xmlregexp.c

xml开源解析代码.版本为libxml2-2.6.29,可支持GB3212.网络消息发送XML时很有用.

            fprintf(output, "NUMBER_DECIMAL "); break;
        case XML_REGEXP_NUMBER_LETTER:
            fprintf(output, "NUMBER_LETTER "); break;
        case XML_REGEXP_NUMBER_OTHERS:
            fprintf(output, "NUMBER_OTHERS "); break;
        case XML_REGEXP_PUNCT:
            fprintf(output, "PUNCT "); break;
        case XML_REGEXP_PUNCT_CONNECTOR:
            fprintf(output, "PUNCT_CONNECTOR "); break;
        case XML_REGEXP_PUNCT_DASH:
            fprintf(output, "PUNCT_DASH "); break;
        case XML_REGEXP_PUNCT_OPEN:
            fprintf(output, "PUNCT_OPEN "); break;
        case XML_REGEXP_PUNCT_CLOSE:
            fprintf(output, "PUNCT_CLOSE "); break;
        case XML_REGEXP_PUNCT_INITQUOTE:
            fprintf(output, "PUNCT_INITQUOTE "); break;
        case XML_REGEXP_PUNCT_FINQUOTE:
            fprintf(output, "PUNCT_FINQUOTE "); break;
        case XML_REGEXP_PUNCT_OTHERS:
            fprintf(output, "PUNCT_OTHERS "); break;
        case XML_REGEXP_SEPAR:
            fprintf(output, "SEPAR "); break;
        case XML_REGEXP_SEPAR_SPACE:
            fprintf(output, "SEPAR_SPACE "); break;
        case XML_REGEXP_SEPAR_LINE:
            fprintf(output, "SEPAR_LINE "); break;
        case XML_REGEXP_SEPAR_PARA:
            fprintf(output, "SEPAR_PARA "); break;
        case XML_REGEXP_SYMBOL:
            fprintf(output, "SYMBOL "); break;
        case XML_REGEXP_SYMBOL_MATH:
            fprintf(output, "SYMBOL_MATH "); break;
        case XML_REGEXP_SYMBOL_CURRENCY:
            fprintf(output, "SYMBOL_CURRENCY "); break;
        case XML_REGEXP_SYMBOL_MODIFIER:
            fprintf(output, "SYMBOL_MODIFIER "); break;
        case XML_REGEXP_SYMBOL_OTHERS:
            fprintf(output, "SYMBOL_OTHERS "); break;
        case XML_REGEXP_OTHER:
            fprintf(output, "OTHER "); break;
        case XML_REGEXP_OTHER_CONTROL:
            fprintf(output, "OTHER_CONTROL "); break;
        case XML_REGEXP_OTHER_FORMAT:
            fprintf(output, "OTHER_FORMAT "); break;
        case XML_REGEXP_OTHER_PRIVATE:
            fprintf(output, "OTHER_PRIVATE "); break;
        case XML_REGEXP_OTHER_NA:
            fprintf(output, "OTHER_NA "); break;
        case XML_REGEXP_BLOCK_NAME:
	    fprintf(output, "BLOCK "); break;
    }
}

static void
xmlRegPrintQuantType(FILE *output, xmlRegQuantType type) {
    switch (type) {
        case XML_REGEXP_QUANT_EPSILON:
	    fprintf(output, "epsilon "); break;
        case XML_REGEXP_QUANT_ONCE:
	    fprintf(output, "once "); break;
        case XML_REGEXP_QUANT_OPT:
	    fprintf(output, "? "); break;
        case XML_REGEXP_QUANT_MULT:
	    fprintf(output, "* "); break;
        case XML_REGEXP_QUANT_PLUS:
	    fprintf(output, "+ "); break;
	case XML_REGEXP_QUANT_RANGE:
	    fprintf(output, "range "); break;
	case XML_REGEXP_QUANT_ONCEONLY:
	    fprintf(output, "onceonly "); break;
	case XML_REGEXP_QUANT_ALL:
	    fprintf(output, "all "); break;
    }
}
static void
xmlRegPrintRange(FILE *output, xmlRegRangePtr range) {
    fprintf(output, "  range: ");
    if (range->neg)
	fprintf(output, "negative ");
    xmlRegPrintAtomType(output, range->type);
    fprintf(output, "%c - %c\n", range->start, range->end);
}

