@hubert-reinterpretcast: Could you please review this PR?

The footnote is indeed obsolete in its clarification on the terminology.

I believe a note remains necessary to clarify that tokens that would otherwise be instances of unsupported conditionally-supported-directive still "otherwise act as preprocessing directives" (because they would have required a diagnostic). In the alternative, implementations should not fail compilation on what may be interpreted as an unsupported conditionally-supported-directive within the list of arguments; they would either need to treat the tokens in question as "plain" tokens, or (to fail compilation) they would instead have to document support for all forms of conditionally-supported-directive which they do not intend to support as having semantics amounting to "causes the program to be ill-formed". @zygoloid: comments?

Hubert, would you be looking for an alternative footnote such as "A conditionally-supported-directive is a directive regardless of whether the implementation supports it." ?

I think that makes sense given the context ("[a sequence of tokens] that would act as a preprocessing directive").

Yes, something like the suggested alternative footnote would be fine.

Force-pushed a new version that replaces the footnote with @zygoloid's wording.