Similar to rS318912 except that ptrace.[sS] was a file in the source tree, not a generated assembly wrapper. I'm not sure of a good way to handle this other than testing for the existence of ptrace.[sS] in the .depend file itself.
This is a bit hackish, and in any case I would not leave this indefinitely. I think one year for NO_CLEAN workarounds is more than enough time; if we persist with this approach I think we could remove the ad-hoc cleanup after say 3 months. Posting the review for discussion of the approach in general.