Continue #2027: throw CoreException instead of returning null

[ruspl-afed]

• see Make connector API provisional and prevent static singleton access #2027 (comment)
• see Make connector API provisional and prevent static singleton access #2027 (comment)

[ruspl-afed]

@laeubi here is the last part of what was suggested for #2027
I plan to work more on non-API part of these types (to avoid double calls for getActiveWorkbenchPage() in some API chains for example)

[github-actions]

Test Results

 1 920 files   -  27   1 920 suites   - 27   1h 31m 12s ⏱️ + 3m 6s
 4 718 tests ±  0   4 693 ✅  -   1   24 💤 ±0  1 ❌ +1 
14 001 runs   - 153  13 833 ✅  - 154  167 💤 ±0  1 ❌ +1 

For more details on these failures, see this check.

Results for commit b8c6c8b. ± Comparison against base commit d4985f2.

♻️ This comment has been updated with latest results.

[laeubi]

Thanks for working on this, I think instead of using Optionals here it would be much more appropriate to throw a CoreException if something goes wrong as we do elsewhere in Platform code and otherwise never return null.

[ruspl-afed]

Thank you for your review @laeubi
I support your vision regarding this kind of signatures and we have much more other signatures to be changed this way in further PRs.

[laeubi]

Looks good

[ruspl-afed]

@laeubi TBH I do not understand what exactly is wrong with pipeline and how can I fix it

[laeubi]

I think this is a usual test failure:

• Random failing ParallelBuildChainTest.testIndividualProjectBuilds_WithManyProjects_ProjectRelaxedRule #825