Error in norm in some condition

I have a simple test case to find the least square solution of
Adec_cvs-x, but it keeps give me this error message:
Error using == (line 4)
Array dimensions must match for binary array op.
Error in norm (line 81)
I am sure the dimension of arrays are correct, and if I change the range of x then it will finish correctly. here is mysimple code
variable dec_cvs(n)

the dimension of A is 201x3 and x is 1x201;
each column of A is a different monolithic increased positive data for example [0,0…10,15…30…300]
When x=-1:0.01 1 it will cause error in norm and fail,but if I make lower bound of x larger than 0 then it will work, ex: x=0.1:0.005:1.1.
Can anyone give me a hint? If that because I don’t have an negative component in A when x is below 0? thanks

I would say this sounds enough like a bug that you should submit a bug report at . Make sure to include the full output of cvx_version, the full text of the error message, and precise conditions to reproduce the problem.

Thanks, just submitted