1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2006, 2009, 2010 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
18 #include "por-file-reader.h"
29 #include <data/casereader-provider.h>
30 #include <data/casereader.h>
31 #include <data/dictionary.h>
32 #include <data/file-handle-def.h>
33 #include <data/file-name.h>
34 #include <data/format.h>
35 #include <data/missing-values.h>
36 #include <data/short-names.h>
37 #include <data/value-labels.h>
38 #include <data/variable.h>
39 #include <libpspp/compiler.h>
40 #include <libpspp/message.h>
41 #include <libpspp/misc.h>
42 #include <libpspp/pool.h>
43 #include <libpspp/str.h>
49 #define _(msgid) gettext (msgid)
50 #define N_(msgid) (msgid)
52 /* portable_to_local[PORTABLE] translates the given portable
53 character into the local character set. */
54 static const char portable_to_local[256] =
57 "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz ."
58 "<(+|&[]!$*);^-/|,%_>?`:$@'=\" ~- 0123456789 -() {}\\ "
62 /* Portable file reader. */
65 struct pool *pool; /* All the portable file state. */
67 jmp_buf bail_out; /* longjmp() target for error handling. */
69 struct file_handle *fh; /* File handle. */
70 struct fh_lock *lock; /* Read lock for file. */
71 FILE *file; /* File stream. */
72 int line_length; /* Number of characters so far on this line. */
73 char cc; /* Current character. */
74 char *trans; /* 256-byte character set translation table. */
75 int var_cnt; /* Number of variables. */
76 int weight_index; /* 0-based index of weight variable, or -1. */
77 struct caseproto *proto; /* Format of output cases. */
78 bool ok; /* Set false on I/O error. */
81 static const struct casereader_class por_file_casereader_class;
84 error (struct pfm_reader *r, const char *msg,...)
88 /* Displays MSG as an error message and aborts reading the
89 portable file via longjmp(). */
91 error (struct pfm_reader *r, const char *msg, ...)
97 ds_init_empty (&text);
98 ds_put_format (&text, _("portable file %s corrupt at offset 0x%llx: "),
99 fh_get_file_name (r->fh), (long long int) ftello (r->file));
100 va_start (args, msg);
101 ds_put_vformat (&text, msg, args);
104 m.category = MSG_C_GENERAL;
105 m.severity = MSG_S_ERROR;
106 m.where.file_name = NULL;
107 m.where.line_number = 0;
108 m.where.first_column = 0;
109 m.where.last_column = 0;
110 m.text = ds_cstr (&text);
116 longjmp (r->bail_out, 1);
119 /* Displays MSG as an warning for the current position in
120 portable file reader R. */
122 warning (struct pfm_reader *r, const char *msg, ...)
128 ds_init_empty (&text);
129 ds_put_format (&text, _("reading portable file %s at offset 0x%llx: "),
130 fh_get_file_name (r->fh), (long long int) ftello (r->file));
131 va_start (args, msg);
132 ds_put_vformat (&text, msg, args);
135 m.category = MSG_C_GENERAL;
136 m.severity = MSG_S_WARNING;
137 m.where.file_name = NULL;
138 m.where.line_number = 0;
139 m.where.first_column = 0;
140 m.where.last_column = 0;
141 m.text = ds_cstr (&text);
146 /* Close and destroy R.
147 Returns false if an error was detected on R, true otherwise. */
149 close_reader (struct pfm_reader *r)
157 if (fn_close (fh_get_file_name (r->fh), r->file) == EOF)
159 msg (ME, _("Error closing portable file `%s': %s."),
160 fh_get_file_name (r->fh), strerror (errno));
170 pool_destroy (r->pool);
175 /* Closes portable file reader R, after we're done with it. */
177 por_file_casereader_destroy (struct casereader *reader, void *r_)
179 struct pfm_reader *r = r_;
180 if (!close_reader (r))
181 casereader_force_error (reader);
184 /* Read a single character into cur_char. */
186 advance (struct pfm_reader *r)
190 /* Read the next character from the file.
