The Cambridge Grammar of the English Language has this (Page 1061):
In , CGEL doesn't analyze the determiner no as part of the antecedent of the relative clause.
Let's compare  with [11a] and [11b]:
[11a] No one who scored 40% or more was ever failed.
[11b] Some who scored 40% or more were failed.
In [11a] and [11b], what's the antecedent of the relative clause who scored 40% or more?
In CGEL's grammar, just like nobody, no one is a compound determinative, which always occurs in a fused-head construction. In traditional grammar, no one is a pronoun. In either grammar, therefore, no one is a syntactically inseparable unit.
Does that mean in [11a] that the antecedent of the relative clause has to be no one, and therefore that the antecedent has a negative meaning to it?
If so, how could you explain the fact that the antecedent in [11a], but not in , has a negative meaning?
If the antecedent is one in [11a], what's the antecedent in [11b]? Is it the covert nominal people?
Background Info on CGEL's grammar
(1) The integrated relative clause in CGEL refers to what traditional grammar calls 'restrictive relative clause'.
(2) CGEL basically says that the antecedent of the integrated relative clause is not an NP but a nominal.
(3) CGEL's definition of the term 'nominal' is completely different from that of traditional grammar as follows (The red underlines indicate what CGEL defines as 'nominals'):