Possible bug in DSolve

Question

Let us consider

DSolve[{f == y[x] + x*(1 - y'[x]^2)/2/y'[x], y[0] == f}, y[x], x]

{}

and a warning "DSolve::bvnul: For some branches of the general solution, the given boundary conditions lead to an empty solution".

Doesn't that result contradict the result of

DSolve[{f == y[x] + x*(1 - y'[x]^2)/2/y'[x]}, y[x], x]

{{y[x] -> 1/2 (E^C[1] + 2 f - E^-C[1] x^2)}, {y[x] -> 1/2 (-E^C[1] + 2 f + E^-C[1] x^2)}}

or I don't understand something?

Answer 1

There is no finite solution for the given IC. To see this more clearly, will use Michael E2 code from

simplifying-dsolve-output-exponentials-raised-to-constant

ClearAll[f, y, x];
constSimplify2 // ClearAll;
constSimplify2[dsol_, rest___] := 
  Activate[
   FixedPoint[#[[First@OrderingBy[#, LeafCount, 1]]] &[{#, 
       Replace[#, 
        s_ /; ! FreeQ[s, Power[_, p_ /; ! FreeQ[p, _C]]] :> 
         Simplify[# /. 
           Cases[#, 
            Power[_, 
              p_ /; ! FreeQ[p, _C]] :> (c : 
                Alternatives @@ Cases[p, _C, {0, Infinity}] :> 
               Log[c]), Infinity], rest]], 
       Replace[#, 
        s_ /; ! FreeQ[s, a_?NumericQ c_C] :> 
         Simplify[# /. 
           Cases[#, a_?NumericQ c_C /; a != 0 :> (c :> c/a), 
            Infinity], rest]], 
       Replace[#, 
        s_ /; ! FreeQ[s, a_?NumericQ + c_C] :> 
         Simplify[# /. 
           Cases[#, a_?NumericQ + c_C :> (c :> c - a), Infinity], 
          rest]]}] &, Inactivate[dsol, Function], 100], Function];

Now

sol = DSolve[{f == y[x] + x*(1 - y'[x]^2)/2/y'[x]}, y[x], x]

Trying first solution. But simplifying it first

sol = sol // constSimplify2

Applying the initial conditions on first solution

sol[[1]] /. {y[x] -> f, x -> 0} /. Rule -> Equal

Solve[%, C[1]]

(* {{C[1] -> 0}} *)

Plugging C[1]=0 into the solution y[x] -> f + x^2/(4 C[1]) - C[1] clearly gives division by zero.

Same for the second solution.

Therefore the initial condition gives no valid solution.

Update

-1. Can you elaborate your "Same for the second solution"?

I thought this was obvious. But here it is

sol[[2]] /. {y[x] -> f, x -> 0} /. Rule -> Equal

Solve[%, C[1]]

(*  {{C[1] -> 0}} *)

Plugging the above in the second solution also gives divide by zero.

Btw, you could just have asked without downvoting at same time. I do not understand why you have to downvote for.