191 Ignore carriage returns entirely.
192 Mostly ignore new-lines, but if a new-line occurs before the
193 line has reached 80 bytes in length, then treat the
194 "missing" bytes as spaces. */
197 while ((c = getc (r->file)) == '\r')
202 if (r->line_length < 80)
205 ungetc ('\n', r->file);
211 error (r, _("unexpected end of file"));
213 if (r->trans != NULL)
219 /* Skip a single character if present, and return whether it was
222 match (struct pfm_reader *r, int c)
233 static void read_header (struct pfm_reader *);
234 static void read_version_data (struct pfm_reader *, struct pfm_read_info *);
235 static void read_variables (struct pfm_reader *, struct dictionary *);
236 static void read_value_label (struct pfm_reader *, struct dictionary *);
237 static void read_documents (struct pfm_reader *, struct dictionary *);
239 /* Reads the dictionary from file with handle H, and returns it in a
240 dictionary structure. This dictionary may be modified in order to
241 rename, reorder, and delete variables, etc. */
243 pfm_open_reader (struct file_handle *fh, struct dictionary **dict,
244 struct pfm_read_info *info)
246 struct pool *volatile pool = NULL;
247 struct pfm_reader *volatile r = NULL;
249 *dict = dict_create ();
251 /* Create and initialize reader. */
252 pool = pool_create ();
253 r = pool_alloc (pool, sizeof *r);
259 r->weight_index = -1;
264 if (setjmp (r->bail_out))
268 /* TRANSLATORS: this fragment will be interpolated into
269 messages in fh_lock() that identify types of files. */
270 r->lock = fh_lock (fh, FH_REF_FILE, N_("portable file"), FH_ACC_READ, false);
275 r->file = fn_open (fh_get_file_name (r->fh), "rb");
278 msg (ME, _("An error occurred while opening `%s' for reading "
279 "as a portable file: %s."),
280 fh_get_file_name (r->fh), strerror (errno));
284 /* Read header, version, date info, product id, variables. */
286 read_version_data (r, info);
287 read_variables (r, *dict);
289 /* Read value labels. */
290 while (match (r, 'D'))
291 read_value_label (r, *dict);
293 /* Read documents. */
295 read_documents (r, *dict);
297 /* Check that we've made it to the data. */
299 error (r, _("Data record expected."));
301 r->proto = caseproto_ref_pool (dict_get_proto (*dict), r->pool);
302 return casereader_create_sequential (NULL, r->proto, CASENUMBER_MAX,
303 &por_file_casereader_class, r);
307 dict_destroy (*dict);
312 /* Returns the value of base-30 digit C,
313 or -1 if C is not a base-30 digit. */
315 base_30_value (unsigned char c)
317 static const char base_30_digits[] = "0123456789ABCDEFGHIJKLMNOPQRST";
318 const char *p = strchr (base_30_digits, c);
319 return p != NULL ? p - base_30_digits : -1;
322 /* Read a floating point value and return its value. */
324 read_float (struct pfm_reader *r)
328 bool got_dot = false; /* Seen a decimal point? */
329 bool got_digit = false; /* Seen any digits? */
330 bool negative = false; /* Number is negative? */
332 /* Skip leading spaces. */
333 while (match (r, ' '))
336 /* `*' indicates system-missing. */
339 advance (r); /* Probably a dot (.) but doesn't appear to matter. */
343 negative = match (r, '-');
346 int digit = base_30_value (r->cc);
351 /* Make sure that multiplication by 30 will not overflow. */
352 if (num > DBL_MAX * (1. / 30.))
353 /* The value of the digit doesn't matter, since we have already
354 gotten as many digits as can be represented in a `double'.
355 This doesn't necessarily mean the result will overflow.
356 The exponent may reduce it to within range.
358 We just need to record that there was another
359 digit so that we can multiply by 10 later. */
362 num = (num * 30.0) + digit;
364 /* Keep track of the number of digits after the decimal point.
365 If we just divided by 30 here, we would lose precision. */
369 else if (!got_dot && r->cc == '.')
370 /* Record that we have found the decimal point. */
373 /* Any other character terminates the number. */
379 /* Check that we had some digits. */
381 error (r, _("Number expected."));
383 /* Get exponent if any. */
384 if (r->cc == '+' || r->cc == '-')
387 bool negative_exponent = r->cc == '-';
390 for (advance (r); (digit = base_30_value (r->cc)) != -1; advance (r))
392 if (exp > LONG_MAX / 30)
397 exp = exp * 30 + digit;
400 /* We don't check whether there were actually any digits, but we
402 if (negative_exponent)
407 /* Numbers must end with `/'. */
409 error (r, _("Missing numeric terminator."));
411 /* Multiply `num' by 30 to the `exponent' power, checking for
414 num *= pow (30.0, (double) exponent);
415 else if (exponent > 0)
417 if (num > DBL_MAX * pow (30.0, (double) -exponent))
420 num *= pow (30.0, (double) exponent);
423 return negative ? -num : num;
426 /* Read an integer and return its value. */
428 read_int (struct pfm_reader *r)
430 double f = read_float (r);
431 if (floor (f) != f || f >= INT_MAX || f <= INT_MIN)
432 error (r, _("Invalid integer."));
436 /* Reads a string into BUF, which must have room for 256
439 read_string (struct pfm_reader *r, char *buf)
441 int n = read_int (r);
442 if (n < 0 || n > 255)
443 error (r, _("Bad string length %d."), n);
454 /* Reads a string into BUF, which must have room for 256
456 Returns the number of bytes read.
459 read_bytes (struct pfm_reader *r, uint8_t *buf)
461 int n = read_int (r);
462 if (n < 0 || n > 255)
463 error (r, _("Bad string length %d."), n);
475 /* Reads a string and returns a copy of it allocated from R's
478 read_pool_string (struct pfm_reader *r)
481 read_string (r, string);
482 return pool_strdup (r->pool, string);
485 /* Reads the 464-byte file header. */
487 read_header (struct pfm_reader *r)
492 /* Read and ignore vanity splash strings. */
493 for (i = 0; i < 200; i++)
496 /* Skip the first 64 characters of the translation table.
