JDK-8021343 : Regression: 2 sun.io converter regression tests fail since JDK 5.0u55 b02
  • Type: Bug
  • Component: core-libs
  • Sub-Component: java.io
  • Affected Version: 5.0u55
  • Priority: P3
  • Status: Resolved
  • Resolution: Won't Fix
  • OS: windows_7
  • CPU: x86
  • Submitted: 2013-07-25
  • Updated: 2014-02-05
  • Resolved: 2014-02-05
Related Reports
Relates :  
Relates :  
Description
Testsuite name: Regression
Test names:
sun/io/CharToByteConverter/4809949/Test4809949.sh
sun/io/Converter/Test4200310.sh
JDK tested: JDK 5.0u55 b02
Is it a regression :YES
Regression introduced in JDK 5.0u55 b02

Platform specific: Windows
Comments
Last three builds do not expose the failure. Those are 5.0u65b04, 5.0u65b05 and 5.0u65b06. I've checked it with the following script: for b in b03 b04 b05 b06 ; do for p in windows-amd64 ; do dr=/net/jre.us.oracle.com/onestop/jdk/5.0u65/promoted/all/$b/binaries/$p/jre/lib if [ -d $dr ] ; then echo "Build $b Platform $p" find $dr \( -name rt.jar -or -name localedata.jar -or -name charsets.jar \) -print -exec jar -tf {} \; | grep '.*\.class' | sort | uniq -d fi done done Closing it with the 'Won't fix' resolution, as it's not clear where the problem is and taking into account that it stopped happening.
05-02-2014

These tests FAIL against: JDK 5.0u55 b09, JDK 5.0u55b10, JDK 5.0u56b01, JDK 5.0u56b03.
28-10-2013

Moreover, as far as I understand there's nothing to fix neither in tests nor in the code. It seems to be a build environment issue with the specific bundle where it was tested
23-10-2013

This is not applicable to 7u. Moreover, as far as I understand there's nothing to fix neither in tests nor in the code. It seems to be a build environment issue.
01-10-2013

SQE OK to defer But what's the intended fix version? Would be fine to fix it in 7u51.
30-09-2013

In b02 this appeared to be the build environment issue. First, it was not reproducible with JPRT builds. Second, log files for b02 differed a lot from the previous builds. Third, the problem happened only for windows x64 - all other platforms were fine.
30-09-2013

Looks like this one is an intermittent issue, was reproduced on b02 but not on b03 and again reproduced on b10. I don't consider it as a stopper. Requesting it to be deferred.
27-09-2013

Reproduced against JDK 5.0u55 b10 Windows 7 x64.
20-09-2013

I have rebuilt the bundle for 5.0u55-b02 with JPRT and the bundle does not have this problem. This must be a problem with building environment. I'm not sure if it's worth filing an issue against spoiled build b02.
20-08-2013

An easy way to check the failure without jtreg: cd ./jre/lib find . \( -name rt.jar -or -name localedata.jar -or -name charsets.jar \) -print -exec jar -tf {} \; | sort | uniq -d Classes are duplicated in 64-bit windows distribution, but not in 32-bit windows neither in linux.
05-08-2013

#section:shell ----------messages:(3/123)---------- command: shell Test4200310.sh [] reason: User specified action: run shell Test4200310.sh elapsed time (seconds): 0.707 ----------System.out:(1/140)---------- FAILED: sun/io/ByteToCharISO8859_1.class sun/io/ByteToCharSingleByte.class are duplicated between rt.jar, charsets.jar, and localedata.jar ----------System.err:(0/0)---------- result: Failed. Execution failed: exit code 1 test result: Failed. Execution failed: exit code 1
25-07-2013

#section:shell ----------messages:(3/123)---------- command: shell Test4809949.sh [] reason: User specified action: run shell Test4809949.sh elapsed time (seconds): 0.733 ----------System.out:(1/140)---------- FAILED: sun/io/ByteToCharISO8859_1.class sun/io/ByteToCharSingleByte.class are duplicated between rt.jar, charsets.jar, and localedata.jar ----------System.err:(1/37)---------- java full version "1.5.0_55-ea-b02" result: Failed. Execution failed: exit code 1 test result: Failed. Execution failed: exit code 1
25-07-2013