ncbi_connutil.h

ncbi源码

 *       and/or read/write timeouts were passed as {0,0}, meaning that both
 *       connection and HTTP header write operation may still be pending in
 *       the resultant socket. It is responsibility of the application to
 *       analyze the actual socket state in this case (see "ncbi_socket.h").
 */

extern NCBI_XCONNECT_EXPORT SOCK URL_Connect
(const char*     host,
 unsigned short  port,
 const char*     path,
 const char*     args,
 EReqMethod      req_method,
 size_t          content_length,
 const STimeout* c_timeout,       /* timeout for the CONNECT stage          */
 const STimeout* rw_timeout,      /* timeout for READ and WRITE             */
 const char*     user_header,
 int/*bool*/     encode_args,     /* URL-encode the "args", if any          */
 ESwitch         data_logging     /* sock.data log.; eDefault in most cases */
 );


/* Discard all input data before(and including) the first occurrence of
 * "pattern". If "buf" is not NULL then add the discarded data(including
 * the "pattern") to it. If "n_discarded" is not NULL then "*n_discarded"
 * will return # of discarded bytes.
 * NOTE: "pattern" == NULL causes stripping to the EOF.
 */
extern NCBI_XCONNECT_EXPORT EIO_Status CONN_StripToPattern
(CONN        conn,
 const void* pattern,
 size_t      pattern_size,
 BUF*        buf,
 size_t*     n_discarded
 );

extern NCBI_XCONNECT_EXPORT EIO_Status SOCK_StripToPattern
(SOCK        sock,
 const void* pattern,
 size_t      pattern_size,
 BUF*        buf,
 size_t*     n_discarded
 );

extern NCBI_XCONNECT_EXPORT EIO_Status BUF_StripToPattern
(BUF         buffer,
 const void* pattern,
 size_t      pattern_size,
 BUF*        buf,
 size_t*     n_discarded
 );


/* URL-decode up to "src_size" symbols(bytes) from buffer "src_buf".
 * Write the decoded data to buffer "dst_buf", but no more than "dst_size"
 * bytes.
 * Assign "*src_read" to the # of bytes successfully decoded from "src_buf".
 * Assign "*dst_written" to the # of bytes written to buffer "dst_buf".
 * Return FALSE only if cannot decode nothing, and an unrecoverable
 * URL-encoding error (such as an invalid symbol or a bad "%.." sequence)
 * has occurred.
 * NOTE:  the unfinished "%.." sequence is fine -- return TRUE, but dont
 *        "read" it.
 */
extern NCBI_XCONNECT_EXPORT int/*bool*/ URL_Decode
(const void* src_buf,    /* [in]     non-NULL */
 size_t      src_size,   /* [in]              */
 size_t*     src_read,   /* [out]    non-NULL */
 void*       dst_buf,    /* [in/out] non-NULL */
 size_t      dst_size,   /* [in]              */
 size_t*     dst_written /* [out]    non-NULL */
 );


/* Act just like URL_Decode (see above) but caller can allow the specified
 * non-standard URL symbols in the input buffer to be decoded "as is".
 * The extra allowed symbols are passed in a '\0'-terminated string
 * "allow_symbols" (it can be NULL or empty -- then this will be an exact
 * equivalent of URL_Decode).
 */
extern NCBI_XCONNECT_EXPORT int/*bool*/ URL_DecodeEx
(const void* src_buf,      /* [in]     non-NULL  */
 size_t      src_size,     /* [in]               */
 size_t*     src_read,     /* [out]    non-NULL  */
 void*       dst_buf,      /* [in/out] non-NULL  */
 size_t      dst_size,     /* [in]               */
 size_t*     dst_written,  /* [out]    non-NULL  */
 const char* allow_symbols /* [in]     '\0'-term */
 );


/* URL-encode up to "src_size" symbols(bytes) from buffer "src_buf".
 * Write the encoded data to buffer "dst_buf", but no more than "dst_size"
 * bytes.
 * Assign "*src_read" to the # of bytes successfully encoded from "src_buf".
 * Assign "*dst_written" to the # of bytes written to buffer "dst_buf".
 */
extern NCBI_XCONNECT_EXPORT void URL_Encode
(const void* src_buf,    /* [in]     non-NULL */
 size_t      src_size,   /* [in]              */
 size_t*     src_read,   /* [out]    non-NULL */
 void*       dst_buf,    /* [in/out] non-NULL */
 size_t      dst_size,   /* [in]              */
 size_t*     dst_written /* [out]    non-NULL */
 );



