postgresql/contrib/dbase/dbf.c

/* Routines to read and write xBase-files (.dbf)

   By Maarten Boekhold, 29th of oktober 1995

   Modified by Frank Koormann (fkoorman@usf.uni-osnabrueck.de), Jun 10 1996
	prepare dataarea with memset
	get systemtime and set filedate
	set formatstring for real numbers
*/

#include "postgres_fe.h"

#include <fcntl.h>
#include <ctype.h>
#include <time.h>
#include <unistd.h>

#include "dbf.h"

/* open a dbf-file, get it's field-info and store this information */

dbhead *
dbf_open(char *file, int flags)
{
	int			file_no;
	dbhead	   *dbh;
	f_descr    *fields;
	dbf_header *head;
	dbf_field  *fieldc;
	int			t;

	if ((dbh = (dbhead *) malloc(sizeof(dbhead))) == NULL)
		return (dbhead *) DBF_ERROR;

	if ((head = (dbf_header *) malloc(sizeof(dbf_header))) == NULL)
	{
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

	if ((fieldc = (dbf_field *) malloc(sizeof(dbf_field))) == NULL)
	{
		free(head);
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

	if ((file_no = open(file, flags)) == -1)
	{
		free(fieldc);
		free(head);
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

/* read in the disk-header */

	if (read(file_no, head, sizeof(dbf_header)) == -1)
	{
		close(file_no);
		free(fieldc);
		free(head);
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

	if (!(head->dbh_dbt & DBH_NORMAL))
	{
		close(file_no);
		free(fieldc);
		free(head);
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

	dbh->db_fd = file_no;
	if (head->dbh_dbt & DBH_MEMO)
		dbh->db_memo = 1;
	else
		dbh->db_memo = 0;
	dbh->db_year = head->dbh_year;
	dbh->db_month = head->dbh_month;
	dbh->db_day = head->dbh_day;
	dbh->db_hlen = get_short((u_char *) &head->dbh_hlen);
	dbh->db_records = get_long((u_char *) &head->dbh_records);
	dbh->db_currec = 0;
	dbh->db_rlen = get_short((u_char *) &head->dbh_rlen);
	dbh->db_nfields = (dbh->db_hlen - sizeof(dbf_header)) / sizeof(dbf_field);

	/*
	 * dbh->db_hlen - sizeof(dbf_header) isn't the correct size, cos dbh->hlen
	 * is in fact a little more cos of the 0x0D (and possibly another byte,
	 * 0x4E, I have seen this somewhere). Because of rounding everything turns
	 * out right :)
	 */

	if ((fields = (f_descr *) calloc(dbh->db_nfields, sizeof(f_descr)))
		== NULL)
	{
		close(file_no);
		free(fieldc);
		free(head);
		free(dbh);
		return (dbhead *) DBF_ERROR;
	}

	for (t = 0; t < dbh->db_nfields; t++)
	{
/* Maybe I have calculated the number of fields incorrectly. This can happen
   when programs reserve lots of space at the end of the header for future
   expansion. This will catch this situation */
		if (fields[t].db_name[0] == 0x0D)
		{
			dbh->db_nfields = t;
			break;
		}
		read(file_no, fieldc, sizeof(dbf_field));
		strncpy(fields[t].db_name, fieldc->dbf_name, DBF_NAMELEN);
		fields[t].db_type = fieldc->dbf_type;
		fields[t].db_flen = fieldc->dbf_flen;
		fields[t].db_dec = fieldc->dbf_dec;
	}

	dbh->db_offset = dbh->db_hlen;
	dbh->db_fields = fields;

	if ((dbh->db_buff = (u_char *) malloc(dbh->db_rlen)) == NULL)
		return (dbhead *) DBF_ERROR;

	free(fieldc);
	free(head);

	return dbh;
}

int
dbf_write_head(dbhead * dbh)
{
	dbf_header	head;
	time_t		now;
	struct tm  *dbf_time;

	if (lseek(dbh->db_fd, 0, SEEK_SET) == -1)
		return DBF_ERROR;

/* fill up the diskheader */

/* Set dataarea of head to '\0' */
	memset(&head, '\0', sizeof(dbf_header));

	head.dbh_dbt = DBH_NORMAL;
	if (dbh->db_memo)
		head.dbh_dbt = DBH_MEMO;

	now = time((time_t *) NULL);
	dbf_time = localtime(&now);
	head.dbh_year = dbf_time->tm_year;
	head.dbh_month = dbf_time->tm_mon + 1;		/* Months since January + 1 */
	head.dbh_day = dbf_time->tm_mday;

	put_long(head.dbh_records, dbh->db_records);
	put_short(head.dbh_hlen, dbh->db_hlen);
	put_short(head.dbh_rlen, dbh->db_rlen);

	if (write(dbh->db_fd, &head, sizeof(dbf_header)) != sizeof(dbf_header))
		return DBF_ERROR;

	return 0;
}

int
dbf_put_fields(dbhead * dbh)
{
	dbf_field	field;
	u_long		t;
	u_char		end = 0x0D;

	if (lseek(dbh->db_fd, sizeof(dbf_header), SEEK_SET) == -1)
		return DBF_ERROR;

