agrep.c

Mehldau和Myer的Agrep3版本

						r1 = Init1 & r2;
						r0 = B[3] | r3;
						B[4] = ((r3 | Next[r0>>hh] | Next1[r0&LL])&r_NO_ERR) | r1 ;  
						if(D == 4) goto Nextchar1mem;
					}
				} /* if(NOT Newline) */
				else {  
					j++;
					if (DELIMITER) CurrentByteOffset -= D_length;
					else CurrentByteOffset -= 1;
					r1 = Init1 & A[D];            /* match against endofline */
					B[D] = ((Next[A[D]>>hh] | Next1[A[D]&LL]) & CMask) | r1;
					if(TAIL) B[D] = ( Next[B[D]>>hh] | Next1[B[D]&LL] ) | B[D]; 
					/* epsilon move */
					if(( B[D] & 1 ) ^ INVERSE) {
						if(FILENAMEONLY && (NEW_FILE || !POST_FILTER)) {
							num_of_matched++;

							if (agrep_finalfp != NULL) 
								fprintf(agrep_finalfp, "%s\n", CurrentFileName);
							else {
								int outindex;
								for(outindex=0; (outindex+agrep_outpointer<agrep_outlen) && 
										(CurrentFileName[outindex] != '\0'); outindex++) {
									agrep_outbuffer[agrep_outpointer+outindex] = CurrentFileName[outindex];
								}
								if ((CurrentFileName[outindex] != '\0') || (outindex+agrep_outpointer+1>=agrep_outlen)) {
									OUTPUT_OVERFLOW;
									free_buf(Text, buffer);
									return -1;
								}
								else agrep_outbuffer[agrep_outpointer+outindex++] = '\n';
								agrep_outpointer += outindex;
							}

							free_buf(Text, buffer);
							NEW_FILE = OFF;
							return 0;
						} 
						if (-1 == r_output(buffer, i, end, j)) {free_buf(Text, buffer); return -1;}
						if (((LIMITOUTPUT > 0) && (LIMITOUTPUT <= num_of_matched)) ||
						    ((LIMITPERFILE > 0) && (LIMITPERFILE <= num_of_matched - prev_num_of_matched))) {
							free_buf(Text, buffer);
							return 0;	/* done */
						}
					}
					for(k=0; k<=D; k++) A[k] = Init0; 
					r1 = Init1 & A[0];
					B[0] = ((Next[A[0]>>hh] | Next1[A[0]&LL]) & CMask) | r1;
					for(k=1; k<=D; k++) {
						r3 = A[k];
						r1 = Init1 & r3;
						r2 = A[k-1] | B[k-1];
						B[k] = ((Next[r3>>hh] | Next1[r3&LL]) & CMask) | ((A[k-1] | Next[r2>>hh] | Next1[r2&LL]) & r_NO_ERR) | r1;
					}
					if (DELIMITER) CurrentByteOffset += 1*D_length;
					else CurrentByteOffset += 1*1;
				}
	Nextchar1mem: 
				i=i+1;
				CurrentByteOffset ++;
			} /* while i < end */

		return 0;
	}
#endif	/*AGREP_POINTER*/
} /* re1 */


int re(Text, M, D)
int Text, M, D;
{
	register unsigned i, c, r1, r2, r3, CMask, k, Newline, Init0, Init1, end; 
	register unsigned r_even, r_odd, r_NO_ERR ;
	unsigned RMask[MAXSYM];
	unsigned A[MaxRerror+1], B[MaxRerror+1];
	int num_read, j=0, lasti, base, ResidueSize; 
	int FIRST_TIME; /* Flag */
	CHAR *buffer;