static void
xmlRegPrintAtom(FILE *output, xmlRegAtomPtr atom) {
    fprintf(output, " atom: ");
    if (atom == NULL) {
	fprintf(output, "NULL\n");
	return;
    }
    if (atom->neg)
        fprintf(output, "not ");
    xmlRegPrintAtomType(output, atom->type);
    xmlRegPrintQuantType(output, atom->quant);
    if (atom->quant == XML_REGEXP_QUANT_RANGE)
	fprintf(output, "%d-%d ", atom->min, atom->max);
    if (atom->type == XML_REGEXP_STRING)
	fprintf(output, "'%s' ", (char *) atom->valuep);
    if (atom->type == XML_REGEXP_CHARVAL)
	fprintf(output, "char %c\n", atom->codepoint);
    else if (atom->type == XML_REGEXP_RANGES) {
	int i;
	fprintf(output, "%d entries\n", atom->nbRanges);
	for (i = 0; i < atom->nbRanges;i++)
	    xmlRegPrintRange(output, atom->ranges[i]);
    } else if (atom->type == XML_REGEXP_SUBREG) {
	fprintf(output, "start %d end %d\n", atom->start->no, atom->stop->no);
    } else {
	fprintf(output, "\n");
    }
}

static void
xmlRegPrintTrans(FILE *output, xmlRegTransPtr trans) {
    fprintf(output, "  trans: ");
    if (trans == NULL) {
	fprintf(output, "NULL\n");
	return;
    }
    if (trans->to < 0) {
	fprintf(output, "removed\n");
	return;
    }
    if (trans->nd != 0) {
	if (trans->nd == 2)
	    fprintf(output, "last not determinist, ");
	else
	    fprintf(output, "not determinist, ");
    }
    if (trans->counter >= 0) {
	fprintf(output, "counted %d, ", trans->counter);
    }
    if (trans->count == REGEXP_ALL_COUNTER) {
	fprintf(output, "all transition, ");
    } else if (trans->count >= 0) {
	fprintf(output, "count based %d, ", trans->count);
    }
    if (trans->atom == NULL) {
	fprintf(output, "epsilon to %d\n", trans->to);
	return;
    }
    if (trans->atom->type == XML_REGEXP_CHARVAL)
	fprintf(output, "char %c ", trans->atom->codepoint);
    fprintf(output, "atom %d, to %d\n", trans->atom->no, trans->to);
}
    
static void
xmlRegPrintState(FILE *output, xmlRegStatePtr state) {
    int i;

    fprintf(output, " state: ");
    if (state == NULL) {
	fprintf(output, "NULL\n");
	return;
    }
    if (state->type == XML_REGEXP_START_STATE)
	fprintf(output, "START ");
    if (state->type == XML_REGEXP_FINAL_STATE)
	fprintf(output, "FINAL ");
    
    fprintf(output, "%d, %d transitions:\n", state->no, state->nbTrans);
    for (i = 0;i < state->nbTrans; i++) {
	xmlRegPrintTrans(output, &(state->trans[i]));
    }
}

#ifdef DEBUG_REGEXP_GRAPH
static void
xmlRegPrintCtxt(FILE *output, xmlRegParserCtxtPtr ctxt) {
    int i;

    fprintf(output, " ctxt: ");
    if (ctxt == NULL) {
	fprintf(output, "NULL\n");
	return;
    }
    fprintf(output, "'%s' ", ctxt->string);
    if (ctxt->error)
	fprintf(output, "error ");
    if (ctxt->neg)
	fprintf(output, "neg ");
    fprintf(output, "\n");
    fprintf(output, "%d atoms:\n", ctxt->nbAtoms);
    for (i = 0;i < ctxt->nbAtoms; i++) {
	fprintf(output, " %02d ", i);
	xmlRegPrintAtom(output, ctxt->atoms[i]);
    }
    if (ctxt->atom != NULL) {
	fprintf(output, "current atom:\n");
	xmlRegPrintAtom(output, ctxt->atom);
    }
    fprintf(output, "%d states:", ctxt->nbStates);
    if (ctxt->start != NULL)
	fprintf(output, " start: %d", ctxt->start->no);
    if (ctxt->end != NULL)
	fprintf(output, " end: %d", ctxt->end->no);
    fprintf(output, "\n");
    for (i = 0;i < ctxt->nbStates; i++) {
	xmlRegPrintState(output, ctxt->states[i]);
    }
    fprintf(output, "%d counters:\n", ctxt->nbCounters);
    for (i = 0;i < ctxt->nbCounters; i++) {
	fprintf(output, " %d: min %d max %d\n", i, ctxt->counters[i].min,
		                                ctxt->counters[i].max);
    }
}
#endif

