rtXmlPull.h

00001 /*
00002  * Copyright (c) 2003-2010 Objective Systems, Inc.
00003  *
00004  * This software is furnished under a license and may be used and copied
00005  * only in accordance with the terms of such license and with the
00006  * inclusion of the above copyright notice. This software or any other
00007  * copies thereof may not be provided or otherwise made available to any
00008  * other person. No title to and ownership of the software is hereby
00009  * transferred.
00010  *
00011  * The information in this software is subject to change without notice
00012  * and should not be construed as a commitment by Objective Systems, Inc.
00013  *
00014  * PROPRIETARY NOTICE
00015  *
00016  * This software is an unpublished work subject to a confidentiality agreement
00017  * and is protected by copyright and trade secret law.  Unauthorized copying,
00018  * redistribution or other use of this work is prohibited.
00019  *
00020  * The above notice of copyright on this source code product does not indicate
00021  * any actual or intended publication of such source code.
00022  *
00023  *****************************************************************************/
00024 #ifndef __RTXMLPULL_H
00025 #define __RTXMLPULL_H
00026 
00027 #include "rtxsrc/rtxCtype.h"
00028 #include "rtxsrc/rtxStream.h"
00029 #include "rtxsrc/rtxStreamBuffered.h"
00030 #include "rtxsrc/rtxErrCodes.h"
00031 #include "rtxsrc/rtxStreamDirectBuf.hh"
00032 #include "rtxmlsrc/osrtxml.h"
00033 
00034 /* _PULL_FROM_BUFFER - if defined, parser pull data from context buffer */
00035 /* #define _PULL_FROM_BUFFER */
00036 
00037 #define MAX_DELAYED_TASK_SIZE 10
00038 
00039 /* default buffer size for qName in ElementName */
00040 #define OSXML_DEFAULT_QNAME_BUF_SIZE 10
00041 
00042 #define OSXMLEVT_NONE 0
00043 #define OSXMLEVT_INITIAL OSXMLEVT_NONE
00044 #define OSXMLEVT_START_DOCUMENT 0x1
00045 #define OSXMLEVT_START_TAG 0x2
00046 #define OSXMLEVT_TEXT 0x4
00047 #define OSXMLEVT_END_TAG 0x8
00048 #define OSXMLEVT_END_DOCUMENT 0x10
00049 #define OSXMLEVT_COMMENT 0x20
00050 #define OSXMLEVT_PI 0x40
00051 #define OSXMLEVT_DTD 0x80
00052 
00053 #define OSXMLEVT_USED_FLAG 0x80000000u
00054 #define OSXMLEVT_ID_MASK (~OSXMLEVT_USED_FLAG)
00055 #define OSXMLEVT_ALL_MASK  OSXMLEVT_ID_MASK
00056 
00057 #define XMLPREADER(pctxt) ((OSXMLCtxtInfo*) pctxt->pXMLInfo)->pXmlPPReader
00058 
00059 typedef enum {
00060    OSXMLS_HEADER = -1,
00061    OSXMLS_COMMENT = -2,
00062    OSXMLS_START_ELEMENT = -3,
00063    OSXMLS_DTD = -4,
00064    OSXMLS_CONTENT = -5,
00065    OSXMLS_PI = -6,
00066    OSXMLS_END_ELEMENT = -7,
00067    OSXMLS_CDATA = -8,
00068 
00069    OSXMLS_LAST = -9
00070 } OSXMLStates;
00071 
00072 typedef enum {
00073    OSXMLSI_HEADER = -OSXMLS_HEADER-1,
00074    OSXMLSI_COMMENT = -OSXMLS_COMMENT-1,
00075    OSXMLSI_START_ELEMENT = -OSXMLS_START_ELEMENT-1,
00076    OSXMLSI_DTD = -OSXMLS_DTD-1,
00077    OSXMLSI_CONTENT = -OSXMLS_CONTENT-1,
00078    OSXMLSI_PI = -OSXMLS_PI-1,
00079    OSXMLSI_END_ELEMENT = -OSXMLS_END_ELEMENT-1,
