That is a nice endgame :)
Endgame study with wrong solution in Karpov book

My favorite CD--The Best of Chess Informant: Anatoly Karpov--has a "play like Karpov" endgame exercise where you must find suboptimal moves. Karpov found a draw in a winning position. If you find the win, you get the problem wrong.
All books have errors; some far more than others. You want names of the worst books: go to Edward Winter's site (he has a list).

That book you own is an absolute masterpiece, I have one exactly identical to yours...but mine has signatures by Karpov, Miles and Spassky
However, there is in fact a small bad print:
Page 247, C.Schlechter - P. Meitner, the 2nd move is hxg4+ not fxg4+

eric c. 8. kh1 for black leads to a draw by either:
8. kh1 9. rxg7 draw
8.kh1 9. Something Else 9. g1=Q 10. rxg1+ 10. kxg1 draw
If anyone else has a better solution for black, please share.
eric c. 8. kh1 for black leads to a draw by either:
8. kh1 9. rxg7 draw
8.kh1 9. Something Else 9. g1=Q 10. rxg1+ 10. kxg1 draw
If anyone else has a better solution for black, please share.
4. ... g2
5. Rf8 Kh2
How the white win?
I could not believe when I found that.
In the The Mosaic Chess book of A. Karpov and E. Guik, Fifty remarkable miniatures chapter, Raduga publishing house, 1984, (My version book is in spanish) case study number 24 there are the following position:
Aparently all is ok, and whithe win in all variants analized here, but one of them has an incomplete analysis
Karpov did not see this variant or I am crazy ?