Bug 14741 – [REG2.066] problems with template overload resolution

Based on https://issues.dlang.org/show_bug.cgi?id=14735#c4 ------------------------------------------------------------------------ template isInputRange(R) { enum bool isInputRange = !is(R : U[n], U, size_t n); } ptrdiff_t indexOf(Range)(Range s, dchar c, bool cs = true) // A if (isInputRange!Range) { return 0; } static if(1) ptrdiff_t indexOf(T, size_t n)(ref T[n] s, dchar c, bool cs = true) // B if (!isInputRange!(T[n])) { return 0; } static if(0) ptrdiff_t indexOf(T)(ref T s, dchar c, bool cs = true) // C if (is(T : U[n], U, size_t n)) { return 0; } void main() { char[10] fixedSizeArray = "0123456789"; pragma(msg, isInputRange!(typeof(fixedSizeArray))); assert(indexOf(fixedSizeArray, '2') == 2); // line 30 pragma(msg, isInputRange!(typeof(fixedSizeArray[]))); assert(indexOf(fixedSizeArray[], '2') == 2); // line 32 } ------------------------------------------------------------------------ The above code fails to compile with current DMD: false true xx.d(32): Error: template xx.indexOf cannot deduce function from argument types !()(char[], char), candidates are: xx.d(6): xx.indexOf(Range)(Range s, dchar c, bool cs = true) if (isInputRange!Range) xx.d(13): xx.indexOf(T, ulong n)(ref T[n] s, dchar c, bool cs = true) if (!isInputRange!(T[n])) Clearly, the template constraint of overload (A) matches, as witnessed by second pragma(msg). However, compilation succeeds if (C) is enabled instead of (B). Further modification (beginning with the original code, (B) is enabled, (C) is disabled) reveals: When the `cs` parameter of (B) is removed completely, the error message becomes: xx.d(32): Error: xx.indexOf called with argument types (char[], char) matches both: xx.d(6): xx.indexOf!(char[]).indexOf(char[] s, dchar c, bool cs) and: xx.d(13): xx.indexOf!(char, 10LU).indexOf(ref char[10] s, dchar c) This is clearly wrong. `char[]` doesn't match overload (B). Furthermore, when the `cs` parameter of (A) is also removed, it suddenly compiles successfully.

This appears to be a regression. Introduced in https://github.com/D-Programming-Language/dmd/pull/3485

The root issue is same with issue 14735. *** This issue has been marked as a duplicate of issue 14735 ***