JDK-6875434 : The color of the two images are not the same during 2D test on solaris10-x64.
  • Type: Bug
  • Component: client-libs
  • Sub-Component: 2d
  • Affected Version: 1.4.2
  • Priority: P3
  • Status: Closed
  • Resolution: Duplicate
  • OS: solaris_10
  • CPU: x86
  • Submitted: 2009-08-25
  • Updated: 2011-03-08
  • Resolved: 2009-09-28
Related Reports
Duplicate :  
Description
JDK/JRE tested: jdkfb1.4.2_24b01
OS/architecture: solaris_10 x64
Reproducible: Always 
Reproducible on machine: stt-30 

Is it a platform specific regression: N 
Is it a Regression: Not sure 
Test log location: /net/stt-13/export/home0/results/1.4.2_24/b01_j4b/2d/solaris10_x64/manual/BufferedFIOSTest/BufferedFIOSTest.log
Steps to reproduce:
Login via VNC to appropriate machine.
Install JDK if required

Copy entire BufferedFIOSTest folder to local machine from /net/stt-13/export/home0/results/1.4.2_24/b01_j4b/2d/solaris10_x64/manual
cd to BufferedFIOSTest
execute:
java -Dworkdir=. -cp .:ttbuilder.jar javasoft.clientsqe.tonga.TTBuilder BufferedFIOSTest manual

and follow on-screen instructions

#Actual behavior:
After step 4,the first motorbike appears in GREEN color,but the second one is in PURPLE on solaris10-x64 with jdkfb1.4.2_24b01.

java.lang.RuntimeException is thrown with message: <br><br>ImageIO fails to read/write the image properly when using  a ImgOutputStream that is BufferedOS(FileOS) with a cache file. 

#Analyze:
In BufferedFIOSTest.java
  witr =
ImageIO.getImageWritersByFormatName(BufferedFIOSTest.resources.getString("pngformat"));

In BufferedFIOSTest.properties , when pngformate = jpeg the problem  exists,
modify pngformate=png , test case is passed.

#Exceptions:
None.

#Validation:
1.same configuration/1.4.2_22b04-j4b
2.same configuration/1.4.2_22b03-j4b
3.same configuration/1.4.2u18 fcs(latest live)
4.same configuration/5u16 fcs(latest live)
5.same JDK ( 32bit ), RedHat-x64
6.same JDK ( test target ),winvista-homebasic-x64
All of above have the same problem.

And Same issue happens on solaris10-x64 with 5u22b01(se).

Comments
EVALUATION After applying fix # 4705399, this problem goes away. It's been the issue since 1.4.2 FCS.
28-09-2009