.+ operator fails when substitute() with fold=true

[RobbesU]

.+ operator throws an error when substituteing a symbolic equation containing (only?) .+ and with fold=true.
Example below: All equations work apart from a1b, which fails with

MethodError: no method matching +(::Vector{Float64}, ::Float64)
For element-wise addition, use broadcasting with dot syntax: array .+ scalar
The function `+` exists, but no method is defined for this combination of argument types.

And oddly, f3 which also contains .+ but preceded with a .* does work!
and, as in f4, if specifying the variable as being of length c[1:3] it also works, but folds irrespective of the fold= flag.

Example:

a0 = [1., 2.25, 3.]
b0 = 1.

@variables a b c[1:3] # `a` fails in .+ when folding
f1 = 2a .+ b
f2 = 2a .* b
f3 = 2a .* b .+ a
f4 = 2c .+ b

a1a = substitute( f1, Dict([a=>a0, b=>b0]); fold=false)  
# this works # 1 + 2[1.0, 2.25, 3.0]	
a1b = substitute( f1, Dict([a=>a0, b=>b0]); fold=true)  
# this fails      
	
a2a = substitute( f2, Dict([a=>a0, b=>b0]); fold=false)  
# this works   # [2.0, 4.5, 6.0][Base.OneTo(3)]
a2b = substitute( f2, Dict([a=>a0, b=>b0]); fold=true)   
# this works   # [2.0, 4.5, 6.0][Base.OneTo(3)]
	
a3a = substitute( f3, Dict([a=>a0, b=>b0]); fold=false)  
# this works  # [1.0, 2.25, 3.0] + 2[1.0, 2.25, 3.0]*1
a3b = substitute( f3, Dict([a=>a0, b=>b0]); fold=true)   
# this works   # [3.0, 6.75, 9.0][Base.OneTo(3)]
	
a4b = substitute( f4, Dict([c=>a0, b=>b0]); fold=false)   
# this works but folds  # [3.0, 5.5, 7.0][Base.OneTo(3)]
a4b = substitute( f4, Dict([c=>a0, b=>b0]); fold=true)   
# this works   # [3.0, 5.5, 7.0][Base.OneTo(3)]