On Mon, Mar 07, 2005 at 08:02:40PM +0100, Steven Bosscher wrote: > [ min/max expr ] > So, maybe the extension is not used very much. Perhaps it should be > removed?
Then we'll just get RMS pissed off at the SC again; he hates it when we remove his extensions. (We've sometimes done so anyway, but I think we should proceed with caution). Maybe there's an approach that could simply make them less bothersome, but transforming them into something else at a very early stage of parsing. And as the primary justification for them was to have a clean way of avoiding side-effects (so that an argument to the max operator is evaluated only once) I don't see why they have to be supported as rvalues.