	base = WORD - M;
	k = 2*exponen(M);
	if(FIRST_IN_RE) {
		compute_next(M, Next, Next1); 
		FIRST_IN_RE = 0;    
	}
	for(i=0; i< MAXSYM; i++) RMask[i] = Mask[i];
	r_NO_ERR = NO_ERR_MASK;
	Newline = '\n';
	Init0 = Init[0] = Bit[base];
	if(HEAD) Init0  = Init[0] = Init0 | Bit[base+1] ;
	for(i=1; i<= D; i++) Init[i] = Init[i-1] | Next[Init[i-1]]; /* can be out? */
	Init1 = Init0 | 1; 
	r2 = r3 = Init0;
	for(k=0; k<= D; k++) { 
		A[k] = B[k] = Init[0]; 
	}  /* can be out? */
	FIRST_TIME = ON;
	alloc_buf(Text, &buffer, BlockSize+Maxline+1);
	if ( D == 0 )
	{
#if	AGREP_POINTER
	    if(Text != -1) {
#endif	/*AGREP_POINTER*/
		lasti = Maxline;
		while ((num_read = fill_buf(Text, buffer + Maxline, BlockSize)) > 0)
		{
			i=Maxline; 
			end = Maxline + num_read ;
			if((num_read < BlockSize)&&buffer[end-1] != '\n') buffer[end] = '\n';
			if(FIRST_TIME) {
				buffer[i-1] = '\n';
				i--;
				CurrentByteOffset --;
				FIRST_TIME = 0;
			}

			/* RE_PROCESS_WHEN_DZERO: the while-loop below */
			while (i < end) 
			{
				c = buffer[i++];
				CurrentByteOffset ++;
				CMask = RMask[c];
				if(c != Newline)
				{  
					r1 = Init1 & r3;
					r2 = (Next[r3] & CMask) | r1;
				}
				else {  
					r1 = Init1 & r3;            /* match against '\n' */
					r2 = (Next[r3] & CMask) | r1;
					j++;
					if (DELIMITER) CurrentByteOffset -= D_length;
					else CurrentByteOffset -= 1;
					if(TAIL) r2 = Next[r2] | r2 ;   /* epsilon move */
					if(( r2 & 1) ^ INVERSE) {
						if(FILENAMEONLY && (NEW_FILE || !POST_FILTER)) {
							num_of_matched++;

							if (agrep_finalfp != NULL) 
								fprintf(agrep_finalfp, "%s\n", CurrentFileName);
							else {
								int outindex;
								for(outindex=0; (outindex+agrep_outpointer<agrep_outlen) && 
										(CurrentFileName[outindex] != '\0'); outindex++) {
									agrep_outbuffer[agrep_outpointer+outindex] = CurrentFileName[outindex];
								}
								if ((CurrentFileName[outindex] != '\0') || (outindex+agrep_outpointer+1>=agrep_outlen)) {
									OUTPUT_OVERFLOW;
									free_buf(Text, buffer);
									return -1;
								}
								else agrep_outbuffer[agrep_outpointer+outindex++] = '\n';
								agrep_outpointer += outindex;
							}

							free_buf(Text, buffer);
							NEW_FILE = OFF;
							return 0;
						} 
						if (-1 == r_output(buffer, i-1, end, j)) {free_buf(Text, buffer); return -1;}
						if (((LIMITOUTPUT > 0) && (LIMITOUTPUT <= num_of_matched)) ||
						    ((LIMITPERFILE > 0) && (LIMITPERFILE <= num_of_matched - prev_num_of_matched))) {
							free_buf(Text, buffer);
							return 0;	/* done */
						}
					}
					lasti = i - 1;
					r3 = Init0;
					r2 = (Next[r3] & CMask) | Init0;
					if (DELIMITER) CurrentByteOffset += 1*D_length;
					else CurrentByteOffset += 1*1;
				}
				c = buffer[i++];
				CurrentByteOffset ++;
				CMask = RMask[c];
				if(c != Newline)
				{
					r1 = Init1 & r2;
					r3 = (Next[r2] & CMask) | r1;
				}
				else {  
					j++;
					if (DELIMITER) CurrentByteOffset -= D_length;
					else CurrentByteOffset -= 1;
					r1 = Init1 & r2;            /* match against endofline */
					r3 = (Next[r2] & CMask) | r1;
					if(TAIL) r3 = Next[r3] | r3;
					if(( r3 & 1) ^ INVERSE) {
						if(FILENAMEONLY && (NEW_FILE || !POST_FILTER)) {
							num_of_matched++;