/************************************************************************
 * 									*
 *		 Finite Automata structures manipulations		*
 * 									*
 ************************************************************************/

static void 
xmlRegAtomAddRange(xmlRegParserCtxtPtr ctxt, xmlRegAtomPtr atom,
	           int neg, xmlRegAtomType type, int start, int end,
		   xmlChar *blockName) {
    xmlRegRangePtr range;

    if (atom == NULL) {
	ERROR("add range: atom is NULL");
	return;
    }
    if (atom->type != XML_REGEXP_RANGES) {
	ERROR("add range: atom is not ranges");
	return;
    }
    if (atom->maxRanges == 0) {
	atom->maxRanges = 4;
	atom->ranges = (xmlRegRangePtr *) xmlMalloc(atom->maxRanges *
		                             sizeof(xmlRegRangePtr));
	if (atom->ranges == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding ranges");
	    atom->maxRanges = 0;
	    return;
	}
    } else if (atom->nbRanges >= atom->maxRanges) {
	xmlRegRangePtr *tmp;
	atom->maxRanges *= 2;
	tmp = (xmlRegRangePtr *) xmlRealloc(atom->ranges, atom->maxRanges *
		                             sizeof(xmlRegRangePtr));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding ranges");
	    atom->maxRanges /= 2;
	    return;
	}
	atom->ranges = tmp;
    }
    range = xmlRegNewRange(ctxt, neg, type, start, end);
    if (range == NULL)
	return;
    range->blockName = blockName;
    atom->ranges[atom->nbRanges++] = range;
    
}

static int
xmlRegGetCounter(xmlRegParserCtxtPtr ctxt) {
    if (ctxt->maxCounters == 0) {
	ctxt->maxCounters = 4;
	ctxt->counters = (xmlRegCounter *) xmlMalloc(ctxt->maxCounters *
		                             sizeof(xmlRegCounter));
	if (ctxt->counters == NULL) {
	    xmlRegexpErrMemory(ctxt, "allocating counter");
	    ctxt->maxCounters = 0;
	    return(-1);
	}
    } else if (ctxt->nbCounters >= ctxt->maxCounters) {
	xmlRegCounter *tmp;
	ctxt->maxCounters *= 2;
	tmp = (xmlRegCounter *) xmlRealloc(ctxt->counters, ctxt->maxCounters *
		                           sizeof(xmlRegCounter));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "allocating counter");
	    ctxt->maxCounters /= 2;
	    return(-1);
	}
	ctxt->counters = tmp;
    }
    ctxt->counters[ctxt->nbCounters].min = -1;
    ctxt->counters[ctxt->nbCounters].max = -1;
    return(ctxt->nbCounters++);
}

static int 
xmlRegAtomPush(xmlRegParserCtxtPtr ctxt, xmlRegAtomPtr atom) {
    if (atom == NULL) {
	ERROR("atom push: atom is NULL");
	return(-1);
    }
    if (ctxt->maxAtoms == 0) {
	ctxt->maxAtoms = 4;
	ctxt->atoms = (xmlRegAtomPtr *) xmlMalloc(ctxt->maxAtoms *
		                             sizeof(xmlRegAtomPtr));
	if (ctxt->atoms == NULL) {
	    xmlRegexpErrMemory(ctxt, "pushing atom");
	    ctxt->maxAtoms = 0;
	    return(-1);
	}
    } else if (ctxt->nbAtoms >= ctxt->maxAtoms) {
	xmlRegAtomPtr *tmp;
	ctxt->maxAtoms *= 2;
	tmp = (xmlRegAtomPtr *) xmlRealloc(ctxt->atoms, ctxt->maxAtoms *
		                             sizeof(xmlRegAtomPtr));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "allocating counter");
	    ctxt->maxAtoms /= 2;
	    return(-1);
	}
	ctxt->atoms = tmp;
    }
    atom->no = ctxt->nbAtoms;
    ctxt->atoms[ctxt->nbAtoms++] = atom;
    return(0);
}

