/cc @pjweinb

That seems to be what it is doing. What should it do? gopls semtok produces
/⇒4,keyword,[]/type /⇒5,type,[definition]/Shape /⇒4,keyword,[]/func(/⇒3,type,[definition]/xyz /⇒3,operator,[]/.../⇒7,type,[]/float64) /⇒7,type,[]/float64
/⇒4,keyword,[]/type /⇒3,type,[definition]/ABC /⇒4,keyword,[]/func(/⇒3,type,[definition]/uvw /⇒7,type,[]/float64) (/⇒3,type,[definition]/err /⇒5,type,[]/error)

/⇒4,keyword,[]/func (/⇒1,variable,[]/a /⇒4,variable,[definition]/Node) /⇒2,member,[definition]/Sh(/⇒3,parameter,[definition]/abc /⇒3,operator,[]/.../⇒3,type,[]/int) {}

So the part that seems off to me is /⇒3,type,[definition]/xyz, as xyz is not really a definition, but really just an argument name suggestion. The user of the Shape type can choose to use xyz as the argument name or just any other name within the Go rules for identifier naming. Which brings me to the next point, xyz is closer to an identifier (parameter semtok).

thanks. let me think about it overnight. (or over-weekend).

There's the same issue in type X struct{A int}, which syntactically just has a StructDecl instead of a FuncDecl, so it need to change both places, i think.

Yes, I was about to create an issue complaining on that! The following declaration is confusing since all of the contents are colored the same!

type rocket struct {
	// detalles constructivos
	caliber, length, massEmpty float64
	// esp. tecnicas
	thrustMax, burnTime, propInitialMass float64
	// rocket state
	apogee, empty, mainParachute bool
}

I believe apogee, empty, mainParachute should be colored differently than bool for clarity.

Inline type definition seems to work as expected. The following example has the expected behaviour in gopls v0.6.9.

v := struct{A int}{}
var w struct{A float64}
var u = struct{A complex128}{}

Does color A and its type differently.

I will produce a fix. Thanks for reporting this.

https://go-review.googlesource.com/c/tools/+/305209

Change https://golang.org/cl/305209 mentions this issue: internal/lsp/semantic: fix some type definitions