/* Set dataarea of field to '\0' */
	memset(&field, '\0', sizeof(dbf_field));

	for (t = 0; t < dbh->db_nfields; t++)
	{
		strncpy(field.dbf_name, dbh->db_fields[t].db_name, DBF_NAMELEN - 1);
		field.dbf_type = dbh->db_fields[t].db_type;
		field.dbf_flen = dbh->db_fields[t].db_flen;
		field.dbf_dec = dbh->db_fields[t].db_dec;

		if (write(dbh->db_fd, &field, sizeof(dbf_field)) != sizeof(dbf_field))
			return DBF_ERROR;
	}

	if (write(dbh->db_fd, &end, 1) != 1)
		return DBF_ERROR;

	return 0;
}

int
dbf_add_field(dbhead * dbh, char *name, u_char type,
			  u_char length, u_char dec)
{
	f_descr    *ptr;
	u_char	   *foo;
	u_long		size,
				field_no;

	size = (dbh->db_nfields + 1) * sizeof(f_descr);
	if (!(ptr = (f_descr *) realloc(dbh->db_fields, size)))
		return DBF_ERROR;
	dbh->db_fields = ptr;

	field_no = dbh->db_nfields;
	strncpy(dbh->db_fields[field_no].db_name, name, DBF_NAMELEN);
	dbh->db_fields[field_no].db_type = type;
	dbh->db_fields[field_no].db_flen = length;
	dbh->db_fields[field_no].db_dec = dec;

	dbh->db_nfields++;
	dbh->db_hlen += sizeof(dbf_field);
	dbh->db_rlen += length;

	if (!(foo = (u_char *) realloc(dbh->db_buff, dbh->db_rlen)))
		return DBF_ERROR;

	dbh->db_buff = foo;

	return 0;
}

dbhead *
dbf_open_new(char *name, int flags)
{
	dbhead	   *dbh;

	if (!(dbh = (dbhead *) malloc(sizeof(dbhead))))
		return (dbhead *) DBF_ERROR;

	if (flags & O_CREAT)
	{
		if ((dbh->db_fd = open(name, flags, DBF_FILE_MODE)) == -1)
		{
			free(dbh);
			return (dbhead *) DBF_ERROR;
		}
	}
	else
	{
		if ((dbh->db_fd = open(name, flags)) == -1)
		{
			free(dbh);
			return (dbhead *) DBF_ERROR;
		}
	}


	dbh->db_offset = 0;
	dbh->db_memo = 0;
	dbh->db_year = 0;
	dbh->db_month = 0;
	dbh->db_day = 0;
	dbh->db_hlen = sizeof(dbf_header) + 1;
	dbh->db_records = 0;
	dbh->db_currec = 0;
	dbh->db_rlen = 1;
	dbh->db_nfields = 0;
	dbh->db_buff = NULL;
	dbh->db_fields = (f_descr *) NULL;

	return dbh;
}

void
dbf_close(dbhead * dbh)
{
	int			t;

	close(dbh->db_fd);

	for (t = 0; t < dbh->db_nfields; t++)
		free(&dbh->db_fields[t]);

	if (dbh->db_buff != NULL)
		free(dbh->db_buff);

	free(dbh);
}

int
dbf_get_record(dbhead * dbh, field * fields, u_long rec)
{
	u_char	   *data;
	int			t,
				i,
				offset;
	u_char	   *dbffield,
			   *end;

/* calculate at which offset we have to read. *DON'T* forget the
   0x0D which seperates field-descriptions from records!

	Note (april 5 1996): This turns out to be included in db_hlen
*/
	offset = dbh->db_hlen + (rec * dbh->db_rlen);

	if (lseek(dbh->db_fd, offset, SEEK_SET) == -1)
	{
		lseek(dbh->db_fd, 0, SEEK_SET);
		dbh->db_offset = 0;
		return DBF_ERROR;
	}

	dbh->db_offset = offset;
	dbh->db_currec = rec;
	data = dbh->db_buff;

	read(dbh->db_fd, data, dbh->db_rlen);

	if (data[0] == DBF_DELETED)
		return DBF_DELETED;

	dbffield = &data[1];
	for (t = 0; t < dbh->db_nfields; t++)
	{
		strncpy(fields[t].db_name, dbh->db_fields[t].db_name, DBF_NAMELEN);
		fields[t].db_type = dbh->db_fields[t].db_type;
		fields[t].db_flen = dbh->db_fields[t].db_flen;
		fields[t].db_dec = dbh->db_fields[t].db_dec;

