Mercury's motion does violate Newtonian gravity. General Relativity however simplifies to Newtonian gravity for various low-dimension applications though (i.e. an Apple falling from a tree, the orbit of the moon for most astronomical guidance purposes over that distance etc.)

In much the same way we don't account for the curvature of the Earth when building a house's foundations.

That's my point though: when you violate a fundamental physical law, it can't just be wrong. You have to somehow also account for why it doesn't happen all the time elsewhere (i.e. the effect is small to within experimental error over common measurement cases) and by necessity the new theory to explain it has to encompass the predictions of the old i.e. the benefit of Newtonian gravity was that it easily recovers the behaviour of nested epicycles of orbital motion (and made new predictions successfully where epicycles could not).