Should we allow null for the parameters to

WriteDebug, WriteWarning, WriteVerbose, etc?

I suspect that we ideally would not, but I'm not sure if making such a change would break anything

IParameterMetadataProvider has 5 dependencies. This complicates understanding design intentions and we will review this later on phase 5.

Should we allow null for the parameters to
WriteDebug, WriteWarning, WriteVerbose, etc?

I suspect that we ideally would not, but I'm not sure if making such a change would break anything

Nullable annotations can not break anything. They only explicitly expose design intentions.

This pull request has been automatically marked as stale because it has been marked as requiring author feedback but has not had any activity for 15 days. It will be closed if no further activity occurs within 10 days of this comment.

Hmm, we never check for null in the WriteDebug() API until low level host implementations where we ignore null messages.
For WriteError() we throw if null.
For WriteObject() we accept null.
For WriteProgress() we throw if null.
For WriteVerbose we accept null (we do the same as for WriteDebug()).
For WriteWarning we accept null (we do the same as for WriteDebug()).
For WriteCommandDetail() we accept null.

This pull request has been automatically marked as stale because it has been marked as requiring author feedback but has not had any activity for 15 days. It will be closed if no further activity occurs within 10 days of this comment.

WriteObject makes sense. The throws make sense. The skips trouble me, since changing them is breaking, but it would be nice to change the type signature. But I'm not sure I see the value there. I guess the best course of action is to label the APIs as you've researched them @iSazonov

The skips trouble me, since changing them is breaking

Nullable annotations are only about design intentions. So we can annotate all the "skips" as non-nullable without adding real null checks and de-facto without breaking changes.
I can the commit if you want.

So we can annotate all the "skips" as non-nullable without adding real null checks and de-facto without breaking changes.

Yeah that's perfect

The PR branch was removed so I opened new PR #15566.