Ambiguous decompiled operator overload / type argument list

[sgkoishi]

https://docs.microsoft.com/en-us/dotnet/csharp/language-reference/proposals/csharp-7.0/pattern-matching#changes-to-syntactic-disambiguation
The generic type arguments may be ambiguous.
Demo on SharpLab

        int num = 0;
        int num2 = 0;
        D d = null;
        F((num < num2), d > (d ?? d));

The last line was decompiled to F(num < num2, d > (d ?? d));(bracket missing), which num2, d will be considered as generic type arguments of num, and d ?? d as parameter.

[siegfriedpammer]

Should we always add extra parentheses around simple binary operator expressions in call argument lists (and tuple expressions?) with operator < and left operand identifier expression?

[dgrunwald]

Hmm this is tricky. It's also possible to construct cases where the < operator isn't directly contained in the argument list, e.g. F(b ? true : num < num2, d > (d ?? d))
This is really more of a lexical issue. We could detect the ambiguity in the token stream, but at that point it's too late to insert the opening parenthesis.