Possible typo in type inference

[jskeet]

From https://github.com/dotnet/csharpstandard/blob/draft-v7/standard/expressions.md#12632-the-first-phase, after an introduction which sets the scene with:

Tₑ M<X₁...Xᵥ>(T₁ p₁ ... Tₓ pₓ)

we have this text:

For each of the method arguments Eᵢ:

• If Eᵢ is an anonymous function, an explicit parameter type inference (§12.6.3.8) is made from Eᵢ to Tᵢ
• Otherwise, if Eᵢ has a type U and xᵢ is a value parameter (§15.6.2.2) then a lower-bound inference (§12.6.3.10) is made from U to Tᵢ.
• Otherwise, if Eᵢ has a type U and xᵢ is a reference (§15.6.2.3) or output (§15.6.2.4) parameter then an exact inference (§12.6.3.9) is made from U to Tᵢ.
• Otherwise, no inference is made for this argument.

Should the xᵢ in the second and third bullets be pᵢ?

I'll probably just create a PR for this, but it was easier to file an issue to start with...

[Nigel-Ecma]

Well a parameter has the grammar:

attributes? parameter_modifier? type identifier default_argument?

Matching Tₑ M<X₁...Xᵥ>(T₁ p₁ ... Tₓ pₓ) to this the Tᵢ are the type and pᵢ the identifier. Neither of these determine whether the parameter is value or reference parameter, that is down to the parameter_modifier. So it seems the writer is using xᵢ as the whole parameter spec but does not define it as such, so the xᵢ are neither pᵢ or Tᵢ…

We could try:

• Otherwise, if Eᵢ has a type U and the i^th parameter is a value parameter (§15.6.2.2) then a lower-bound inference (§12.6.3.10) is made from U to Tᵢ.
• Otherwise, if Eᵢ has a type U and the i^th parameter is a reference (§15.6.2.3) or output (§15.6.2.4) parameter then an exact inference (§12.6.3.9) is made from U to Tᵢ.

Which skirts introducing another letter and requires HTML <sup> tags in the source :-)

[jskeet]

Fixed in #975