I recognize that your example is artificial but it does not support arbitrary "ls" options. It seems to me that ~meino is asking the wrong question. On Sun, Jan 26, 2014 at 8:20 PM, Bart Schaefer wrote: > On Jan 26, 8:59pm, Phil Pennock wrote: > } > } I'd use the print builtin of zsh, so that there are no argv length > } limitations, combined with zsh's very powerful glob operators. > > Meino mentioned in private email that he has particular reasons for > using "ls" instead of "find". If the options of "ls" are important > (rather than just the file names) and you don't need to worry too > much about CPU consumption: > > print -N -- **/*(.e:'reply=("$(ls -l "$REPLY")")':) | xargs -0 ... > > That's going to run "ls -l" once for each plain file (.) below the > current directory, so it's a lot more process-intensive than a single > "ls -lR", and it may not sort in the order you expect, but it will > give you "ls" output with $'\0'-terminated lines. > -- Kurtis Rader Caretaker of the exceptional canines Junior and Hank