Comment 3 for bug 308914

Nikodemus Siivola (nikodemus) wrote :

After another bout of wrestling with control-analysis, I'm inclined to believe the problem is not there after all.

The dead bits remain in the component because there is a XEP for the lambda from EVERY => (MAP NIL (LAMBDA ...) ...) -- this is quite correct as far as control analysis is concerned.

The existence of the XEP is not a great mystery either, since MAP doesn't get fully optimized there.

What is a mystery so far is why that XEP sticks around even though we should have by that point know that the (SETF AREF) will never return.