static void 
xmlRegStateAddTransTo(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr target,
                      int from) {
    if (target->maxTransTo == 0) {
	target->maxTransTo = 8;
	target->transTo = (int *) xmlMalloc(target->maxTransTo *
		                             sizeof(int));
	if (target->transTo == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding transition");
	    target->maxTransTo = 0;
	    return;
	}
    } else if (target->nbTransTo >= target->maxTransTo) {
	int *tmp;
	target->maxTransTo *= 2;
	tmp = (int *) xmlRealloc(target->transTo, target->maxTransTo *
		                             sizeof(int));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding transition");
	    target->maxTransTo /= 2;
	    return;
	}
	target->transTo = tmp;
    }
    target->transTo[target->nbTransTo] = from;
    target->nbTransTo++;
}

static void 
xmlRegStateAddTrans(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr state,
	            xmlRegAtomPtr atom, xmlRegStatePtr target,
		    int counter, int count) {

    int nrtrans;

    if (state == NULL) {
	ERROR("add state: state is NULL");
	return;
    }
    if (target == NULL) {
	ERROR("add state: target is NULL");
	return;
    }
    /*
     * Other routines follow the philosophy 'When in doubt, add a transition'
     * so we check here whether such a transition is already present and, if
     * so, silently ignore this request.
     */

    for (nrtrans = state->nbTrans - 1; nrtrans >= 0; nrtrans--) {
	xmlRegTransPtr trans = &(state->trans[nrtrans]);
	if ((trans->atom == atom) &&
	    (trans->to == target->no) &&
	    (trans->counter == counter) &&
	    (trans->count == count)) {
#ifdef DEBUG_REGEXP_GRAPH
	    printf("Ignoring duplicate transition from %d to %d\n",
		    state->no, target->no);
#endif
	    return;
	}
    }

    if (state->maxTrans == 0) {
	state->maxTrans = 8;
	state->trans = (xmlRegTrans *) xmlMalloc(state->maxTrans *
		                             sizeof(xmlRegTrans));
	if (state->trans == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding transition");
	    state->maxTrans = 0;
	    return;
	}
    } else if (state->nbTrans >= state->maxTrans) {
	xmlRegTrans *tmp;
	state->maxTrans *= 2;
	tmp = (xmlRegTrans *) xmlRealloc(state->trans, state->maxTrans *
		                             sizeof(xmlRegTrans));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding transition");
	    state->maxTrans /= 2;
	    return;
	}
	state->trans = tmp;
    }
#ifdef DEBUG_REGEXP_GRAPH
    printf("Add trans from %d to %d ", state->no, target->no);
    if (count == REGEXP_ALL_COUNTER)
	printf("all transition\n");
    else if (count >= 0)
	printf("count based %d\n", count);
    else if (counter >= 0)
	printf("counted %d\n", counter);
    else if (atom == NULL)
	printf("epsilon transition\n");
    else if (atom != NULL) 
        xmlRegPrintAtom(stdout, atom);
#endif

    state->trans[state->nbTrans].atom = atom;
    state->trans[state->nbTrans].to = target->no;
    state->trans[state->nbTrans].counter = counter;
    state->trans[state->nbTrans].count = count;
    state->trans[state->nbTrans].nd = 0;
    state->nbTrans++;
    xmlRegStateAddTransTo(ctxt, target, state->no);
}

static int
xmlRegStatePush(xmlRegParserCtxtPtr ctxt, xmlRegStatePtr state) {
    if (state == NULL) return(-1);
    if (ctxt->maxStates == 0) {
	ctxt->maxStates = 4;
	ctxt->states = (xmlRegStatePtr *) xmlMalloc(ctxt->maxStates *
		                             sizeof(xmlRegStatePtr));
	if (ctxt->states == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding state");
	    ctxt->maxStates = 0;
	    return(-1);
	}
    } else if (ctxt->nbStates >= ctxt->maxStates) {
	xmlRegStatePtr *tmp;
	ctxt->maxStates *= 2;
	tmp = (xmlRegStatePtr *) xmlRealloc(ctxt->states, ctxt->maxStates *
		                             sizeof(xmlRegStatePtr));
	if (tmp == NULL) {
	    xmlRegexpErrMemory(ctxt, "adding state");
	    ctxt->maxStates /= 2;
	    return(-1);
	}
	ctxt->states = tmp;
    }
    state->no = ctxt->nbStates;
    ctxt->states[ctxt->nbStates++] = state;
    return(0);
}

/**
 * xmlFAGenerateAllTransition: