On Tuesday 26 February 2008 21:03, Pádraig Brady <[EMAIL PROTECTED]> wrote: > > That depends on your definition of "works". > > If you don't mind retaining the first 2GiB of content in > > a preexisting output file, then it works fine. But the initial > > truncation is required if you want to be sure it's a big sparse > > file with nothing but NUL bytes. > > > > Try this: > > > > echo foo > k > > dd bs=1 seek=2G of=k < /dev/null > > head -c4 k > > > > and note that it prints "foo\n". > > That's how I would expect it to work, > and is how I'm currently implementing it. > Given the name "truncate" I think this > operation should be obvious to users.
What exactly do you mean by "how I expect it to work", do you mean "reducing the file size" or "removing all content and extending it to size N"? The latter is not truncation, it's replacement.