-
Notifications
You must be signed in to change notification settings - Fork 131
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Switch SWT to Java 17 runtime #627
Conversation
Test Results 299 files ±0 299 suites ±0 5m 42s ⏱️ -15s For more details on these failures, see this check. Results for commit 717cc0d. ± Comparison against base commit 04678a2. ♻️ This comment has been updated with latest results. |
@HannesWell : I assume something Jenkins related doesn't work as expected.
but later on, it looks like something related to
Is that a known Jenkins build issue in SWT? I don't see any relationship to Java 17 changes. |
That's indeed very likely not related to this change. I assume the parameter in the (now) GString block has to be accessed through the params object, but that somehow did not show up in the verification of #628 and the subsequent master branch builds. IIRC those params are also injected as environment variables, but maybe not on the first run of a PR? |
As expected a second build succeeded, even without #613. It looks like the build parameters are only injected as environment variables on the second run. From my previous work on the SWT pipeline I believe to remember that the value of the parameter value was only available in a subsequent run as env. For whatever reason. Using the params-object seems to make it available immediately. Since you changed the way how the natives are build, without changing a file in a natives source directory (and therefore did not implicitly trigger a build of the natives), the natives build has to be enforced, by re-triggering the job in Jenkins via |
What do you mean by that? Where did I changed native build? |
You changed the JDKs that provide the header files for JNI and if there is a no JDK at the new path the SWT natives build for that platform likely fails because of the missing header, unless there is some auto-search like in the Windows build.bat. |
So should I revert changes on Jenkinsfile? |
At least for the platforms
So unless you are fine with a mixture of Java-17 and Java-11 headers you should revert all, yes. |
Note: there are still build files requiring Java 11 for execution (Jenkinsfile and buildSWT.xml). See eclipse-platform#626 Fixes eclipse-platform#625
So I did now.
I've updated #626 |
@HannesWell : in case you asked already what to install and where, can you please ask infra team to install Java 17 variants in same workstations / hosts / whatever and link to #626? |
Yes I can do that this evening. After thinking about the problem again, what maybe could also work is to collect/stash the JNI headers from the JVM at the 'main' Jenkins executor before the matrix is forked (this will likely be a linux VM), using the headers at the location returned by |
Thanks.
I have no opinion as I have no real experience with "right" Jenkins builds configuration way. |
Note: there are still build files requiring Java 11 for execution (Jenkinsfile and buildSWT.xml).
See #626
Fixes #625