		if (fields[t].db_type == 'C')
		{
			end = &dbffield[fields[t].db_flen - 1];
			i = fields[t].db_flen;
			while (i > 0 && !isprint(*end))
			{
				end--;
				i--;
			}
			strncpy((char *) fields[t].db_contents, (char *) dbffield, i);
			fields[t].db_contents[i] = '\0';
		}
		else
		{
			end = dbffield;
			i = fields[t].db_flen;
			while (i > 0 && !isprint(*end))
			{
				end++;
				i--;
			}
			strncpy((char *) fields[t].db_contents, (char *) end, i);
			fields[t].db_contents[i] = '\0';
		}

		dbffield += fields[t].db_flen;
	}

	dbh->db_offset += dbh->db_rlen;

	return DBF_VALID;
}

field *
dbf_build_record(dbhead * dbh)
{
	int			t;
	field	   *fields;

	if (!(fields = (field *) calloc(dbh->db_nfields, sizeof(field))))
		return (field *) DBF_ERROR;

	for (t = 0; t < dbh->db_nfields; t++)
	{
		if (!(fields[t].db_contents =
			  (u_char *) malloc(dbh->db_fields[t].db_flen + 1)))
		{
			for (t = 0; t < dbh->db_nfields; t++)
			{
				if (fields[t].db_contents != 0)
				{
					free(fields[t].db_contents);
					free(fields);
				}
				return (field *) DBF_ERROR;
			}
		}
		strncpy(fields[t].db_name, dbh->db_fields[t].db_name, DBF_NAMELEN);
		fields[t].db_type = dbh->db_fields[t].db_type;
		fields[t].db_flen = dbh->db_fields[t].db_flen;
		fields[t].db_dec = dbh->db_fields[t].db_dec;
	}

	return fields;
}

void
dbf_free_record(dbhead * dbh, field * rec)
{
	int			t;

	for (t = 0; t < dbh->db_nfields; t++)
		free(rec[t].db_contents);

	free(rec);
}

int
dbf_put_record(dbhead * dbh, field * rec, u_long where)
{
	u_long		offset,
				new,
				idx,
				t,
				h,
				length;
	u_char	   *data,
				end = 0x1a;
	double		fl;
	char		foo[128],
				format[32];

/*	offset: offset in file for this record
	new:	real offset after lseek
	idx:	index to which place we are inside the 'hardcore'-data for this
			record
	t:		field-counter
	data:	the hardcore-data that is put on disk
	h:		index into the field-part in the hardcore-data
	length: length of the data to copy
	fl:		a float used to get the right precision with real numbers
	foo:	copy of db_contents when field is not 'C'
	format: sprintf format-string to get the right precision with real numbers

	NOTE: this declaration of 'foo' can cause overflow when the contents-field
	is longer the 127 chars (which is highly unlikely, because it is not used
	in text-fields).
*/
/*	REMEMBER THAT THERE'S A 0x1A AT THE END OF THE FILE, SO DON'T
	DO A SEEK_END WITH 0!!!!!! USE -1 !!!!!!!!!!
*/

	if (where > dbh->db_records)
	{
		if ((new = lseek(dbh->db_fd, -1, SEEK_END)) == -1)
			return DBF_ERROR;
		dbh->db_records++;
	}
	else
	{
		offset = dbh->db_hlen + (where * dbh->db_rlen);
		if ((new = lseek(dbh->db_fd, offset, SEEK_SET)) == -1)
			return DBF_ERROR;
	}

	dbh->db_offset = new;

	data = dbh->db_buff;

/* Set dataarea of data to ' ' (space) */
	memset(data, ' ', dbh->db_rlen);

/*	data[0] = DBF_VALID; */

	idx = 1;
	for (t = 0; t < dbh->db_nfields; t++)
	{
/* if field is empty, don't do a thing */
		if (rec[t].db_contents[0] != '\0')
		{
/*	Handle text */
			if (rec[t].db_type == 'C')
			{
				if (strlen((char *) rec[t].db_contents) > rec[t].db_flen)
					length = rec[t].db_flen;
				else
					length = strlen((char *) rec[t].db_contents);
				strncpy((char *) data + idx, (char *) rec[t].db_contents,
						length);
			}
			else
			{
/* Handle the rest */
/* Numeric is special, because of real numbers */
				if ((rec[t].db_type == 'N') && (rec[t].db_dec != 0))
				{
					fl = atof((char *) rec[t].db_contents);
					snprintf(format, 32, "%%.%df", rec[t].db_dec);
					snprintf(foo, 128, format, fl);
				}
				else
					strncpy(foo, (char *) rec[t].db_contents, 128);
				if (strlen(foo) > rec[t].db_flen)
					length = rec[t].db_flen;
				else
					length = strlen(foo);
				h = rec[t].db_flen - length;
				strncpy((char *) (data + idx + h), foo, length);
			}
		}
		idx += rec[t].db_flen;
	}

	if (write(dbh->db_fd, data, dbh->db_rlen) != dbh->db_rlen)
		return DBF_ERROR;

	/* There's a 0x1A at the end of a dbf-file */
	if (where == dbh->db_records)
	{
		if (write(dbh->db_fd, &end, 1) != 1)
			return DBF_ERROR;
	}

	dbh->db_offset += dbh->db_rlen;

	return 0;
}