nanohttp.c

libxml,在UNIX/LINUX下非常重要的一个库,为XML相关应用提供方便.目前上载的是最新版本,若要取得最新版本,请参考里面的readme.

/*
 * nanohttp.c: minimalist HTTP GET implementation to fetch external subsets.
 *             focuses on size, streamability, reentrancy and portability
 *
 * This is clearly not a general purpose HTTP implementation
 * If you look for one, check:
 *         http://www.w3.org/Library/
 *
 * See Copyright for the status of this software.
 *
 * [e-mail]
 */
 
/* TODO add compression support, Send the Accept- , and decompress on the
        fly with ZLIB if found at compile-time */

#define NEED_SOCKETS
#define IN_LIBXML
#include "libxml.h"

#ifdef LIBXML_HTTP_ENABLED
#include <string.h>

#ifdef HAVE_STDLIB_H
#include <stdlib.h>
#endif
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_SYS_TYPES_H
#include <sys/types.h>
#endif
#ifdef HAVE_SYS_SOCKET_H
#include <sys/socket.h>
#endif
#ifdef HAVE_NETINET_IN_H
#include <netinet/in.h>
#endif
#ifdef HAVE_ARPA_INET_H
#include <arpa/inet.h>
#endif
#ifdef HAVE_NETDB_H
#include <netdb.h>
#endif
#ifdef HAVE_RESOLV_H
#ifdef HAVE_ARPA_NAMESER_H
#include <arpa/nameser.h>
#endif
#include <resolv.h>
#endif
#ifdef HAVE_FCNTL_H
#include <fcntl.h> 
#endif
#ifdef HAVE_ERRNO_H
#include <errno.h>
#endif
#ifdef HAVE_SYS_TIME_H
#include <sys/time.h>
#endif
#ifdef HAVE_SYS_SELECT_H
#include <sys/select.h>
#endif
#ifdef HAVE_STRINGS_H
#include <strings.h>
#endif
#ifdef SUPPORT_IP6
#include <resolv.h>
#endif

#ifdef VMS
#include <stropts>
#define XML_SOCKLEN_T unsigned int
#define SOCKET int
#endif


#ifdef __MINGW32__
#define _WINSOCKAPI_
#include <wsockcompat.h>
#include <winsock2.h>
#undef XML_SOCKLEN_T
#define XML_SOCKLEN_T unsigned int
#endif


#include <libxml/globals.h>
#include <libxml/xmlerror.h>
#include <libxml/xmlmemory.h>
#include <libxml/parser.h> /* for xmlStr(n)casecmp() */
#include <libxml/nanohttp.h>
#include <libxml/globals.h>
#include <libxml/uri.h>

/**
 * A couple portability macros
 */
#ifndef _WINSOCKAPI_
#ifndef __BEOS__
#define closesocket(s) close(s)
#endif
#define SOCKET int
#endif

#ifdef __BEOS__
#ifndef PF_INET
#define PF_INET AF_INET
#endif
#endif

#ifndef XML_SOCKLEN_T
#define XML_SOCKLEN_T unsigned int
#endif
#ifndef SOCKET
#define SOCKET int
#endif

#ifdef STANDALONE
#define DEBUG_HTTP
#define xmlStrncasecmp(a, b, n) strncasecmp((char *)a, (char *)b, n)
#define xmlStrcasecmpi(a, b) strcasecmp((char *)a, (char *)b)
#endif

#define XML_NANO_HTTP_MAX_REDIR	10

#define XML_NANO_HTTP_CHUNK	4096

#define XML_NANO_HTTP_CLOSED	0
#define XML_NANO_HTTP_WRITE	1
#define XML_NANO_HTTP_READ	2
#define XML_NANO_HTTP_NONE	4

typedef struct xmlNanoHTTPCtxt {
    char *protocol;	/* the protocol name */
    char *hostname;	/* the host name */
    int port;		/* the port */
    char *path;		/* the path within the URL */
    SOCKET fd;		/* the file descriptor for the socket */
    int state;		/* WRITE / READ / CLOSED */
    char *out;		/* buffer sent (zero terminated) */
    char *outptr;	/* index within the buffer sent */
    char *in;		/* the receiving buffer */
    char *content;	/* the start of the content */
    char *inptr;	/* the next byte to read from network */
    char *inrptr;	/* the next byte to give back to the client */
    int inlen;		/* len of the input buffer */
    int last;		/* return code for last operation */
    int returnValue;	/* the protocol return value */
    int ContentLength;  /* specified content length from HTTP header */
    char *contentType;	/* the MIME type for the input */
    char *location;	/* the new URL in case of redirect */
    char *authHeader;	/* contents of {WWW,Proxy}-Authenticate header */
    char *encoding;	/* encoding extracted from the contentType */
    char *mimeType;	/* Mime-Type extracted from the contentType */
} xmlNanoHTTPCtxt, *xmlNanoHTTPCtxtPtr;

