How to read $$\int_0^x \frac{1}{1+t^2}dt= \int_0^1 ( 1-t^2+ \cdot \cdot \cdot + (-1)^t t^{2n})dt$$
Specifically, why the integral switches from being $\int_0^x$ to being $\int_0^1$? Is this because $\int_0^x$ is "integral over all $x$" and since the series expansion for $\frac{1}{1+x^2}$ is defined for $|x|<1$, then one can get "all $x$" by integrating only the interval [0,1]?
However, since $\frac{1}{1+x^2}$ is surely defined for $[-1,1]$, then shouldn't the above rather be
$$2 \cdot \int_0^1 ( 1-t^2+ \cdot \cdot \cdot + (-1)^t t^{2n})dt$$
This appears in p.11 of this document.
Clearly the upper bound of integration should be the same. This is a typo.
Moreover the two expressions are not equal: in the second there should be the error term appearing at the very end for these to be equal, which is then bounded just below.
In formulas:
$$\int_0^x \frac{1}{1+t^2}dt= \int_0^x( 1-t^2+ \cdot \cdot \cdot + (-1)^t t^{2n})dt+R_n(x)$$