Declaration of W3C Sample Code Library Streams

/*
**	(c) COPYRIGHT MIT 1995.
**	Please first read the full copyright statement in the file COPYRIGH.
*/

The stream part of the Library is a set of streams that can be used in many contexts throughout the Library. It also contains conversions between streams and other memory containers, for example Chunks.

#ifndef WWWSTREAM_H
#define WWWSTREAM_H
#ifdef __cplusplus
extern "C" { 
#endif

System dependencies

The wwwsys.h file includes system-specific include files and flags for I/O to network and disk. The only reason for this file is that the Internet world is more complicated than Posix and ANSI.

#include "wwwsys.h"

Content Length Counter

This stream also buffers the result to find out the content length. If a maximum buffer limit is reached Content-Length is calculated for logs but it is not sent to the client -- rather the buffer is flushed right away.

#include "HTConLen.h"			/* Content Length Counter */

File Writer Streams

This module contains a set of basic file writer streams that are used to dump data objects to disk at various places within the Library core. Most notably, we use these streams in the Format Manager in order to handle external presenters, for example post script viewers etc. These streams can of course also be used in other contexts by the application.

#include "HTFWrite.h"
#include "HTFSave.h"

Content Guess Stream

This interface provides functionality for guessing unknown media types from magic words. The stream is a one that reads first a chunk of stuff, tries to figure out the format, and calls HTStreamStack(). This is a kind of lazy-evaluation of HTStreamStack(). This could be extended arbitrarily to recognize all the possible file formats in the world, if someone only had time to do it.

#include "HTGuess.h"			/* Guess stream */

Tee Stream

The Tee stream just writes everything you put into it into two oter streams. One use (the only use?!) is for taking a cached copey on disk while loading the main copy, without having to wait for the disk copy to be finished and reread it.

#include "HTTee.h"

Merge Stream

The Merge stream can be used to merge multiple streams into a single target stream. The Merge stream does not prevent any of the streams from writing and no ordering is imposed. The main feature is basically that the free and abort methods can be called n times where n equals the number of feeds that put data to the stream.

#include "HTMerge.h"

Stream to Chunk Conversion

If you do not like the stream model in libwww, then you can use this stream to convert a stream object into a Chunk object which is a dynamic character string buffer in memory.

#include "HTSChunk.h"

Stream to Memory Object Conversion

This version of the stream object is a hook for clients that want an unparsed stream from libwww. The HTXParse_put_* and HTXParse_write routines copy the content of the incoming buffer into a buffer that is realloced whenever necessary. This buffer is handed over to the client in HTXParse_free.

#include "HTXParse.h"			/* External parse stream */
#include "HTEPtoCl.h"			/* Client callbacks */

Network Telnet To Internal Character Text

This is a filter stream suitable for taking text from a socket and passing it into a stream which expects text in the local C representation.

#include "HTNetTxt.h"
#ifdef __cplusplus
} /* end extern C definitions */
#endif

#endif


@(#) $Id: WWWStream.html,v 2.10 1999/02/14 04:00:10 frystyk Exp $