							if (agrep_finalfp != NULL) 
								fprintf(agrep_finalfp, "%s\n", CurrentFileName);
							else {
								int outindex;
								for(outindex=0; (outindex+agrep_outpointer<agrep_outlen) && 
										(CurrentFileName[outindex] != '\0'); outindex++) {
									agrep_outbuffer[agrep_outpointer+outindex] = CurrentFileName[outindex];
								}
								if ((CurrentFileName[outindex] != '\0') || (outindex+agrep_outpointer+1>=agrep_outlen)) {
									OUTPUT_OVERFLOW;
									free_buf(Text, buffer);
									return -1;
								}
								else agrep_outbuffer[agrep_outpointer+outindex++] = '\n';
								agrep_outpointer += outindex;
							}

							free_buf(Text, buffer);
							NEW_FILE = OFF;
							return 0;
						} 
						if (-1 == r_output(buffer, i-1, end, j)) {free_buf(Text, buffer); return -1;}
						if (((LIMITOUTPUT > 0) && (LIMITOUTPUT <= num_of_matched)) ||
						    ((LIMITPERFILE > 0) && (LIMITPERFILE <= num_of_matched - prev_num_of_matched))) {
							free_buf(Text, buffer);
							return 0;	/* done */
						}
					}
					lasti = i - 1;
					r2 = Init0; 
					r3 = (Next[r2] & CMask) | Init0;  /* match the newline */
					if (DELIMITER) CurrentByteOffset += 1*D_length;
					else CurrentByteOffset += 1*1;
				}
			} /* while */

			ResidueSize = Maxline + num_read - lasti;
			if(ResidueSize > Maxline) {
				ResidueSize = Maxline;  
			}
			strncpy(buffer+Maxline-ResidueSize, buffer+lasti, ResidueSize);
			lasti = Maxline - ResidueSize;
		} /* while fill_buf() */
		free_buf(Text, buffer);
		return 0;
#if	AGREP_POINTER
	    }
	    else {
		num_read = agrep_inlen;
		buffer = (CHAR *)agrep_inbuffer;
		end = num_read;
		/* buffer[end-1] = '\n';*/	/* at end of the text. */
		/* buffer[0] = '\n';*/		/* in front of the  text. */
		i = 0;
		lasti = 1;

			/* An exact copy of the above RE_PROCESS_WHEN_DZERO: the while-loop below */
			while (i < end) 
			{
				c = buffer[i++];
				CurrentByteOffset ++;
				CMask = RMask[c];
				if(c != Newline)
				{  
					r1 = Init1 & r3;
					r2 = (Next[r3] & CMask) | r1;
				}
				else {  
					r1 = Init1 & r3;            /* match against '\n' */
					r2 = (Next[r3] & CMask) | r1;
					j++;
					if (DELIMITER) CurrentByteOffset -= D_length;
					else CurrentByteOffset -= 1;
					if(TAIL) r2 = Next[r2] | r2 ;   /* epsilon move */
					if(( r2 & 1) ^ INVERSE) {
						if(FILENAMEONLY && (NEW_FILE || !POST_FILTER)) {
							num_of_matched++;

							if (agrep_finalfp != NULL) 
								fprintf(agrep_finalfp, "%s\n", CurrentFileName);
							else {
								int outindex;
								for(outindex=0; (outindex+agrep_outpointer<agrep_outlen) && 
										(CurrentFileName[outindex] != '\0'); outindex++) {
									agrep_outbuffer[agrep_outpointer+outindex] = CurrentFileName[outindex];
								}
								if ((CurrentFileName[outindex] != '\0') || (outindex+agrep_outpointer+1>=agrep_outlen)) {
									OUTPUT_OVERFLOW;
									free_buf(Text, buffer);
									return -1;
								}
								else agrep_outbuffer[agrep_outpointer+outindex++] = '\n';
								agrep_outpointer += outindex;
							}

