In Nibble's answer, it is only by inference that we see that the division 25/99 poduces an infinitely recurring decimal and is therefore not a proof. Bamberger's proof seems incomplete and Hymie's, if correct, seems long winded.
The following is the classic way of proving this:
Let X = .2525252525... (1)
then 100X = 25.2525252525... (2)
Subract equation (2) from equation (1)
99X = 25, so X = 99/25
This works because we can see that for any position after the decimal point, the digit in each expression (1) and (2) is equal, so subtraction of one from the other leaves zero after the decimal point.
This proof can be used to find a fraction equal to any recurring decimal, using higher powers of 10 to match length of the repeating decimal. Where we used 100 above for .25 recurring, use 1000 for .258, 10000 for .2587 etc
Prove .25872587 recurring = 2587/99999
Let X = .25872587... (1)
then 10000X = 2587.25872587... (2)
Subtract (1) from (2)
9999X = 2587, so X = 2587/9999