[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-2d-dev
Subject:    Re: [OpenJDK 2D-Dev] RFR: 8257809: JNI warnings from Toolkit JPEG image decoding
From:       Sergey Bylokhov <serb () openjdk ! java ! net>
Date:       2020-12-23 22:08:54
Message-ID: 7IU037_RvHNLATTCuyAHp-6OJbi5qZmFof1_sGN921w=.b1d3e51a-e3bb-4029-87df-84ae1b5cbd7a () github ! com
[Download RAW message or body]

On Sun, 20 Dec 2020 01:37:47 GMT, Phil Race <prr@openjdk.org> wrote:

> The fix is to reverse the order of acquisition to get dst before src so that the \
> call to GetArrayLength() comes first. This also necessitates moving the \
> RELEASE_ARRAYS() call on an error condition to the new "2nd" block. 
> The new regression test passes on all platforms and all the other headless tests \
> passed too. So do the automated headful tests.

Marked as reviewed by serb (Reviewer).

test/jdk/java/awt/image/GetImageJNICheck/GetImageJNICheck.sh line 50:

> 48:     -cp "${CP}" -Xcheck:jni GetImageJNICheck | grep \
>                 ReleasePrimitiveArrayCritical > "${CP}"/log.txt
> 49: 
> 50: #if [ $? -ne 0 ]

I always wonder why it is not possible to report such warnings as fatal errors or at \
least report the exit code as non zero. Probably it is meant to create an \
enhancement?

-------------

PR: https://git.openjdk.java.net/jdk/pull/1850


[prev in list] [next in list] [prev in thread] [next in thread]