static int initialized = 0;
static char *proxy = NULL;	 /* the proxy name if any */
static int proxyPort;	/* the proxy port if any */
static unsigned int timeout = 60;/* the select() timeout in seconds */

static int xmlNanoHTTPFetchContent( void * ctx, char ** ptr, int * len );

/**
 * xmlHTTPErrMemory:
 * @extra:  extra informations
 *
 * Handle an out of memory condition
 */
static void
xmlHTTPErrMemory(const char *extra)
{
    __xmlSimpleError(XML_FROM_HTTP, XML_ERR_NO_MEMORY, NULL, NULL, extra);
}

/**
 * A portability function
 */
static int socket_errno(void) {
#ifdef _WINSOCKAPI_
    return(WSAGetLastError());
#else
    return(errno);
#endif
}

#ifdef SUPPORT_IP6
static
int have_ipv6(void) {
    int s;

    s = socket (AF_INET6, SOCK_STREAM, 0);
    if (s != -1) {
	close (s);
	return (1);
    }
    return (0);
}
#endif

/**
 * xmlNanoHTTPInit:
 *
 * Initialize the HTTP protocol layer.
 * Currently it just checks for proxy informations
 */

void
xmlNanoHTTPInit(void) {
    const char *env;
#ifdef _WINSOCKAPI_
    WSADATA wsaData;    
#endif

    if (initialized)
	return;

#ifdef _WINSOCKAPI_
    if (WSAStartup(MAKEWORD(1, 1), &wsaData) != 0)
	return;
#endif

    if (proxy == NULL) {
	proxyPort = 80;
	env = getenv("no_proxy");
	if (env && ((env[0] == '*') && (env[1] == 0)))
	    goto done;
	env = getenv("http_proxy");
	if (env != NULL) {
	    xmlNanoHTTPScanProxy(env);
	    goto done;
	}
	env = getenv("HTTP_PROXY");
	if (env != NULL) {
	    xmlNanoHTTPScanProxy(env);
	    goto done;
	}
    }
done:
    initialized = 1;
}

/**
 * xmlNanoHTTPCleanup:
 *
 * Cleanup the HTTP protocol layer.
 */

void
xmlNanoHTTPCleanup(void) {
    if (proxy != NULL)
	xmlFree(proxy);
#ifdef _WINSOCKAPI_
    if (initialized)
	WSACleanup();
#endif
    initialized = 0;
    return;
}

/**
 * xmlNanoHTTPScanURL:
 * @ctxt:  an HTTP context
 * @URL:  The URL used to initialize the context
 *
 * (Re)Initialize an HTTP context by parsing the URL and finding
 * the protocol host port and path it indicates.
 */

static void
xmlNanoHTTPScanURL(xmlNanoHTTPCtxtPtr ctxt, const char *URL) {
    xmlURIPtr uri;
    /*
     * Clear any existing data from the context
     */
    if (ctxt->protocol != NULL) { 
        xmlFree(ctxt->protocol);
	ctxt->protocol = NULL;
    }
    if (ctxt->hostname != NULL) { 
        xmlFree(ctxt->hostname);
	ctxt->hostname = NULL;
    }
    if (ctxt->path != NULL) { 
        xmlFree(ctxt->path);
	ctxt->path = NULL;
    }
    if (URL == NULL) return;

    uri = xmlParseURI(URL);
    if (uri == NULL)
	return;

    if ((uri->scheme == NULL) || (uri->server == NULL)) {
	xmlFreeURI(uri);
	return;
    }
    
    ctxt->protocol = xmlMemStrdup(uri->scheme);
    ctxt->hostname = xmlMemStrdup(uri->server);
    if (uri->path != NULL)
	ctxt->path = xmlMemStrdup(uri->path);
    else
	ctxt->path = xmlMemStrdup("/");
    if (uri->port != 0)
	ctxt->port = uri->port;

