OpenJDK AIX FileEncodingTest testFileEncodingToDefault Default Charset: ISO-8859-1, expected: US-ASCII #14472
Comments
pshipton
changed the title
|
@tajila still a jdk18 file encoding issue on AIX |
|
Might be worth testing an Adoptium build to see if it's the same. |
|
Similar w/ #14227, will investigate them together. |
|
@JasonFengJ9 can you add #14473 to that list as well |
|
Not sure how similar it is to #14227. In #14227 the test is making wrong assumptions because of the change to use UTF8 by default. This one is an OpenJDK test, which should be correct, and I didn't see it failing on any other platform. There could be an issue with AIX itself, I'm not sure how well maintained this platforms is at OpenJDK any more. |
|
#14227 is a test issue as per #14227 (comment), also I noticed that the non-ascii command line option (tested via |
|
RI failed as well at Will exclude it first. Edit: actually it has been excluded for RI already. |
|
I'm concerned about there being a real problem here, you can exclude it but we shouldn't ignore it. My understanding is that AIX isn't maintained at OpenJDK any more the same way it was in the past, and may need fixes to support jdk18 properly. |
|
@pushkarnk this test fails on jdk18 Temurin builds from Adoptium, as well as OpenJ9. It may be related to https://openjdk.java.net/jeps/400. Note the JDK18 release schedule, we are running out of time to resolve problems for the initial release. http://openjdk.java.net/projects/jdk/18/ |
|
@JasonFengJ9 I believe #14227 to be a problem with the OpenJ9 test itself (at least on Windows), which I suspect is different from this problem. |
That's correct, the test makes assumption not applying to Java 18 w/ JEP 400 implemented. |
|
I was able to reproduce this JTREG failure with Temurin on AIX: |
|
Hello 👋, I just wanted to give you a head's up that I am working on these failures on the OpenJDK side. My initial assessment of the failure in ProcessBuilder/Basic.java is that there is a testbug related to the LIBPATH variable on AIX. The OpenJDK issue is created here, and I expect to have a PR for it shortly. I also merged this change related to FileEncodingTest. [edits for clarity about which tests I am referring to] |
|
@backwaterred can you please also create a PR with the fix against https://github.com/ibmruntimes/openj9-openjdk-jdk18. Unless you expect your OpenJDK fix is going to get into a jdk18 release build for the jdk18 initial release. Even if it does, there is no harm in delivering the same fix directly to https://github.com/ibmruntimes/openj9-openjdk-jdk18, the official fix will just be a no-op when merged. |
|
Did you request the OpenJDK fix to be backported to jdk18? |
|
@pshipton I suspect it's too late for the change to be included in jdk18, as the codebase has entered the Initial Release phase. Unfortunately, it might also be too early for the change to be backported as it will GA 2022/03/22. I have opened a PR for FileEncodingTest that mirrors the changes to the OpenJDK repo. |
|
@backwaterred thanks. We'd still like to have the OpenJDK change backported to jdk18 at the appropriate time. |
You're welcome. I am making a note to submit a backport after GA. |
|
Unexcluded the test adoptium/aqa-tests#3367 |
|
I received approval in the JBS to backport this change. I expect the request to be merged shortly. openjdk/jdk18u#96 |
|
Awesome, thanks. We can remove our patch once the official change gets merged (to the openj9 extensions). |
https://openj9-jenkins.osuosl.org/job/Test_openjdk18_j9_sanity.openjdk_ppc64_aix_Nightly/2 - p8-java1-ibm06
jdk_lang_0, jdk_lang_1
java/lang/System/FileEncodingTest.java
https://openj9-artifactory.osuosl.org/artifactory/ci-openj9/Test/Test_openjdk18_j9_sanity.openjdk_ppc64_aix_Nightly/2/openjdk_test_output.tar.gz
The text was updated successfully, but these errors were encountered: