Bump SWT bundle versions for 4.33

[iloveeclipse]

See eclipse-platform/eclipse.platform.releng.aggregator#2085

[iloveeclipse]

Looks like SWT build hangs in signing native macosx.aarch code:

https://ci.eclipse.org/releng/job/eclipse.platform.swt/job/PR-1261/1/pipeline-console/?selected-node=191

[iloveeclipse]

OK, looks like we (https://ci.eclipse.org/releng/view/SWT%20Natives/) are waiting for this machine https://ci.eclipse.org/releng/computer/nc1ht-macos11-arm64/ that currently runs SDK tests https://ci.eclipse.org/releng/job/AutomatedTests/job/ep433I-unit-mac64-java17/

[github-actions]

Test Results

   313 files  ±0  313 suites  ±0   1h 48m 12s ⏱️ +15s
 4 067 tests ±0  107 ✅ ±0  2 💤 ±0  1 ❌ ±0   3 957 🔥 ±0 
23 397 runs  ±0  358 ✅ ±0  8 💤 ±0  1 ❌ ±0  23 030 🔥 ±0 

For more details on these failures and errors, see this check.

Results for commit 7acb4a8. ± Comparison against base commit fdde98e.

[HannesWell]

This is the right thing to do and should have been done in #1259

The last verification build failed due to the occasional OOM error. I restarted it hoping it passes this time.

OK, looks like we (https://ci.eclipse.org/releng/view/SWT%20Natives/) are waiting for this machine https://ci.eclipse.org/releng/computer/nc1ht-macos11-arm64/ that currently runs SDK tests https://ci.eclipse.org/releng/job/AutomatedTests/job/ep433I-unit-mac64-java17/

The jobs in the SWT natives view are not used anymore and will be deleted soon: eclipse-platform/eclipse.platform.releng.aggregator#1060

But your observation about the job waiting for the build machine is correct. The mac native builder is unfortunately over-occupied (hopefully this can be solved soon): eclipse-platform/eclipse.platform.releng.aggregator#2005 (comment)

[iloveeclipse]

The build fails, can you see why?

[HannesWell]

The build fails, can you see why?

I only see these obscure 801 Killed messages:
/home/jenkins/agent/workspace/eclipse.platform.swt_PR-1261/eclipse.platform.swt@tmp/durable-9deab4c8/script.sh.copy: line 2: 801 Killed mvn clean verify --batch-mode --threads 1C -V -U -e -DforkCount=0 -Papi-check -Dcompare-version-with-baselines.skip=false -Dorg.eclipse.swt.tests.junit.disable.test_isLocal=true -Dmaven.test.failure.ignore=true -Dmaven.test.error.ignore=true

The number often varies. I was told that this is an out of memory issue and it happened more often in the past. But I cannot really say why. The api-check was a suspect to cause it, but disabling wasn't the silver bullet to fix it :/
I have now disabled it for the verification build in a replay and also disabled test execution. Maybe we are lucky this time.

[HannesWell]

I have now disabled it for the verification build in a replay and also disabled test execution. Maybe we are lucky this time.

Maven build passed now.
Only the subsequent test-analyzer failed because the test

00:13:29.851  Recording test results
00:13:29.946  No test report files were found. Configuration error?
Error when executing always post condition:

I suggest to submit this nevertheless to have the natives re-built.

[iloveeclipse]

Should we trigger build now, or wait for the nightlyin 3 hours?

[HannesWell]

Should we trigger build now, or wait for the nightlyin 3 hours?

We first have to wait until the master built triggered by submitting this change has passed and the rebuilt natives have been pushed.
I'm currently monitoring the built and will retrigger it without API check if it fails again.

[merks]

Thank you both for caring for this !!

[HannesWell]

Should we trigger build now, or wait for the nightlyin 3 hours?

We first have to wait until the master built triggered by submitting this change has passed and the rebuilt natives have been pushed.

Once that is done we could trigger an I-build. But I would be fine waiting for it too. I don't know if anyone needs it in the (European) night or will take action if there is again an issue.

Thank you both for caring for this !!

Your welcome 😃

[iloveeclipse]

OK, let's wait for a regular build then.

[HannesWell]

The master build succeeded immediately and the new binaries were built and pushed to:
a6f4cc6
Everything regarding the SWT natives should now be fine again.