[#40] CI update, builds with IBM Semeru and Java 11, 17

[bmarwell]

Some hints:

• Because these are a lot of builds, I did a pre-step with only Java 8 Temurin. If that fails, no need to build the others.
• The others use the matrix as expected. Excluded: The build which was already done.
• Build with IBM Semeru is a bit special. In contrast to Zulu and Temurin, it is not based on hotspot but Eclipse OpenJ9 and might have different encryption settings: algorithm aliases, defaults, etc.
  Therefore I consider it highly valuable for anything security-related.
• Build with recent JDKs should be a no-brainer.
• I recommend to exclude some Mac builds to speed up the build.
• Semeru is not ready yet (IBM Semeru (OpenJ9) Support actions/setup-java#289), using adopt-openj9 until then.

[bmarwell]

The failing checks in 11/17:

 80 [INFO] --- keytool-maven-plugin:1.6-SNAPSHOT:generateKeyPair (default) @ MKEYTOOL-10 ---
 81 [INFO] /bin/sh -c cd '$HOME/git/mojohaus/keytool-maven-plugin/keytool-maven-plugin/target/it/MKEYTOOL-10' && '$HOME/.sdkman/candidates/java/17.0.2-tem/bin/keytool' '-genkeypair' '-v' '-keystore' '$HOME/git/mojohaus/keytool-maven-plugin/keytool-maven-plugin/target/it/MKEYTOOL-10/ta    rget/keystore.jks' '-storepass' 'secret' '-alias' 'alias' '-dname' 'cn=Test Development, ou=Development, L=Herzele, ST=VL, o=FrankCornelis, c=BE' '-keypass' 'secret'
 82 [INFO] Keytool-Fehler: java.lang.Exception: The -keyalg option must be specified.                                                                                                                                                                                                                                   
 83 [INFO] java.lang.Exception: The -keyalg option must be specified.
 84 [INFO]  at java.base/sun.security.tools.keytool.Main.doCommands(Main.java:1168)
 85 [INFO]  at java.base/sun.security.tools.keytool.Main.run(Main.java:415)
 86 [INFO]  at java.base/sun.security.tools.keytool.Main.main(Main.java:408)

Fix: -keyalg parameter.

[slachiewicz]

Use 11 here

[bmarwell]

Can do. Any particular reason though?

[slachiewicz]

Decrease usage of Java 8 😜 - 11 should be just faster. Personally i would use 17

[bmarwell]

@slachiewicz what do you think about excluding Zulu on Mac?
Mac jobs tend to fail due to timeouts frequently.

[slachiewicz]

maybe let's use defaults for MacOS ie. Temurin?
https://github.com/actions/virtual-environments/blob/main/images/macos/macos-11-Readme.md#java

[olamy]

why do you merge that when the checks are still failing?

[bmarwell]

Because this PR was not about solving the Java 17 incompatibility, just about setting up the CI properly. See above, we identified the reason which is out of the scope of this PR.

Will create another issue for the.
But.
See the other issues: we are thinking about using the Keystore API instead of the executable. So even less reason to fix this one test here.

[olamy]

still doesn't make sense to merge a PR which breaks.
just fix the issues then merge it but the master in a working status.

[bmarwell]

It didn't break it. It wasn't working with JDK 17 before my PR. It just didn't show up.

[olamy]

I mean why adding 17 if it break??(9e0adf3)
if you make a requirement to build with 17 by adding it in the build file you must fix it as well or you do not add it.
if it's not in the build file it means it's not requirement (yet) but we know it doesn't work.
Now with the current situation if someone make a PR the PR will fail.

[bmarwell]

Oh I see what you mean.

I meant to fix it afterwards right away but had to go spontaneously. Will fix it later today. 😉