Since elections in our current environment are turnout-based, in terms of gauging popular support, "eligible voter population" is probably the right metric. By that metric, Trump got just over a quarter of the vote --- and, by any metric, less than Clinton.

Since some fraction of eligible non-voters wouldn't have voted no matter who the candidate was, the truth of the matter is obviously somewhere in between a quarter and a half. But there's no sense in which "half the vote" isn't misleading.