Albert Meyer says he experiences something like this occurring several times a year in his role as editor and indicated that he thought one should take the issue to the editor and "offer your opinion of the publishability of the result" but not state that the result should be published?
Caroline Whitbeck: What is the difference that you see between "tell the editor the main result is worth publishing" and "offer your opinion of the publishability of the result"?
Albert Meyer: A theoretical result by itself is not intrinsically publishable. There is often a tradeoff between the significance of a result and the simplicity of proof in making a publication decision: a nonbreakthrough, incremental result is probably not worth publishing unless its proof is reasonably short and accessible. Also, the fact that the reviewer came up (quickly?) with her own simpler proof raises the possibility the author's result is too routine to be publishable. Finally, the reviewer ought to be sensitive to the conflict of interest and nonobjectivity to which she is now subject. She should put an emphasis on explaining to, and seeking the advice of, the editor, rather than making a simple publication recommendation. (BTW, it's a common pitfall for an expert reviewer who comes up with her own simpler proof to overly devalue the submitted manuscript as routine.)