On Thu, 2003-07-10 at 08:50, Tim Waugh wrote: > On Thu, Jul 10, 2003 at 10:47:00AM -0500, Bill Gradwohl wrote: > > > I have a file full of random IP addresses, 1 per line. When I sort > > it with no options, I don't get a pure ASCII sort. > > You do in the C locale. In any other locale, the locale's sorting > rules apply. > > For pure ASCII sorting, you need to do: LC_COLLATE=C sort Does anyone know where the reasoning behind the strange collating behavior in the en_US.UTF-8 locale? No matter how I pre-sort the IP list, en_US.UTF-8 collating still wants to put the "128.*" IP address between two "12.*" ones. This is even true on Solaris, so the problem appears to be with the en_US.UTF-8 collation table, not the OS or its collation code. I thought the whole point of having a locale handle collation is so things are sorted in the way that seems most natural to the people who live there. Apparently I live in "C", not California. Tom
