rtXmlPull.h

00001 /*
00002  * Copyright (c) 2003-2017 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00016  * This software is an unpublished work subject to a confidentiality agreement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 struct {
00092    size_t offset;
00093    size_t length;
00094 } OSXMLStrFragOffset;
00095 
00096 typedef struct {
00097    OSXMLStrFragment     mQName;
00098    OSXMLStrFragment     mLocalName;
00099    OSXMLStrFragment     mPrefix;
00100 
00101    OSUTF8CHAR           mBuffer[OSXML_DEFAULT_QNAME_BUF_SIZE];
00102 } OSXMLElementName;
00103 
00104 typedef struct OSXMLEvent {
00105    OSUINT32 mId;
00106    OSINT32 mLevel;
00107 } OSXMLEvent;
00108 
00109 typedef struct OSXMLDataCtxt {
00110    OSINT32 mDataLevel;
00111    OSXMLDataMode mDataMode;
00112 
00113    /* output */
00114    int mnChunk;
00115    OSBOOL mbLastChunk;
00116    OSBOOL mbCDATA;
00117    OSXMLStrFragment mData;
00118 
00119    /* for attr only */
00120    OSXMLStrFragment mSrcData;
00121    size_t           mSrcDataOffset;
00122    OSBOOL           mbInsTokenSeparator;
00123 } OSXMLDataCtxt;
00124 
00125 typedef struct {
00126    OSXMLStrFragOffset   mQName;
00127    OSXMLStrFragOffset   mLocalName;
00128    OSXMLStrFragOffset   mPrefix;
00129 } OSXMLElemNameOffset;
00130 
00131 typedef struct OSXMLStack {
00132    void*                mpStackArray; /* units specified in params */
00133    size_t               mSize;        /* allocated size in units */
00134    size_t               mUnitSize;    /* size of the unit, in octets */
00135    size_t               mCount;       /* number of used items    */
00136    OSBOOL               mbDynamic;    /* is dynamic memory allowed */
00137 } OSXMLStack;
00138 
00139 typedef struct OSXMLRewindPos {
00140    size_t      mPos;
00141    OSXMLEvent  mEvent;
00142    OSXMLStrFragment mData;
00143    OSBOOL      mbListMode;
00144    OSBOOL      mbLastChunk;
00145 } OSXMLRewindPos;
00146 
00147 typedef struct OSXMLSrcPos {
00148    OSUINT32    mLine;
00149    OSUINT32    mColumn;
00150    OSUINT32    mByteIdx;
00151    OSBOOL      mbCR;
00152 } OSXMLSrcPos;
00153 
00154 typedef struct OSXMLReader {
00155    /* input */
00156    OSCTXT*              mpCtxt;
00157    OSBOOL               mbNoTransform;
00158    /* OSBOOL               mbWSOnlyContext; */
00159    OSBOOL               mbMixedContext;
00160    OSBOOL               mbListMode;
00161    OSBOOL               mbSkipPullListElem;
00162    OSBOOL               mbDecodeAsGroup;
00163    OSXMLWhiteSpaceMode  mWhiteSpaceMode;
00164    OSXMLWhiteSpaceMode  mStringWhiteSpaceMode;
00165    OSXMLEncoding        mEncoding;  /* XML document encoding (ex. UTF-8) */
00166    OSXMLBOM             mBOM;
00167 
00168    /* output */
00169    OSXMLEvent           mLastEvent;
00170 
00171    OSXMLElemNameOffset  mElementName;
00172    OSBOOL               mbHasAttributes;
00173    OSBOOL               mbEmptyElement;
00174 
00175    OSXMLStack           mAttributes; /* contains OSXMLAttrOffset */
00176 
00177    /*OSXMLStrFragment     mCharacters;*/
00178    OSXMLStrFragment     mData;
00179    OSBOOL               mbLastChunk;
00180    OSBOOL               mbCDATA;
00181    OSINT32              mLevel;
00182    OSINT32              mDecodeLevel;
00183 
00184    int                  mError;
00185 
00186    OSINT16              mElementNsIndex;
00187 
00188    /* internals */
00189    int mState, mPrevState, mNewState;
00190    int mLocalStates [OSXMLSI_LAST];
00191    OSUTF8CHAR  mCharBuf[10];
00192    OSUTF8CHAR* mpEscapeChar;
00193    OSUTF8CHAR* mpChars;
00194    size_t      mMarkedByteIndex;
00195    OSXMLDataMode mDataMode;
00196 
00197    /* contains OSXMLStrFragment for start tags to test against end tags */
00198    OSXMLStack  mTagNamesStack;
00199 
00200    /* contains namespaces (OSXMLNamespace) */
00201    OSXMLStack  mNamespacesStack;
00202 #ifdef OSXMLNS12
00203    const OSUTF8CHAR** mNamespaceTable;
00204    int         mNamespacesNumber;
00205    const OSUTF8CHAR** mPrevNamespaceTable;
00206    int         mPrevNamespacesNumber;
00207 #else
00208    const char** mNamespaceTable;
00209    int         mNamespacesNumber;
00210    const char** mPrevNamespaceTable;
00211    int         mPrevNamespacesNumber;
00212 #endif
00213 
00214    OSOCTET     mDelayedTasks [MAX_DELAYED_TASK_SIZE];
00215    size_t      mDelayedTaskCount;
00216    size_t      mDelayedTaskIndex;
