[8u] RFR 8209113: Use WeakReference for lastFontStrike for created Fonts

[Hohensee, Paul]

Sorry for empty email, fumble-fingers.



Now the whole hunk is indented one space too much. Should be:



@@ -105,7 +107,22 @@

      * This pre-supposes that a FontStrike is a shareable object, which

      * it should.

      */

-    protected Reference lastFontStrike = new SoftReference(null);

+    protected Reference<FontStrike> lastFontStrike = new WeakReference<>(null);

+

+    /*

+     * if useWeak is true, proactively clear the cache after this

+     * many strikes are present. 0 means leave it alone.

+     */

+    private int strikeCacheMax = 0;

+    /*

+     * Whether to use weak refs for this font, even if soft refs is the default.

+     */

+    private boolean useWeak;

+

+    void setUseWeakRefs(boolean weak, int maxStrikes) {

+        this.useWeak = weak;

+        this.strikeCacheMax = weak && maxStrikes > 0 ? maxStrikes : 0;

+    }

     /*

      * POSSIBLE OPTIMISATION:



Paul



On 3/23/20, 1:51 PM, "Hohensee, Paul" <hohensee at amazon.com> wrote:



    On 3/23/20, 11:38 AM, "Zhengyu Gu" <zgu at redhat.com> wrote:



        Thanks, Paul.



        Fixed indents:



        http://cr.openjdk.java.net/~zgu/JDK-8209113-8u/webrev.01/



        -Zhengyu





        On 3/23/20 2:20 PM, Hohensee, Paul wrote:

        > Lgtm, except there's an over-indentation issue in the 2nd hunk for Font2D.java.

        >

        > Paul

        >

        > On 3/20/20, 3:02 PM, "jdk8u-dev on behalf of Zhengyu Gu" <jdk8u-dev-bounces at openjdk.java.net on behalf of zgu at redhat.com> wrote:

        >

        >      Hi,

        >

        >      I would like to backport this change to JDK8u. This patch fixes a memory

        >      leak that can be reproduced in 8u.

        >

        >      The original patch does not apply cleanly. Other than location mismatch,

        >      I changed

        >      FontStrikeDisposer.strikeCache type in original patch

        >      ConcurrentHashMap<FontStrikeDesc, Reference<FontStrike>> =>

        >      ConcurrentHashMap<FontStrikeDesc, Reference>

        >      to match Font2D.strikeCache

        >

        >      Original bug: https://bugs.openjdk.java.net/browse/JDK-8209113

        >      Original code review thread:

        >      https://mail.openjdk.java.net/pipermail/2d-dev/2019-December/010470.html

        >

        >

        >      8u Webrev:

        >      http://cr.openjdk.java.net/~zgu/JDK-8209113-8u/webrev.00/index.html

        >

        >      Test:

        >         The testcase in bug.

        >

        >      Thanks,

        >

        >      -Zhengyu

        >

        >

        >