Set o.e.j.c.compiler.problem.nullUncheckedConversion to ignore by default

[snjeza]

Fixes redhat-developer/vscode-java#3501

[rgrunber]

Wouldn't it make more sense to put it within the if/else at

eclipse.jdt.ls/org.eclipse.jdt.ls.core/src/org/eclipse/jdt/ls/core/internal/preferences/Preferences.java

Line 2498 in 2016a78

private Map<String, String> generateProjectNullAnalysisOptions(String nonnullType, String nullableType, String nonnullbydefaultType) {

? Otherwise, looks fine.

[rgrunber]

Also, looks like testMixedNullAnalysis needs to be adjusted :

eclipse.jdt.ls/org.eclipse.jdt.ls.tests/src/org/eclipse/jdt/ls/core/internal/preferences/NullAnalysisTest.java

Line 70 in 2016a78

IMarker marker3 = getWarningMarker(project, "Null type mismatch: required '@Nonnull List<String>' but the provided value is specified as @Nullable");

.. or is this an issue.

[snjeza]

@rgrunber I have updated the PR..

[agriffis]

I promise I'm not trying to be a troublemaker 😂

I'm still chasing a problem in #2655 where org.eclipse.jdt.core.compiler.problem.nullUncheckedConversion=ignore and it's not clear to me that this PR is doing the right thing. (Although it might be! I'm theorizing a bunch in an area that isn't very familiar to me!)

See my comment at redhat-developer/vscode-java#3501 (comment)

I'd love if someone could authoritatively say that this change makes sense whenever null analysis is enabled, or if possibly it's a hammer fix to a problem that could/should be solved differently. Thanks!