`invalid-name-notation` lint excessively reports `ρ` character in receiver names

eo:special in src/main/resources/org/eolang/funcs/special-name.xsl:10 exempts φ, the lambda, ξ and the cactus, but not ρ. invalid-name-notation filters candidates through that function, so every formation that writes its receiver down as [^ index opened bits] has the resulting <o name="ρ"/> measured against [a-z]+(-[a-z]+)* and reported. In objectionary/eo at eo-runtime, that is 629 of the lint's 764 reports, all of them noise. $eo:rho is already defined in the _specials.xsl this file imports through _funcs.xsl, so or $name = $eo:rho alongside the $eo:xi term is the whole fix.

The remaining 135 reports are a different question, and they are the reason the lint cannot be switched on in eo-runtime even once ρ is exempt. The motive says "no digits", and the standard library names its own numeric types with them: u8, i64, as-i16 through as-u64, win32 and stat64 all fail the expression, and none of them can be renamed without renaming the language's types. Either the expression should admit a digit that does not lead the name, or the lint needs to stay off wherever such names are the domain.

@yegor256

Start with eo:special in src/main/resources/org/eolang/funcs/special-name.xsl and trace how invalid-name-notation filters candidates through it, including the imported _specials.xsl and _funcs.xsl. Run the lint against eo-runtime to verify that receiver-name reports for ρ disappear; the remaining numeric-name reports require a separate decision about the expression or lint scope.