    xmlFreeURI(uri);
}

/**
 * xmlNanoHTTPScanProxy:
 * @URL:  The proxy URL used to initialize the proxy context
 *
 * (Re)Initialize the HTTP Proxy context by parsing the URL and finding
 * the protocol host port it indicates.
 * Should be like http://myproxy/ or http://myproxy:3128/
 * A NULL URL cleans up proxy informations.
 */

void
xmlNanoHTTPScanProxy(const char *URL) {
    xmlURIPtr uri;

    if (proxy != NULL) { 
        xmlFree(proxy);
	proxy = NULL;
    }
    proxyPort = 0;

#ifdef DEBUG_HTTP
    if (URL == NULL)
	xmlGenericError(xmlGenericErrorContext,
		"Removing HTTP proxy info\n");
    else
	xmlGenericError(xmlGenericErrorContext,
		"Using HTTP proxy %s\n", URL);
#endif
    if (URL == NULL) return;

    uri = xmlParseURI(URL);
    if ((uri == NULL) || (uri->scheme == NULL) ||
	(strcmp(uri->scheme, "http")) || (uri->server == NULL)) {
	__xmlIOErr(XML_FROM_HTTP, XML_HTTP_URL_SYNTAX, "Syntax Error\n");
	if (uri != NULL)
	    xmlFreeURI(uri);
	return;
    }
    
    proxy = xmlMemStrdup(uri->server);
    if (uri->port != 0)
	proxyPort = uri->port;

    xmlFreeURI(uri);
}

/**
 * xmlNanoHTTPNewCtxt:
 * @URL:  The URL used to initialize the context
 *
 * Allocate and initialize a new HTTP context.
 *
 * Returns an HTTP context or NULL in case of error.
 */

static xmlNanoHTTPCtxtPtr
xmlNanoHTTPNewCtxt(const char *URL) {
    xmlNanoHTTPCtxtPtr ret;

    ret = (xmlNanoHTTPCtxtPtr) xmlMalloc(sizeof(xmlNanoHTTPCtxt));
    if (ret == NULL) {
        xmlHTTPErrMemory("allocating context");
        return(NULL);
    }

    memset(ret, 0, sizeof(xmlNanoHTTPCtxt));
    ret->port = 80;
    ret->returnValue = 0;
    ret->fd = -1;
    ret->ContentLength = -1;

    xmlNanoHTTPScanURL(ret, URL);

    return(ret);
}

/**
 * xmlNanoHTTPFreeCtxt:
 * @ctxt:  an HTTP context
 *
 * Frees the context after closing the connection.
 */

static void
xmlNanoHTTPFreeCtxt(xmlNanoHTTPCtxtPtr ctxt) {
    if (ctxt == NULL) return;
    if (ctxt->hostname != NULL) xmlFree(ctxt->hostname);
    if (ctxt->protocol != NULL) xmlFree(ctxt->protocol);
    if (ctxt->path != NULL) xmlFree(ctxt->path);
    if (ctxt->out != NULL) xmlFree(ctxt->out);
    if (ctxt->in != NULL) xmlFree(ctxt->in);
    if (ctxt->contentType != NULL) xmlFree(ctxt->contentType);
    if (ctxt->encoding != NULL) xmlFree(ctxt->encoding);
    if (ctxt->mimeType != NULL) xmlFree(ctxt->mimeType);
    if (ctxt->location != NULL) xmlFree(ctxt->location);
    if (ctxt->authHeader != NULL) xmlFree(ctxt->authHeader);
    ctxt->state = XML_NANO_HTTP_NONE;
    if (ctxt->fd >= 0) closesocket(ctxt->fd);
    ctxt->fd = -1;
    xmlFree(ctxt);
}

