The MWE below triggers the same error that has been causing Hypatia's tests to fail (see https://github.com/chriscoey/Hypatia.jl/runs/3599357717). The issue started a month ago and was unrelated to any changes in Hypatia, so I guess something changed somewhere in MP/DP/PolyJuMP etc.
using DynamicPolynomials
@polyvar x
@polyvar y
m = monomials(vcat(x, y), 0:2)
p = rand(length(m))' * m
p2 = subs(p, y => 0.0)
julia> p2(0.5)
ERROR: ArgumentError: args must be equal in length
Stacktrace:
[1] pairzip(#unused#::Tuple{PolyVar{true}}, #unused#::Tuple{})
@ MultivariatePolynomials ~/.julia/packages/MultivariatePolynomials/vqcb5/src/zip.jl:9
[2] pairzip(t::Tuple{PolyVar{true}, PolyVar{true}}, u::Tuple{Float64})
@ MultivariatePolynomials ~/.julia/packages/MultivariatePolynomials/vqcb5/src/zip.jl:10
[3] pairzip(p::Pair{Vector{PolyVar{true}}, Tuple{Float64}})
@ MultivariatePolynomials ~/.julia/packages/MultivariatePolynomials/vqcb5/src/zip.jl:13
[4] substitute
@ ~/.julia/packages/MultivariatePolynomials/vqcb5/src/substitution.jl:31 [inlined]
[5] (::Polynomial{true, Float64})(x::Float64)
@ DynamicPolynomials ~/.julia/packages/DynamicPolynomials/zw67w/src/subs.jl:122
[6] top-level scope
@ REPL[9]:1
The MWE below triggers the same error that has been causing Hypatia's tests to fail (see https://github.com/chriscoey/Hypatia.jl/runs/3599357717). The issue started a month ago and was unrelated to any changes in Hypatia, so I guess something changed somewhere in MP/DP/PolyJuMP etc.