00080    OSXMLSI_CDATA = -OSXMLS_CDATA-1,
00081 
00082    OSXMLSI_LAST = -OSXMLS_LAST-1
00083 } OSXMLStatesIndex;
00084 
00085 typedef enum {
00086    OSXMLDM_NONE = 0,
00087    OSXMLDM_SIMULATED,
00088    OSXMLDM_CONTENT
00089 } OSXMLDataMode;
00090 
00091 typedef enum {
00092    OSXMLNSI_UNQUALIFIED = 0,
00093    OSXMLNSI_UNKNOWN = -1,
00094    OSXMLNSI_UNCHECKED = -2,
00095    OSXMLNSI_XSI = -3,
00096    OSXMLNSI_XMLNS = -4,
00097    OSXMLNSI_XML = -5,
00098    OSXMLNSI_SOAP_ENVELOPE = -6,
00099    OSXMLNSI_XSD = -7
00100 } OSXMLNsIndex;
00101 
00102 typedef struct {
00103    size_t offset;
00104    size_t length;
00105 } OSXMLStrFragOffset;
00106 
00107 typedef struct {
00108    OSXMLStrFragment     mQName;
00109    OSXMLStrFragment     mLocalName;
00110    OSXMLStrFragment     mPrefix;
00111 
00112    OSUTF8CHAR           mBuffer[OSXML_DEFAULT_QNAME_BUF_SIZE];
00113 } OSXMLElementName;
00114 
00115 typedef struct OSXMLEvent {
00116    OSUINT32 mId;
00117    OSINT32 mLevel;
00118 } OSXMLEvent;
00119 
00120 typedef struct OSXMLDataCtxt {
00121    OSINT32 mDataLevel;
00122    OSXMLDataMode mDataMode;
00123 
00124    /* output */
00125    int mnChunk;
00126    OSBOOL mbLastChunk;
00127    OSBOOL mbCDATA;
00128    OSXMLStrFragment mData;
00129 
00130    /* for attr only */
00131    OSXMLStrFragment mSrcData;
00132    size_t           mSrcDataOffset;
00133    OSBOOL           mbInsTokenSeparator;
00134 } OSXMLDataCtxt;
00135 
00136 typedef struct {
00137    OSXMLStrFragOffset   mQName;
00138    OSXMLStrFragOffset   mLocalName;
00139    OSXMLStrFragOffset   mPrefix;
00140 } OSXMLElemNameOffset;
00141 
00142 typedef struct OSXMLStack {
00143    void*                mpStackArray; /* units specified in params */
00144    size_t               mSize;        /* allocated size in units */
00145    size_t               mUnitSize;    /* size of the unit, in octets */
00146    size_t               mCount;       /* number of used items    */
00147    OSBOOL               mbDynamic;    /* is dynamic memory allowed */
00148 } OSXMLStack;
00149 
00150 typedef struct OSXMLRewindPos {
00151    size_t      mPos;
00152    OSXMLEvent  mEvent;
00153    OSXMLStrFragment mData;
00154    OSBOOL      mbListMode;
00155    OSBOOL      mbLastChunk;
00156 } OSXMLRewindPos;
00157 
00158 typedef struct OSXMLSrcPos {
00159    OSUINT32    mLine;
00160    OSUINT32    mColumn;
00161    OSUINT32    mByteIdx;
00162    OSBOOL      mbCR;
00163 } OSXMLSrcPos;
00164 
00165 typedef struct OSXMLReader {
00166    /* input */
00167    OSCTXT*              mpCtxt;
00168    OSBOOL               mbNoTransform;
00169    OSBOOL               mbWSOnlyContext;
00170    OSBOOL               mbMixedContext;
00171    OSBOOL               mbListMode;
00172    OSBOOL               mbSkipPullListElem;
00173    OSBOOL               mbDecodeAsGroup;
00174    OSXMLWhiteSpaceMode  mWhiteSpaceMode;
00175    OSXMLWhiteSpaceMode  mStringWhiteSpaceMode;
00176    OSXMLEncoding        mEncoding;  /* XML document encoding (ex. UTF-8) */
00177    OSXMLBOM             mBOM;