/****************************************************************************
 * NCBI-specific MIME content type and sub-types
 * (the API to compose and parse them)
 *    Content-Type: <type>/<MIME_ComposeSubType()>\r\n
 *
 *    Content-Type: <type>/<subtype>-<encoding>\r\n
 *
 * where  MIME_ComposeSubType(EMIME_SubType subtype, EMIME_Encoding encoding):
 *   "x-<subtype>-<encoding>":
 *     "x-<subtype>",   "x-<subtype>-urlencoded",   "x-<subtype>-<encoding>",
 *     "x-dispatch",    "x-dispatch-urlencoded",    "x-dispatch-<encoding>
 *     "x-asn-text",    "x-asn-text-urlencoded",    "x-asn-text-<encoding>
 *     "x-asn-binary",  "x-asn-binary-urlencoded",  "x-asn-binary-<encoding>"
 *     "x-www-form",    "x-www-form-urlencoded",    "x-www-form-<encoding>"
 *     "html",          "html-urlencoded",          "html-<encoding>"
 *     "x-unknown",     "x-unknown-urlencoded",     "x-unknown-<encoding>"
 *
 *  Note:  <subtype> and <encoding> are expected to contain only
 *         alphanumeric symbols, '-' and '_'. They are case-insensitive.
 ****************************************************************************/


/* Type
 */
typedef enum {
    eMIME_T_NcbiData = 0,  /* "x-ncbi-data"  (NCBI specific data) */
    eMIME_T_Text,          /* "text"                              */
    eMIME_T_Application,   /* "application"                       */
    /* eMIME_T_???, "<type>" here go other types                  */
    eMIME_T_Unknown        /* "unknown"                           */
} EMIME_Type;


/* SubType
 */
typedef enum {
    eMIME_Dispatch = 0,  /* "x-dispatch"    (dispatcher info)          */
    eMIME_AsnText,       /* "x-asn-text"    (text ASN.1 data)          */
    eMIME_AsnBinary,     /* "x-asn-binary"  (binary ASN.1 data)        */
    eMIME_Fasta,         /* "x-fasta"       (data in FASTA format)     */
    eMIME_WwwForm,       /* "x-www-form"                               */
    /* standard MIMEs */
    eMIME_Html,          /* "html"                                     */
    eMIME_Plain,         /* "plain"                                    */
    eMIME_Xml,           /* "xml"                                      */
    eMIME_XmlSoap,       /* "xml+soap"                                 */
    /* eMIME_???,           "<subtype>" here go other NCBI subtypes    */
    eMIME_Unknown        /* "x-unknown"     (an arbitrary binary data) */
} EMIME_SubType;


/* Encoding
 */
typedef enum {
    eENCOD_None = 0, /* ""              (the content is passed "as is") */
    eENCOD_Url,      /* "-urlencoded"   (the content is URL-encoded)    */
    /* eENCOD_???,      "-<encoding>" here go other NCBI encodings      */
    eENCOD_Unknown   /* "-encoded"      (unknown encoding)              */
} EMIME_Encoding;


/* Write up to "buflen" bytes to "buf":
 *   Content-Type: <type>/[x-]<subtype>-<encoding>\r\n
 * Return pointer to the "buf".
 */
#define MAX_CONTENT_TYPE_LEN 64
extern NCBI_XCONNECT_EXPORT char* MIME_ComposeContentTypeEx
(EMIME_Type     type,
 EMIME_SubType  subtype,
 EMIME_Encoding encoding,
 char*          buf,
 size_t         buflen    /* must be at least MAX_CONTENT_TYPE_LEN */
 );

/* Exactly equivalent to MIME_ComposeContentTypeEx(eMIME_T_NcbiData, ...)
 */
extern NCBI_XCONNECT_EXPORT char* MIME_ComposeContentType
(EMIME_SubType  subtype,
 EMIME_Encoding encoding,
 char*          buf,
 size_t         buflen
 );


