Rook v Pawns

REACHED another tricky ending in today's round seven at the Hastings Masters.

White to make his 41st move in Spanton (1954) - Laurence Butt (1635), but what should the result be: win, lose or draw?

I have just given up a rook on f1 to capture a promoted pawn, and Black has recaptured with his king.
A simple material-count in the resulting position gives Black a big advantage, but Black's king is well away from the centre of the action on the queenside.
One factor to always bear in mind is that two pawns safely abreast on the sixth rank will beat a rook even if the pawns' king cannot help them (and provided the rook's king cannot interfere).
If you do not know this rule, to which there are NO exceptions, try placing white pawns on e6 and d6 on an otherwise empty board (including no kings). It is White to move, but before he can do so Black is allowed to place a rook anywhere on the board. You will quickly find the black rook cannot prevent a white pawn from queening.
The position in the diagram is complicated by Black still having a pawn left, but, with correct play, should White win, lose or draw?
41.b5
This is non-controversial - almost any sensible-looking White move keeps the game moving towards its proper conclusion.
41...axb5
Not 41...a5? as it wastes a tempo that White can use to get a second pawn to the fifth rank with 42.c5.
42.cxb5
It seemed sensible to keep the white pawns as far from the black king as possible, but it 42.axb5 was probably equally good.
42...Re3+!?
This surprised me, but analysis with the engines Komodo9 and Stockfish10 suggests the move is just as legitimate as the perhaps-more obvious 42...Ra2. Both moves clear the e2 square for the black king to take a straight diagonal towards the white pawns.
43.Kb4 Re4+?
But this drives the white king in the direction it wants to go while not giving any benefit to Black.
Correct was 43...Ke2, when Black is a tempo up on the game continuation. As we shall see, White eventually wins the game by a single tempo.
44.Ka5 Ke2 45.b6 Kd3 46.b7 Re8 47.Kb6 Kc4 48.a5 Re6+ 49.Ka7
More precise is 49.Kc7.
Komodo9 reckons the position is drawn after the text, but it is not as Black cannot prevent White returning to the correct plan.
49...Re7 50.Kb6
Not 50.a6? as 50...Kb5 51.Ka8 Re8+ leaves White with nothing better than  a draw after 52.b8=Q Rxb8+ 53.Kxb8 Kxa6.
50...Re5
Setting a little trap.
Equally losing. but requiring accurate play by White, was 50...Re6+ (returning to the position after Black's 48th move) 51.Kc7 Re7+ 52.Kc6! (but not 52.Kc8? Kc5! when White cannot make progress, eg 53.a6 Kc6 54.Kb8 Kb6=) Re6+ 53.Kd7 etc (analysis by the engines).
51.a6
51b8=Q?? Rb5+ is a draw.
51...Re6+ 52.Ka7 1-0
So, to get back to the original question, the diagram position seems to be a draw with correct play, but one wasted tempo (43...Re4+?) tips the balance.