00178 
00179    /* output */
00180    OSXMLEvent           mLastEvent;
00181 
00182    OSXMLElemNameOffset  mElementName;
00183    OSBOOL               mbHasAttributes;
00184    OSBOOL               mbEmptyElement;
00185 
00186    OSXMLStack           mAttributes; /* contains OSXMLAttrOffset */
00187 
00188    /*OSXMLStrFragment     mCharacters;*/
00189    OSXMLStrFragment     mData;
00190    OSBOOL               mbLastChunk;
00191    OSBOOL               mbCDATA;
00192    OSINT32              mLevel;
00193    OSINT32              mDecodeLevel;
00194 
00195    int                  mError;
00196 
00197    OSINT16              mElementNsIndex;
00198 
00199    /* internals */
00200    int mState, mPrevState, mNewState;
00201    int mLocalStates [OSXMLSI_LAST];
00202    OSUTF8CHAR  mCharBuf[10];
00203    OSUTF8CHAR* mpEscapeChar;
00204    OSUTF8CHAR* mpChars;
00205    size_t      mMarkedByteIndex;
00206    OSXMLDataMode mDataMode;
00207 
00208    /* contains OSXMLStrFragment for start tags to test against end tags */
00209    OSXMLStack  mTagNamesStack;
00210 
00211    /* contains namespaces (OSXMLNamespace) */
00212    OSXMLStack  mNamespacesStack;
00213 #ifdef OSXMLNS12
00214    const OSUTF8CHAR** mNamespaceTable;
00215    int         mNamespacesNumber;
00216    const OSUTF8CHAR** mPrevNamespaceTable;
00217    int         mPrevNamespacesNumber;
00218 #else
00219    const char** mNamespaceTable;
00220    int         mNamespacesNumber;
00221    const char** mPrevNamespaceTable;
00222    int         mPrevNamespacesNumber;
00223 #endif
00224 
00225    OSOCTET     mDelayedTasks [MAX_DELAYED_TASK_SIZE];
00226    size_t      mDelayedTaskCount;
00227    size_t      mDelayedTaskIndex;
00228 
00229    /* transport */
00230    OSUTF8CHAR* mpBuffer;
00231    size_t      mByteIndex;
00232    size_t      mBufSize;
00233    size_t      mReadSize;
00234    size_t      mMarkedPos;
00235 
00236    size_t      mLastBlockSize;
00237    size_t      mLastByteIndex;
00238 
00239    OSBOOL      mbBackoffEnabled;
00240    OSBOOL      mbSysMemBuf;
00241 
00242    OSXMLStack  mRewindPosStack;
00243 
00244    /* position in source XML */
00245    OSXMLSrcPos mSrcPos;
00246    size_t      mMarkedAttrCount;
00247 
00248 } OSXMLReader;
00249 
00250 #ifdef __cplusplus
00251 extern "C" {
00252 #endif
00253 
00257 EXTXMLMETHOD struct OSXMLReader* rtXmlRdCreateXmlReader (OSCTXT* pctxt);
00258 
00268 EXTXMLMETHOD OSXMLWhiteSpaceMode rtXmlRdSetWhiteSpaceMode
00269    (struct OSXMLReader* pReader, OSXMLWhiteSpaceMode whiteSpaceMode);
00270 
00288 EXTXMLMETHOD long rtXmlRdFirstData (struct OSXMLReader* pReader, OSXMLDataCtxt* pDataCtxt);
00289 
00303 EXTXMLMETHOD long rtXmlRdNextData (struct OSXMLReader* pReader, OSXMLDataCtxt* pDataCtxt);
00304 
00311 EXTXMLMETHOD int rtXmlRdSkipCurrentLevel (struct OSXMLReader* pReader);
00312 
00321 EXTXMLMETHOD int rtXmlRdNext (struct OSXMLReader* pReader);
00322 
00329 EXTXMLMETHOD int rtXmlRdNextTag (struct OSXMLReader* pReader);
00330 
00336 EXTXMLMETHOD int rtXmlRdNextEndTag (struct OSXMLReader* pReader);
