Add envelopeS / envelopeSMay for querying scalar displacements from envelopes

[mgsloan]

I realize these shouldn't be encouraged, but they seem potentially useful!

Discussing the "intrudeEnvelope" in the haddocks led to figuring out how to express surprise at the fact that "envelopeS v x - envelopeS (negateV v) x < 0" is possible.

[byorgey]

I find the use of the subjunctive and the discussion of "moving" the bounding hyperplane confusing. Isn't this just returning the magnitude of whatever is returned by envelopeV[May]? If so, the wording of the documentation ought to be similar as well.

Also, there seems to be a type error -- see the little red dot next to 3ce357b for a link to the Travis build log.

[byorgey]

Looks good to me. Just to be clear, note that according to this patch, these new functions are exported from Diagrams.Core.Envelope but not re-exported from Diagrams.Core. I think that's an acceptable state of affairs, given how (as you say) these "shouldn't be encouraged" (though I don't know if you intended it or just didn't think to add exports to Diagrams.Core). However, I could be convinced to re-export them. In any case, I'll go ahead and merge and the question of re-exporting can be dealt with later.