00217 
00218    /* transport */
00219    OSUTF8CHAR* mpBuffer;
00220    size_t      mByteIndex;
00221    size_t      mBufSize;
00222    size_t      mReadSize;
00223    size_t      mMarkedPos;
00224 
00225    size_t      mLastBlockSize;
00226    size_t      mLastByteIndex;
00227 
00228    OSBOOL      mbBackoffEnabled;
00229    OSBOOL      mbSysMemBuf;
00230 
00231    OSXMLStack  mRewindPosStack;
00232 
00233    /* position in source XML */
00234    OSXMLSrcPos mSrcPos;
00235    size_t      mMarkedAttrCount;
00236 
00237 } OSXMLReader;
00238 
00239 #ifdef __cplusplus
00240 extern "C" {
00241 #endif
00242 
00246 EXTXMLMETHOD struct OSXMLReader* rtXmlRdCreateXmlReader (OSCTXT* pctxt);
00247 
00257 EXTXMLMETHOD OSXMLWhiteSpaceMode rtXmlRdSetWhiteSpaceMode
00258    (struct OSXMLReader* pReader, OSXMLWhiteSpaceMode whiteSpaceMode);
00259 
00277 EXTXMLMETHOD long rtXmlRdFirstData (struct OSXMLReader* pReader, OSXMLDataCtxt* pDataCtxt);
00278 
00292 EXTXMLMETHOD long rtXmlRdNextData (struct OSXMLReader* pReader, OSXMLDataCtxt* pDataCtxt);
00293 
00300 EXTXMLMETHOD int rtXmlRdSkipCurrentLevel (struct OSXMLReader* pReader);
00301 
00310 EXTXMLMETHOD int rtXmlRdNext (struct OSXMLReader* pReader);
00311 
00318 EXTXMLMETHOD int rtXmlRdNextTag (struct OSXMLReader* pReader);
00319 
00325 EXTXMLMETHOD int rtXmlRdNextEndTag (struct OSXMLReader* pReader);
00326 
00332 EXTXMLMETHOD OSXMLEvent rtXmlRdGetLastEvent (struct OSXMLReader* pReader);
00333 
00344 EXTXMLMETHOD int rtXmlRdGetTagName (struct OSXMLReader* pReader,
00345                        OSXMLStrFragment* localName,
00346                        OSINT16* namespaceIndex);
00347 
00356 EXTXMLMETHOD int rtXmlRdGetAttributeCount (struct OSXMLReader* pReader);
00357 
00368 EXTXMLMETHOD int rtXmlRdGetAttributeName (struct OSXMLReader* pReader,
00369                              OSXMLNameFragments* pAttr,
00370                              OSINT16* pNsidx,
00371                              size_t index);
00372 
00382 EXTXMLMETHOD int rtXmlRdFirstAttr (struct OSXMLReader* pReader,
00383                              OSXMLStrFragment* pAttrName,
00384                              OSXMLStrFragment* pAttrValue);
00385 
00395 EXTXMLMETHOD int rtXmlRdNextAttr (struct OSXMLReader* pReader,
00396                              OSXMLStrFragment* pAttrName,
00397                              OSXMLStrFragment* pAttrValue);
00398 
00408 EXTXMLMETHOD int rtXmlRdSelectAttribute (struct OSXMLReader* pReader, size_t index);
00409 
00420 EXTXMLMETHOD OSINT32 rtXmlRdGetCurrentLevel (struct OSXMLReader* pReader);
00421 
00436 EXTXMLMETHOD int rtXmlRdNextEvent (struct OSXMLReader* pReader,
00437                       OSUINT32 eventMask,
00438                       OSINT32 maxLevel,
00439                       OSXMLEvent* pLastEvent);
00440 
00446 EXTXMLMETHOD OSXMLDataMode rtXmlRdGetDataMode (struct OSXMLReader* pReader);
00447 
00454 EXTXMLMETHOD int rtXmlRdMarkLastEventDone (struct OSXMLReader* pReader);
00455 
00462 EXTXMLMETHOD int rtXmlRdMarkLastEventActive (struct OSXMLReader* pReader);
00463 
00467 EXTXMLMETHOD void rtXmlRdResetXmlReader (struct OSXMLReader* pReader);
00468 
00469 EXTXMLMETHOD void rtXmlRdMarkPos (struct OSXMLReader* pReader);
00470 EXTXMLMETHOD void rtXmlRdRewindToMarkedPos (struct OSXMLReader* pReader);
00471 EXTXMLMETHOD void rtXmlRdResetMarkedPos (struct OSXMLReader* pReader);
00472 
00473 EXTXMLMETHOD int rtXmlRdGetXSITypeAttr (struct OSXMLReader* pReader,
00474                            const OSUTF8CHAR** ppAttrValue,
00475                            OSINT16* nsidx, size_t* pLocalOffs);
00476 
00480 EXTXMLMETHOD OSBOOL rtXmlRdIsEmpty (struct OSXMLReader* pReader);
00481 
00482 EXTXMLMETHOD void rtXmlRdGetSourcePosition (struct OSXMLReader* pReader,
00483                                OSUINT32* pLine,
00484                                OSUINT32* pColumn,
00485                                OSUINT32* pByteIndex,
00486                                OSBOOL nextPos);
00487 
00488 EXTXMLMETHOD void rtXmlRdErrAddSrcPos (struct OSXMLReader* pReader,
00489                           OSBOOL nextPos);
00490 
00491 EXTXMLMETHOD void rtXmlRdErrAddDataSrcPos (struct OSXMLReader* pReader,
00492                               OSXMLDataCtxt* pDataCtxt,
00493                               OSSIZE offset);
00494 
00495 #ifdef __cplusplus
00496 }
00497 #endif
00498 
00499 #endif /* __RTXMLPULL_H */
00500