Hi Theo, Theo de Raadt: > This seems misguided. We have a horrible program called "file", but > in general people identify what a file is what what purpose it serves > not just by the filename, but also by how it starts. The "untrusted > comment" has become the way to identify a signify file. It has become > colloquial. > > Yes, there is a magic number immediately after that, but it is at > unknown byte offset. It isn't a offset-addressed file like gzip. So > your proposal doesn't actually help solve anything, in fact it > increases the ambiguity.
Nice catch, I haven't thought about file(1). I consider file(1) really useful. But at the same time it's way too empirical/phenomenological. Sometimes it's unable to reliably tell the truth. Only corresponding program/codec can tell if a file really contains data if this type. [in out case it's signify(1)]. Should we duplicate codec logic and put in into separate file(1)? Not the best idea. Can we include all the types info/magic just into one file(1) utility (/etc/magic)? No. Should we create filetypes to be detectable by file(1)? Don't think so. > So why not consider that call it a day, and leave it alone? Just because. -- Ivan Markin