Matlab floating point problem

Hello. I am using Matlab R2007b. I have a floating point problem in it. I am using x=rotx(45*pi/180)*[1 2 3]' and its value is x =[1.0000; 3.5355; 0.7071]. And i am doing reverse rotation of x which is y=rotx(-45*pi/180)*x. And i am doing y==[1; 2; 3]. Logically it is showing 1 1 0. I want all 1's when y==[1; 2; 3]. Can any one help with this problem.

Answer 1

[per isakson]

The script

%%
x = rotx(45*pi/180)*[1 2 3]';
y = rotx(-45*pi/180)*x;
is1 = abs((y-[1;2;3])<eps(10))

returns

is1 =
     1
     1
     1

with R2018b