReference) {
        }
    }

+    /**
+     * Checks whether an array of WeakReference objects can be collected.


I think it should rather say something like "whether content of all references in the array", rather than array itself.

Done (+ others)

andy-goryachev-oracle · 2023-05-05T16:13:54Z

modules/javafx.base/src/test/java/test/util/memory/JMemoryBuddy.java

+    }
+
+    /**
+     * Checks whether a Collection of WeakReference objects can be collected.


same comment: "content of all references in a collection"

andy-goryachev-oracle · 2023-05-05T16:17:28Z

modules/javafx.base/src/test/java/test/util/memory/JMemoryBuddy.java

+    }
+
+    /**
+     * Checks whether provided WeakReference objects can be collected.


same comment: "content of references..."

andy-goryachev-oracle

LGTM, thanks!

lukostyra · 2023-05-08T15:20:15Z

/integrate

openjdk · 2023-05-08T15:21:27Z

@lukostyra
Your change (at version 98e7251) is now ready to be sponsored by a Committer.

andy-goryachev-oracle · 2023-05-08T15:22:14Z

/sponsor

openjdk · 2023-05-08T15:23:15Z

Going to push as commit d9a82d1.
Since your change was applied there have been 7 commits pushed to the master branch:

27ee7e3: 8297071: Provide gradle "TEST_ONLY" flag to completely suppress building the sdk and shims
c50ce60: 8307208: Add GridPane constructor that accepts hGap and vGap values
0a1fd6e: 8299335: Monkey Tester Application
50bafc2: 8302816: Refactor sorting-related classes
97405de: 8307076: gradle test should always run tests
1975165: 8296590: StraightLineTest fails always on Linux and sometimes on other platforms
0b36052: 8306590: Add Windows/macOS system files to .gitignore

Your commit was automatically rebased without conflicts.

openjdk · 2023-05-08T15:23:40Z

@andy-goryachev-oracle @lukostyra Pushed as commit d9a82d1.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

lukostyra added 2 commits April 26, 2023 14:37

Use JMemoryBuddy in tests currently using System.gc()

9bd7ab7

Based on preliminary patch by @aghaisas

Use JMemoryBuddy in Web leak tests

fbee1d2

openjdk bot added the rfr Ready for review label May 1, 2023

andy-goryachev-oracle reviewed May 1, 2023

View reviewed changes

Address review issues

158ed40

* JMemoryBuddy now has helpers for arrays of WeakReference, ArrayList of WeakReference and vararg WeakReferences * Updated copyright years in some files which were missed in previous changes

andy-goryachev-oracle reviewed May 2, 2023

View reviewed changes

Minor review updates

0b8b058

* Changed ArrayList<> to Collection<> in one of assertCollectable-s * Updated LeakTest to use new assertCollectable helpers

andy-goryachev-oracle suggested changes May 3, 2023

View reviewed changes

Update javafx.web classpath to include JMemoryBuddy

5ea23cf

kevinrushforth reviewed May 4, 2023

View reviewed changes

lukostyra added 2 commits May 5, 2023 08:34

Remove thread sleeps before JMemoryBuddy calls

6444cfe

JMemoryBuddy performs sleeps to check if objects are collectable by itself, there is no reason to do those.

JMemoryBuddy: Add javadoc comments

492d670

kevinrushforth reviewed May 5, 2023

View reviewed changes

Add an EDT no-op and sleep for a bit

1230c8b

On some platforms Swing memory leak tests occasionally failed, this might help to mitigate this problem.

kevinrushforth approved these changes May 5, 2023

View reviewed changes

openjdk bot added the ready Ready to be integrated label May 5, 2023

kevinrushforth requested a review from andy-goryachev-oracle May 5, 2023 16:03

andy-goryachev-oracle approved these changes May 5, 2023

View reviewed changes

JMemoryBuddy: Update javadoc comments

98e7251

openjdk bot removed the ready Ready to be integrated label May 8, 2023

kevinrushforth approved these changes May 8, 2023

View reviewed changes

openjdk bot added the ready Ready to be integrated label May 8, 2023

andy-goryachev-oracle approved these changes May 8, 2023

View reviewed changes

openjdk bot added the sponsor Ready to sponsor label May 8, 2023

openjdk bot added the integrated Pull request has been integrated label May 8, 2023

openjdk bot closed this May 8, 2023

openjdk bot removed ready Ready to be integrated rfr Ready for review sponsor Ready to sponsor labels May 8, 2023

lukostyra deleted the jmemorybuddy-8296919 branch May 8, 2023 15:39

8296919: Make system tests that detect memory leaks more robust by using JMemoryBuddy utility #1121

8296919: Make system tests that detect memory leaks more robust by using JMemoryBuddy utility #1121

Conversation

lukostyra commented May 1, 2023 • edited by openjdk bot

Progress

Issue

Reviewers

Reviewing

Webrev

bridgekeeper bot commented May 1, 2023

mlbridge bot commented May 1, 2023 • edited

Webrevs

kevinrushforth commented May 1, 2023

andy-goryachev-oracle left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lukostyra commented May 2, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andy-goryachev-oracle commented May 3, 2023 • edited

andy-goryachev-oracle left a comment

Choose a reason for hiding this comment

lukostyra commented May 4, 2023

kevinrushforth left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

lukostyra May 5, 2023 • edited

Choose a reason for hiding this comment

lukostyra commented May 5, 2023

kevinrushforth left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kevinrushforth left a comment

Choose a reason for hiding this comment

openjdk bot commented May 5, 2023 • edited

lukostyra commented May 5, 2023

andy-goryachev-oracle left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andy-goryachev-oracle left a comment

Choose a reason for hiding this comment

lukostyra commented May 8, 2023

openjdk bot commented May 8, 2023

andy-goryachev-oracle commented May 8, 2023

openjdk bot commented May 8, 2023

openjdk bot commented May 8, 2023

lukostyra commented May 1, 2023 •

edited by openjdk bot

mlbridge bot commented May 1, 2023 •

edited

andy-goryachev-oracle commented May 3, 2023 •

edited

lukostyra May 5, 2023 •

edited

openjdk bot commented May 5, 2023 •

edited