This section describes the Common Lisp interface to file systems. The model used by this interface assumes that files are named by filenames, that a filename can be represented by a pathname object, and that given a pathname a stream can be constructed that connects to a file whose filename it represents.
For information about opening and closing files, and manipulating their contents, see Chapter 21 (Streams).
Figure 20–1 lists some operators that are applicable to files and directories.
A stream associated with a file is either a file stream or a synonym stream whose target is a stream associated with a file. Such streams can be used as pathname designators.
Normally, when a stream associated with a file is used as a pathname designator, it denotes the pathname used to open the file; this may be, but is not required to be, the actual name of the file.
Some functions, such as truename and delete-file, coerce streams to pathnames in a different way that involves referring to the actual file that is open, which might or might not be the file whose name was opened originally. Such special situations are always notated specifically and are not the default.
Many functions that perform file operations accept either open or closed streams as arguments; see Section 21.1.3 (Stream Arguments to Standardized Functions).
Of these, the functions in Figure 20–2 treat open and closed streams differently.
Since treatment of open streams by the file system may vary considerably between implementations, however, a closed stream might be the most reliable kind of argument for some of these functions — in particular, those in Figure 20–3. For example, in some file systems, open files are written under temporary names and not renamed until closed and/or are held invisible until closed. In general, any code that is intended to be portable should use such functions carefully.
Many file systems permit more than one filename to designate a particular file.
Even where multiple names are possible, most file systems have a convention for generating a canonical filename in such situations. Such a canonical filename (or the pathname representing such a filename) is called a truename.
The truename of a file may differ from other filenames for the file because of symbolic links, version numbers, logical device translations in the file system, logical pathname translations within Common Lisp, or other artifacts of the file system.
The truename for a file is often, but not necessarily, unique for each file. For instance, a Unix file with multiple hard links could have several truenames.
For example, a DEC TOPS-20 system with files
PS:<JOE>FOO.TXT.2 might permit the second file to be referred to as
PS:<JOE>FOO.TXT.0, since the “
.0” notation denotes “newest” version of several files. In the same file system, a “logical device” “
JOE:” might be taken to refer to
PS:<JOE>” and so the names
JOE:FOO.TXT.0 might refer to
PS:<JOE>FOO.TXT.2. In all of these cases, the truename of the file would probably be
If a file is a symbolic link to another file (in a file system permitting such a thing), it is conventional for the truename to be the canonical name of the file after any symbolic links have been followed; that is, it is the canonical name of the file whose contents would become available if an input stream to that file were opened.
In the case of a file still being created (that is, of an output stream open to such a file), the exact truename of the file might not be known until the stream is closed. In this case, the function truename might return different values for such a stream before and after it was closed. In fact, before it is closed, the name returned might not even be a valid name in the file system — for example, while a file is being written, it might have version :newest and might only take on a specific numeric value later when the file is closed even in a file system where all files have numeric versions.