Since people agreed with $\leq$ vs $\leqslant$ is a style issue. I checked a list of books from various disciplines,
the "$\leq$" group:
R. Shankar, Principle of Quantum Mechanics.
O.C. Zienkiewicz, et al. The Finite Element Method.
J. Munkres, Topology.
S. Boyd, et al. Convex Optimization.
S.M. Carroll, Spacetime and Geometry
R.W. Hamming, Numerical Methods for Scientists and Engineers
T. Sauer, Numerical Analysis.
R.A. Horn, et al. Matrix Analysis
G.H. Golub, et al. Matrix Computations
W. Fulton, et al. Representation Theory
T.W. Hungerford, Algebra
N. Jacobson, Basic Algebra
I. Goodfellow, Deep Learning
the "$\leqslant$" group:
D.V. Hutton, Fundamentals of Finite Element Analysis.
S.J. Farlow, Partial Differential Equations for Scientistes and Engineers
R.L. Graham, Concrete Mathematics
C.M. Bishop, Pattern Recognition and Machine Learning
It seems that most gurus preferred "$\leq$".