A stream is an object that can be used with an input or output function to identify an appropriate source or sink of characters or bytes for that operation. A character stream is a source or sink of characters. A binary stream is a source or sink of bytes.
Some operations may be performed on any kind of stream; Figure 21–1 provides a list of standardized operations that are potentially useful with any kind of stream.
| close | stream-element-type |
| input-stream-p | streamp |
| interactive-stream-p | with-open-stream |
| output-stream-p |
Other operations are only meaningful on certain stream types. For example, read-char is only defined for character streams and read-byte is only defined for binary streams.
A stream, whether a character stream or a binary stream, can be an input stream (source of data), an output stream (sink for data), both, or (e.g., when “:direction :probe” is given to open) neither.
Figure 21–2 shows operators relating to input streams.
| clear-input | read-byte | read-from-string |
| listen | read-char | read-line |
| peek-char | read-char-no-hang | read-preserving-whitespace |
| read | read-delimited-list | unread-char |
Figure 21–3 shows operators relating to output streams.
| clear-output | prin1 | write |
| finish-output | prin1-to-string | write-byte |
| force-output | princ | write-char |
| format | princ-to-string | write-line |
| fresh-line | write-string | |
| pprint | terpri | write-to-string |
A stream that is both an input stream and an output stream is called a bidirectional stream. See the functions input-stream-p and output-stream-p.
Any of the operators listed in Figure 21–2 or Figure 21–3 can be used with bidirectional streams. In addition, Figure 21–4 shows a list of operators that relate specificaly to bidirectional streams.
Streams are either open or closed.
Except as explicitly specified otherwise, operations that create and return streams return open streams.
The action of closing a stream marks the end of its use as a source or sink of data, permitting the implementation to reclaim its internal data structures, and to free any external resources which might have been locked by the stream when it was opened.
Except as explicitly specified otherwise, the consequences are undefined when a closed stream is used where a stream is called for.
Coercion of streams to pathnames is permissible for closed streams; in some situations, such as for a truename computation, the result might be different for an open stream and for that same stream once it has been closed.
An interactive stream is one on which it makes sense to perform interactive querying.
The precise meaning of an interactive stream is implementation-defined, and may depend on the underlying operating system. Some examples of the things that an implementation might choose to use as identifying characteristics of an interactive stream include:
The stream is connected to a person (or equivalent) in such a way that the program can prompt for information and expect to receive different input depending on the prompt.
The program is expected to prompt for input and support “normal input editing”.
read-char might wait for the user to type something before returning instead of immediately returning a character or end-of-file.
The general intent of having some streams be classified as interactive streams is to allow them to be distinguished from streams containing batch (or background or command-file) input. Output to batch streams is typically discarded or saved for later viewing, so interactive queries to such streams might not have the expected effect.
Terminal I/O might or might not be an interactive stream.
Some streams, called file streams, provide access to files. An object of class file-stream is used to represent a file stream.
The basic operation for opening a file is open, which typically returns a file stream (see its dictionary entry for details). The basic operation for closing a stream is close. The macro with-open-file is useful to express the common idiom of opening a file for the duration of a given body of code, and assuring that the resulting stream is closed upon exit from that body.
The class stream has a number of subclasses defined by this specification. Figure 21–5 shows some information about these subclasses.
Variables whose values must be streams are sometimes called stream variables.
Certain stream variables are defined by this specification to be the proper source of input or output in various situations where no specific stream has been specified instead. A complete list of such standardized stream variables appears in Figure 21–6. The consequences are undefined if at any time the value of any of these variables is not an open stream.
Note that, by convention, standardized stream variables have names ending in “-input*” if they must be input streams, ending in “-output*” if they must be output streams, or ending in “-io*” if they must be bidirectional streams.
User programs may assign or bind any standardized stream variable except *terminal-io*.
The operators in Figure 21–7 accept stream arguments that might be either open or closed streams.
The operators in Figure 21–8 accept stream arguments that must be open streams.
The consequences are undefined if any component of a composite stream is closed before the composite stream is closed.
The consequences are undefined if the synonym stream symbol is not bound to an open stream from the time of the synonym stream’s creation until the time it is closed.