00337 
00343 EXTXMLMETHOD OSXMLEvent rtXmlRdGetLastEvent (struct OSXMLReader* pReader);
00344 
00355 EXTXMLMETHOD int rtXmlRdGetTagName (struct OSXMLReader* pReader,
00356                        OSXMLStrFragment* localName,
00357                        OSINT16* namespaceIndex);
00358 
00367 EXTXMLMETHOD int rtXmlRdGetAttributeCount (struct OSXMLReader* pReader);
00368 
00379 EXTXMLMETHOD int rtXmlRdGetAttributeName (struct OSXMLReader* pReader,
00380                              OSXMLNameFragments* pAttr,
00381                              OSINT16* pNsidx,
00382                              size_t index);
00383 
00393 EXTXMLMETHOD int rtXmlRdFirstAttr (struct OSXMLReader* pReader, 
00394                              OSXMLStrFragment* pAttrName,
00395                              OSXMLStrFragment* pAttrValue);
00396  
00406 EXTXMLMETHOD int rtXmlRdNextAttr (struct OSXMLReader* pReader, 
00407                              OSXMLStrFragment* pAttrName,
00408                              OSXMLStrFragment* pAttrValue);
00409  
00419 EXTXMLMETHOD int rtXmlRdSelectAttribute (struct OSXMLReader* pReader, size_t index);
00420 
00431 EXTXMLMETHOD OSINT32 rtXmlRdGetCurrentLevel (struct OSXMLReader* pReader);
00432 
00447 EXTXMLMETHOD int rtXmlRdNextEvent (struct OSXMLReader* pReader,
00448                       OSUINT32 eventMask,
00449                       OSINT32 maxLevel,
00450                       OSXMLEvent* pLastEvent);
00451 
00457 EXTXMLMETHOD OSXMLDataMode rtXmlRdGetDataMode (struct OSXMLReader* pReader);
00458 
00465 EXTXMLMETHOD int rtXmlRdMarkLastEventDone (struct OSXMLReader* pReader);
00466 
00473 EXTXMLMETHOD int rtXmlRdMarkLastEventActive (struct OSXMLReader* pReader);
00474 
00478 EXTXMLMETHOD void rtXmlRdResetXmlReader (struct OSXMLReader* pReader);
00479 
00480 EXTXMLMETHOD void rtXmlRdMarkPos (struct OSXMLReader* pReader);
00481 EXTXMLMETHOD void rtXmlRdRewindToMarkedPos (struct OSXMLReader* pReader);
00482 EXTXMLMETHOD void rtXmlRdResetMarkedPos (struct OSXMLReader* pReader);
00483 
00484 EXTXMLMETHOD int rtXmlRdGetXSITypeAttr (struct OSXMLReader* pReader,
00485                            const OSUTF8CHAR** ppAttrValue,
00486                            OSINT16* nsidx, size_t* pLocalOffs);
00487 
00491 EXTXMLMETHOD OSBOOL rtXmlRdIsEmpty (struct OSXMLReader* pReader);
00492 
00493 EXTXMLMETHOD void rtXmlRdGetSourcePosition (struct OSXMLReader* pReader,
00494                                OSUINT32* pLine,
00495                                OSUINT32* pColumn,
00496                                OSUINT32* pByteIndex,
00497                                OSBOOL nextPos);
00498 
00499 EXTXMLMETHOD void rtXmlRdErrAddSrcPos (struct OSXMLReader* pReader,
00500                           OSBOOL nextPos);
00501 
00502 EXTXMLMETHOD void rtXmlRdErrAddDataSrcPos (struct OSXMLReader* pReader,
00503                               OSXMLDataCtxt* pDataCtxt,
00504                               OSUINT32 offset);
00505 
00506 #ifdef __cplusplus
00507 }
00508 #endif
00509 
00510 #endif /* __RTXMLPULL_H */
00511