/**
 * xmlNanoHTTPSend:
 * @ctxt:  an HTTP context
 *
 * Send the input needed to initiate the processing on the server side
 * Returns number of bytes sent or -1 on error.
 */

static int
xmlNanoHTTPSend(xmlNanoHTTPCtxtPtr ctxt, const char * xmt_ptr, int outlen) {

    int 	total_sent = 0;

    if ( (ctxt->state & XML_NANO_HTTP_WRITE) && (xmt_ptr != NULL ) ) {
        while (total_sent < outlen) {
            int nsent = send(ctxt->fd, xmt_ptr + total_sent,
                                      outlen - total_sent, 0);
            if (nsent>0)
                total_sent += nsent;
	    else if ( ( nsent == -1 ) && 
#if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
	    	      ( socket_errno( ) != EAGAIN ) &&
#endif
		        ( socket_errno( ) != EWOULDBLOCK ) ) {
		__xmlIOErr(XML_FROM_HTTP, 0, "send failed\n");
		if ( total_sent == 0 )
		    total_sent = -1;
		break;
	    }
	    else {
	        /*
		**  No data sent
		**  Since non-blocking sockets are used, wait for 
		**  socket to be writable or default timeout prior
		**  to retrying.
		*/

		struct timeval	tv;
		fd_set		wfd;

		tv.tv_sec = timeout;
		tv.tv_usec = 0;
		FD_ZERO( &wfd );
		FD_SET( ctxt->fd, &wfd );
		(void)select( ctxt->fd + 1, NULL, &wfd, NULL, &tv );
	    }
	}
    }

    return total_sent;
}

/**
 * xmlNanoHTTPRecv:
 * @ctxt:  an HTTP context
 *
 * Read information coming from the HTTP connection.
 * This is a blocking call (but it blocks in select(), not read()).
 *
 * Returns the number of byte read or -1 in case of error.
 */

static int
xmlNanoHTTPRecv(xmlNanoHTTPCtxtPtr ctxt) {
    fd_set rfd;
    struct timeval tv;


    while (ctxt->state & XML_NANO_HTTP_READ) {
	if (ctxt->in == NULL) {
	    ctxt->in = (char *) xmlMallocAtomic(65000 * sizeof(char));
	    if (ctxt->in == NULL) {
		xmlHTTPErrMemory("allocating input");
	        ctxt->last = -1;
		return(-1);
	    }
	    ctxt->inlen = 65000;
	    ctxt->inptr = ctxt->content = ctxt->inrptr = ctxt->in;
	}
	if (ctxt->inrptr > ctxt->in + XML_NANO_HTTP_CHUNK) {
	    int delta = ctxt->inrptr - ctxt->in;
	    int len = ctxt->inptr - ctxt->inrptr;
	    
	    memmove(ctxt->in, ctxt->inrptr, len);
	    ctxt->inrptr -= delta;
	    ctxt->content -= delta;
	    ctxt->inptr -= delta;
	}
        if ((ctxt->in + ctxt->inlen) < (ctxt->inptr + XML_NANO_HTTP_CHUNK)) {
	    int d_inptr = ctxt->inptr - ctxt->in;
	    int d_content = ctxt->content - ctxt->in;
	    int d_inrptr = ctxt->inrptr - ctxt->in;
	    char *	tmp_ptr = ctxt->in;

	    ctxt->inlen *= 2;
            ctxt->in = (char *) xmlRealloc(tmp_ptr, ctxt->inlen);
	    if (ctxt->in == NULL) {
		xmlHTTPErrMemory("allocating input buffer");
		xmlFree( tmp_ptr );
	        ctxt->last = -1;
		return(-1);
	    }
            ctxt->inptr = ctxt->in + d_inptr;
            ctxt->content = ctxt->in + d_content;
            ctxt->inrptr = ctxt->in + d_inrptr;
	}
	ctxt->last = recv(ctxt->fd, ctxt->inptr, XML_NANO_HTTP_CHUNK, 0);
	if (ctxt->last > 0) {
	    ctxt->inptr += ctxt->last;
	    return(ctxt->last);
	}
	if (ctxt->last == 0) {
	    return(0);
	}
	if (ctxt->last == -1) {
	    switch (socket_errno()) {
		case EINPROGRESS:
		case EWOULDBLOCK:
#if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
		case EAGAIN:
#endif
		    break;

		case ECONNRESET:
		case ESHUTDOWN:
		    return ( 0 );

		default:
		    __xmlIOErr(XML_FROM_HTTP, 0, "recv failed\n");
		    return(-1);
	    }
	}

	tv.tv_sec = timeout;
	tv.tv_usec = 0;
	FD_ZERO(&rfd);
	FD_SET(ctxt->fd, &rfd);
	
	if ( (select(ctxt->fd+1, &rfd, NULL, NULL, &tv)<1)
#if defined(EINTR)
		&& (errno != EINTR)
#endif
	)
		return(0);
    }
    return(0);
}

/**