added 'semver' as core type accelerator for S.M.A.SemanticVersion - #4142

Appveyor temporary failed.

Yeah, it seems one of the New-TimeSpan tests is failing. It must have already been broken when I merged to rebase.

Paul Higinbotham (@PaulHigin) currently SemanticVersion is permitted in constrained language, please review if that's OK.

Oisin Grehan (@oising) PowerShell core types are intended to be the absolute minimum needed in a constrained language interactive session. Can you tell us the scenario where this type is needed and why constrained language is required?

Paul Higinbotham (@PaulHigin) I guess I included it in core types for the same reasons you guys included [ModuleSpecification]. Given that NuGet packages are a core (no pun intended) building block of CoreFX, and [semver] being a large part of their identity, it seemed like something that may be more useful than not. I know this is all a bit fuzzy, but it's just what made sense to me.

Paul Higinbotham (@PaulHigin) I think the rationale is that SemanticVersion is replacing System.Version ( the type of $PSVersionTable.PSVersion is SemanticVersion), so I think it makes sense to allow it in constrained language. But we need your expertise to inspect the implementation of SemanticVersion and see if it's safe to be exposed in constrained language.

Dongbo Wang (@daxian-dbw) Yes, also good points. I took a cursory glance at the implementation and didn't see any red flags, but I'm also not privy to your criteria.

Ok, I think this is reasonable. I'll review for security safety. Thanks.

Filed #4221 to track adding tests for type accelerators in PowerShell.