/* Parse the NCBI-specific content-type; the (case-insensitive) "str"
 * can be in the following two formats:
 *   Content-Type: <type>/x-<subtype>-<encoding>
 *   <type>/x-<subtype>-<encoding>
 *
 * NOTE:  all leading spaces and all trailing spaces (and any trailing symbols,
 *        if they separated from the content type by at least one space) will
 *        be ignored, e.g. these are valid content type strings:
 *           "   Content-Type: text/plain  foobar"
 *           "  text/html \r\n  barfoo coocoo ....\n boooo"
 *
 * If it does not match any of NCBI MIME type/subtypes/encodings, then
 * return TRUE, eMIME_T_Unknown, eMIME_Unknown or eENCOD_None, respectively.
 * If the passed "str" has an invalid (non-HTTP ContentType) format
 * (or if it is NULL/empty), then
 * return FALSE, eMIME_T_Unknown, eMIME_Unknown, and eENCOD_Unknown
 */
extern NCBI_XCONNECT_EXPORT int/*bool*/ MIME_ParseContentTypeEx
(const char*     str,      /* the HTTP "Content-Type:" header to parse */
 EMIME_Type*     type,     /* can be NULL */
 EMIME_SubType*  subtype,  /* can be NULL */
 EMIME_Encoding* encoding  /* can be NULL */
 );

/* Requires the MIME type be "x-ncbi-data"
 */
extern NCBI_XCONNECT_EXPORT int/*bool*/ MIME_ParseContentType
(const char*     str,      /* the HTTP "Content-Type:" header to parse */
 EMIME_SubType*  subtype,  /* can be NULL */
 EMIME_Encoding* encoding  /* can be NULL */
 );


/* Read (skipping leading blanks) "[host][:port]" from a string.
 * On success, return the advanced pointer past the host/port read.
 * If no host/port detected, return 'str'.
 * On format error, return 0.
 * If host and/or port fragments are missing,
 * then corresponding 'host'/'port' value returned as 0.
 * Note that 'host' returned is in network byte order,
 * unlike 'port', which always comes out in host (native) byte order.
 */
extern NCBI_XCONNECT_EXPORT const char* StringToHostPort
(const char*     str,   /* must not be NULL */
 unsigned int*   host,  /* must not be NULL */
 unsigned short* port   /* must not be NULL */
 );


/* Print host:port into provided buffer string, not to exceed 'buflen' size.
 * Suppress printing host if parameter 'host' is zero.
 * Return the number of bytes printed.
 */
extern NCBI_XCONNECT_EXPORT size_t HostPortToString
(unsigned int   host,
 unsigned short port,
 char*          buf,
 size_t         buflen
 );


