What a file is made of

You saw some of it back in chapter 1. There was quite a lot packed into one line of ls -l.

The - on the far left is the mark for "an ordinary file" (a box would be d). After that come the permissions, then 1 is the link count, learner is the owner and the group, 6 is the size, and then the time it was last changed.

stat is the tool that spreads all of that out instead of cramming it into one line.

Give stat a file name and you get something like this.

Size is the size and regular file is the kind. Inode is a number attached to the file, which identifies it separately from its name (that gets used in the chapter on the filesystem).

The interesting part is that there are three times: Access, Modify and Change. It remembers when it was read, when the contents changed, and when the information changed, each on its own.

When you do not want all of it, -c gets you just the piece you asked for. You say which with a short code starting with %.

• %s — size, in bytes
• %F — kind (regular file / directory)
• %a — permissions in octal (644)
• %h — link count
• %n — name

What comes out is the bare number, which is the good part: you can hand it straight to a calculation or the next command.

cat turns the contents into something a person can read before it puts them out. But sometimes you want to know what is really in there.

od -c lays it out one byte at a time. od is short for octal dump.

Look and you can see that café is c a f 303 251 — the plain letters take one byte each, and the é takes two. You can check that with your own eyes.

The same job in hexadecimal is hexdump -C and xxd.

Down the right you get "the character if it is readable, a . if it is not", which often makes them easier to read than od.

The point to notice is that a newline shows up as 0a. Characters you never normally see can be checked this way.

An ordinary Linux has a command called file that looks at the contents and works out what kind of thing it is. Type file photo.jpg and it tells you JPEG image data.

But it is not in this environment. Working out a kind from the contents needs about 10MB of lookup tables, which will not fit in something sent to a browser.

So here we use stat (the facts about it) and od (the bytes). When you graduate and type on your own computer, remember that file is available too.

The times and the Inode number change every time you type, so marking never uses them.

All it looks at are the values that are fixed — size, kind, permissions and link count. Type away.