On Thu, Jul 15, 2004 at 10:58:06PM -0500, david nicol wrote: > On Tue, 2004-07-13 at 14:34, Nicholas Clark wrote: > > > > I'm not sure how worthwhile this really is, but I think it means that > > there's no argument about what's the most efficient way to reverse sort > > a list - don't write a reversed comparison function, simply write > > reverse sort .... > I wonder if this could be turned into yet another context -- "reverse > list" context -- causing any returning array to get reversed on the > stack before being returned, in a lighter way than calling the reverse > function. Just another idea. Based on the number of regression tests that I broke while implementing this, I don't think that it's a very common scenario, so would be complicating the general case for something rather specific and very rare. Plus it's unclear how to make any of it work in a backwards compatible way. Nicholas Clark