J.C. Hull derives the following relation $$Ke^{-rT} - S \le p \le Ke^{-rT}$$
where $p$ is european put option price, $K$ is strike price, $S$ is stock spot price,$r$ rate of interest and $T$ time to maturity. The above relation holds for no arbitrage.
The book states european put option price does not necessarily increase with increase in time to maturity. But just using the above relation as $T$ increases doesn't it mean $p$ will always go to 0 for large $T$ ?
There is no contradiction at all here. $Ke^{-rT}$ goes to zero for $T$ going to $\infty$ so the relation you mention suggests that $-S\leq p \leq 0$ as $T$ gets bigger. If $r>0$. Put prices can (in theory) not be negative so $p$ goes also to zero according to the relation you mention.
Is this consistent with the pricing function of Black-Scholes? Yes. Consider this graph:
When Time to maturity $T$ gets large then enough then the Put value is not longer increasing in $T$
The next graph shows that the inequalities you mention hold for the parameters ($r$, $S$, $K$, $\sigma$) being the same as in the above graph:
