Debian – difference between useradd -b and useradd -d

A tty is a native terminal device, the backend is either hardware or kernel emulated.

A pty (pseudo terminal device) is a terminal device which is emulated by an other program (example: xterm, screen, or ssh are such programs). A pts is the slave part of a pty.

(More info can be found in man pty.)

Short summary:

A pty is created by a process through posix_openpt() (which usually opens the special device /dev/ptmx), and is constituted by a pair of bidirectional character devices:

1. The master part, which is the file descriptor obtained by this process through this call, is used to emulate a terminal. After some initialization, the second part can be unlocked with unlockpt(), and the master is used to receive or send characters to this second part (slave).

2. The slave part, which is anchored in the filesystem as /dev/pts/x (the real name can be obtained by the master through ptsname() ) behaves like a native terminal device (/dev/ttyx). In most cases, a shell is started that uses it as a controlling terminal.

The behaviour you experience depends most likely on differences in the environment variable $PATH. The $PATH is essentially a colon-separated list of directories, which are searched in order for a particular executable when a program is invoked using anexec operating system call. The $PATH can contain relative path components, typically . or an empty string, which both refer to the current working directory. If the current directory is part of $PATH, files in the current working directory can be executed by just their name, e.g. a.out. If the current directory is not in $PATH, one must specify a relative or absolute path to the executable, e.g. ./a.out.

Having relative path components in $PATH has potential security implications as executables in directories earlier in $PATH overshadow executables in directories later in the list. Consider for example an attack on a system where the current working directory path . preceeds /bin in $PATH. If an attacker manages to place a malicious script sharing a name with a commonly used system utility, for instance ls, in the current directory (which typically is far easier that replacing binaries in root-owned /bin), the user will inadvertently invoke the malicious script when the intention is to invoke the system ls. Even if . is only appended at the end of $PATH, a user could be tricked to inadvertently invoke an executable in the current directory which shares a name with a common utility not found on that particular system. This is why it is common not to have relative path components as part of the default $PATH.