							free_buf(Text, buffer);
							NEW_FILE = OFF;
							return 0;
						} 
						if (-1 == r_output(buffer, i-1, end, j)) {free_buf(Text, buffer); return -1;}
						if (((LIMITOUTPUT > 0) && (LIMITOUTPUT <= num_of_matched)) ||
						    ((LIMITPERFILE > 0) && (LIMITPERFILE <= num_of_matched - prev_num_of_matched))) {
							free_buf(Text, buffer);
							return 0;	/* done */
						}
					}
					lasti = i - 1;
					r3 = Init0;
					r2 = (Next[r3] & CMask) | Init0;
					if (DELIMITER) CurrentByteOffset += 1*D_length;
					else CurrentByteOffset += 1*1;
				}
				c = buffer[i++];   
				CurrentByteOffset ++;
				CMask = RMask[c];
				if(c != Newline)
				{
					r1 = Init1 & r2;
					r3 = (Next[r2] & CMask) | r1;
				}
				else {  
					j++;
					if (DELIMITER) CurrentByteOffset -= D_length;
					else CurrentByteOffset -= 1;
					r1 = Init1 & r2;            /* match against endofline */
					r3 = (Next[r2] & CMask) | r1;
					if(TAIL) r3 = Next[r3] | r3;
					if(( r3 & 1) ^ INVERSE) {
						if(FILENAMEONLY && (NEW_FILE || !POST_FILTER)) {
							num_of_matched++;

							if (agrep_finalfp != NULL) 
								fprintf(agrep_finalfp, "%s\n", CurrentFileName);
							else {
								int outindex;
								for(outindex=0; (outindex+agrep_outpointer<agrep_outlen) && 
										(CurrentFileName[outindex] != '\0'); outindex++) {
									agrep_outbuffer[agrep_outpointer+outindex] = CurrentFileName[outindex];
								}
								if ((CurrentFileName[outindex] != '\0') || (outindex+agrep_outpointer+1>=agrep_outlen)) {
									OUTPUT_OVERFLOW;
									free_buf(Text, buffer);
									return -1;
								}
								else agrep_outbuffer[agrep_outpointer+outindex++] = '\n';
								agrep_outpointer += outindex;
							}

							free_buf(Text, buffer);
							NEW_FILE = OFF;
							return 0;
						} 
						if (-1 == r_output(buffer, i-1, end, j)) {free_buf(Text, buffer); return -1;}
						if (((LIMITOUTPUT > 0) && (LIMITOUTPUT <= num_of_matched)) ||
						    ((LIMITPERFILE > 0) && (LIMITPERFILE <= num_of_matched - prev_num_of_matched))) {
							free_buf(Text, buffer);
							return 0;	/* done */
						}
					}
					lasti = i - 1;
					r2 = Init0; 
					r3 = (Next[r2] & CMask) | Init0;  /* match the newline */
					if (DELIMITER) CurrentByteOffset += 1*D_length;
					else CurrentByteOffset += 1*1;
				}
			} /* while */

		/* If a residue is left for within-memory-buffer, since nothing can be "read" after that, we can ignore it: as if only 1 iteration of while */
		return 0;
	    }
#endif	/*AGREP_POINTER*/
	} /* end if(D==0) */

#if	AGREP_POINTER
	if (Text != -1) {
#endif	/*AGREP_POINTER*/
		while ((num_read = fill_buf(Text, buffer + Maxline, BlockSize)) > 0)
		{
			i=Maxline; 
			end = Maxline+num_read;
			if((num_read < BlockSize) && buffer[end-1] != '\n') buffer[end] = '\n';
			if(FIRST_TIME) {
				buffer[i-1] = '\n';
				i--;
				CurrentByteOffset --;
				FIRST_TIME = 0;
			}

			/* RE_PROCESS_WHEN_DNOTZERO: the while-loop below */
			while (i < end)
			{   
				c = buffer[i++];
				CurrentByteOffset ++;
				CMask = RMask[c];
				if (c != Newline)
				{  
					r_even = B[0];
					r1 = Init1 & r_even;
					A[0] = (Next[r_even] & CMask) | r1;
					r_odd = B[1];
					r1 = Init1 & r_odd;
					r2 = (r_even | Next[r_even|A[0]]) &r_NO_ERR;
					A[1] = (Next[r_odd] & CMask) | r2 | r1 ;  
					if(D == 1) goto Nextcharfile;
					r_even = B[2];
					r1 = Init1 & r_even;