Dr. Mustafa Zayd, in his book النسخ في القرآن الكريم, volume 2, page 294 (item 1208), says something strange. He says that in order to know that a verse has truly abrogated a Sunna, a hadeeth must so state. On what basis does he make this rule?
Why would we need two uncertain narrations, and most hadeeths are uncertain in their occurrence (ظني الورود), to enforce a verse, which is certain in its occurrence (قطعي الورود)? Isn't it more logical to conclude that a hadeeth that orders something contrary to what a verse orders, that this hadeeth must either be unauthentic or has been abrogated by the verse?
As examples, Dr. Zayd cites 2:144 as abrogating the Sunna of praying toward Jerusalem, and the prohibition, by 2:238, of talking to others while praying. See
this topic for a detailed discussion of the Qibla redirection issue. As for 2:238, since when is the prohibition of something about which nothing before was specified, since when is that considered abrogation? Any command in Islam that was not issued on day 1 would be considered abrogation under that strange rule!
For completion, here is the verse,
Dr. Zayd repeats his theory about a confirming hadeeth, later on pages 305-306, as he cites two cases where the Quran abrogated the Sunna. The first case is 2:185 abrogating the fasting in `Aashooraa' and the second is 2:187 abrogating the practice of not mating with wives on fasting eves. He relies heavily on authentic narrations, none of which is attributed to the Prophet (PBUH), interpreting the matter as abrogation. That is not proof of abrogation; it's the opinion of a highly esteemed scholar.