#ifdef __cplusplus
}  /* extern "C" */
#endif


/* @} */


/*
 * --------------------------------------------------------------------------
 * $Log: ncbi_connutil.h,v $
 * Revision 1000.1  2004/02/12 21:51:37  gouriano
 * PRODUCTION: UPGRADED [CORE_001] Dev-tree R6.35
 *
 * Revision 6.35  2004/01/14 18:51:41  lavr
 * +eMIME_XmlSoap
 *
 * Revision 6.34  2004/01/07 19:24:40  lavr
 * Added MIME subtype eMIME_Xml
 *
 * Revision 6.33  2003/09/23 21:00:33  lavr
 * Reorder included header files
 *
 * Revision 6.32  2003/08/25 14:48:50  lavr
 * ConnNetInfo_SetUserHeader():  to return completion status
 *
 * Revision 6.31  2003/05/29 17:56:53  lavr
 * More (clarified) comments for URL_Connect()
 *
 * Revision 6.30  2003/05/20 21:24:01  lavr
 * Limit SConnNetInfo::max_try by reasonable "short" value
 *
 * Revision 6.29  2003/04/09 17:58:47  siyan
 * Added doxygen support
 *
 * Revision 6.28  2003/01/17 19:44:20  lavr
 * Reduce dependencies
 *
 * Revision 6.27  2003/01/08 01:59:32  lavr
 * DLL-ize CONNECT library for MSVC (add NCBI_XCONNECT_EXPORT)
 *
 * Revision 6.26  2002/11/19 19:19:24  lavr
 * +ConnNetInfo_ExtendUserHeader()
 *
 * Revision 6.25  2002/11/12 05:49:47  lavr
 * Expand host names to hold 256 chars (instead of 64)
 *
 * Revision 6.24  2002/10/21 18:30:27  lavr
 * +ConnNetInfo_AppendArg()
 * +ConnNetInfo_PrependArg()
 * +ConnNetInfo_DeleteArg()
 * +ConnNetInfo_PreOverrideArg()
 * +ConnNetInfo_PostOverrideArg()
 *
 * Revision 6.23  2002/10/11 19:41:40  lavr
 * +ConnNetInfo_AppendUserHeader()
 * +ConnNetInfo_OverrideUserHeader()
 * +ConnNetInfo_DeleteUserHeader()
 *
 * Revision 6.22  2002/09/19 18:00:21  lavr
 * Header file guard macro changed; log moved to the end
 *
 * Revision 6.21  2002/05/06 19:07:25  lavr
 * -#include <stdlib>; -ConnNetInfo_Print(); +ConnNetInfo_Log()
 *
 * Revision 6.20  2002/02/20 19:12:03  lavr
 * Swapped eENCOD_Url and eENCOD_None; eENCOD_Unknown introduced
 *
 * Revision 6.19  2001/12/30 19:39:36  lavr
 * +ConnNetInfo_ParseURL()
 *
 * Revision 6.18  2001/09/28 20:45:26  lavr
 * SConnNetInfo::max_try equal to 0 is now treated the same way as equal to 1
 *
 * Revision 6.17  2001/09/19 15:58:37  lavr
 * Cut trailing blanks in blank lines
 *
 * Revision 6.16  2001/09/10 21:14:47  lavr
 * Added functions: StringToHostPort()
 *                  HostPortToString()
 *
 * Revision 6.15  2001/06/01 16:01:58  vakatov
 * MIME_ParseContentTypeEx() -- extended description
 *
 * Revision 6.14  2001/05/29 21:15:42  vakatov
 * + eMIME_Plain
 *
 * Revision 6.13  2001/04/24 21:21:38  lavr
 * Special text value "infinite" accepted as infinite timeout from environment
 *
 * Revision 6.12  2001/03/07 23:00:15  lavr
 * Default value for SConnNetInfo::stateless set to empty (FALSE)
 *
 * Revision 6.11  2001/03/02 20:07:07  lavr
 * Typos fixed
 *
 * Revision 6.10  2001/02/26 16:56:41  vakatov
 * Comment SConnNetInfo.
 *
 * Revision 6.9  2001/01/23 23:06:15  lavr
 * SConnNetInfo.debug_printout converted from boolean to enum
 * BUF_StripToPattern() introduced
 *
 * Revision 6.8  2001/01/11 23:05:13  lavr
 * ConnNetInfo_Create() fully documented
 *
 * Revision 6.7  2001/01/08 23:46:10  lavr
 * REQUEST_METHOD -> REQ_METHOD to be consistent with SConnNetInfo
 *
 * Revision 6.6  2001/01/08 22:47:13  lavr
 * ReqMethod constants changed (to conform to coding standard)
 * ClientMode removed; replaced by 2 booleans: stateless and firewall
 * in SConnInfo structure
 *
 * Revision 6.5  2000/12/29 17:47:46  lavr
 * NCBID stuff removed; ClientMode enum added;
 * ConnNetInfo_SetUserHeader added; http_user_header is now
 * included in ConnInfo structure. ConnNetInfo_Destroy parameter
 * changed to be a pointer (was a double pointer).
 *
 * Revision 6.4  2000/11/07 23:23:15  vakatov
 * In-sync with the C Toolkit "connutil.c:R6.15", "connutil.h:R6.13"
 * (with "eMIME_Dispatch" added).
 *
 * Revision 6.3  2000/10/05 22:39:21  lavr
 * SConnNetInfo modified to contain 'client_mode' instead of just 'firewall'
 *
 * Revision 6.2  2000/09/26 22:01:30  lavr
 * Registry entries changed, HTTP request method added
 *
 * Revision 6.1  2000/03/24 22:52:48  vakatov
 * Initial revision
 *
 * ==========================================================================
 */

#endif /* CONNECT___NCBI_CONNUTIL__H */