497 We don't care about these. They are probably all set to
498 '0', marking them as untranslatable, and that would screw
499 up our actual translation of the real '0'. */
500 for (i = 0; i < 64; i++)
503 /* Read the rest of the translation table. */
504 trans = pool_malloc (r->pool, 256);
505 memset (trans, 0, 256);
514 trans[c] = portable_to_local[i];
517 /* Set up the translation table, then read the first
518 translated character. */
522 /* Skip and verify signature. */
523 for (i = 0; i < 8; i++)
524 if (!match (r, "SPSSPORT"[i]))
526 msg (SE, _("%s: Not a portable file."), fh_get_file_name (r->fh));
527 longjmp (r->bail_out, 1);
531 /* Reads the version and date info record, as well as product and
532 subproduct identification records if present. */
534 read_version_data (struct pfm_reader *r, struct pfm_read_info *info)
536 static const char empty_string[] = "";
538 const char *product, *author, *subproduct;
543 error (r, _("Unrecognized version code `%c'."), r->cc);
544 date = read_pool_string (r);
545 time = read_pool_string (r);
546 product = match (r, '1') ? read_pool_string (r) : empty_string;
547 author = match (r, '2') ? read_pool_string (r) : empty_string;
548 subproduct = match (r, '3') ? read_pool_string (r) : empty_string;
551 if (strlen (date) != 8)
552 error (r, _("Bad date string length %zu."), strlen (date));
553 if (strlen (time) != 6)
554 error (r, _("Bad time string length %zu."), strlen (time));
556 /* Save file info. */
560 for (i = 0; i < 8; i++)
562 static const int map[] = {6, 7, 8, 9, 3, 4, 0, 1};
563 info->creation_date[map[i]] = date[i];
565 info->creation_date[2] = info->creation_date[5] = ' ';
566 info->creation_date[10] = 0;
569 for (i = 0; i < 6; i++)
571 static const int map[] = {0, 1, 3, 4, 6, 7};
572 info->creation_time[map[i]] = time[i];
574 info->creation_time[2] = info->creation_time[5] = ' ';
575 info->creation_time[8] = 0;
578 str_copy_trunc (info->product, sizeof info->product, product);
579 str_copy_trunc (info->subproduct, sizeof info->subproduct, subproduct);
583 /* Translates a format specification read from portable file R as
584 the three integers INTS into a normal format specifier FORMAT,
585 checking that the format is appropriate for variable V. */
586 static struct fmt_spec
587 convert_format (struct pfm_reader *r, const int portable_format[3],
588 struct variable *v, bool *report_error)
590 struct fmt_spec format;
593 if (!fmt_from_io (portable_format[0], &format.type))
596 warning (r, _("%s: Bad format specifier byte (%d). Variable "
597 "will be assigned a default format."),
598 var_get_name (v), portable_format[0]);
602 format.w = portable_format[1];
603 format.d = portable_format[2];
606 ok = (fmt_check_output (&format)
607 && fmt_check_width_compat (&format, var_get_width (v)));
614 char fmt_string[FMT_STRING_LEN_MAX + 1];
615 fmt_to_string (&format, fmt_string);
616 if (var_is_numeric (v))
617 warning (r, _("Numeric variable %s has invalid format "
619 var_get_name (v), fmt_string);
621 warning (r, _("String variable %s with width %d has "
622 "invalid format specifier %s."),
623 var_get_name (v), var_get_width (v), fmt_string);
631 *report_error = false;
632 return fmt_default_for_width (var_get_width (v));
635 static void parse_value (struct pfm_reader *, int width, union value *);
637 /* Read information on all the variables. */
639 read_variables (struct pfm_reader *r, struct dictionary *dict)
641 char *weight_name = NULL;
645 error (r, _("Expected variable count record."));
647 r->var_cnt = read_int (r);
649 error (r, _("Invalid number of variables %d."), r->var_cnt);
651 /* Purpose of this value is unknown. It is typically 161. */
656 weight_name = read_pool_string (r);
657 if (strlen (weight_name) > SHORT_NAME_LEN)
658 error (r, _("Weight variable name (%s) truncated."), weight_name);
661 for (i = 0; i < r->var_cnt; i++)
667 struct missing_values miss;
668 struct fmt_spec print, write;
669 bool report_error = true;
673 error (r, _("Expected variable record."));
675 width = read_int (r);
677 error (r, _("Invalid variable width %d."), width);
679 read_string (r, name);
680 for (j = 0; j < 6; j++)
681 fmt[j] = read_int (r);
683 if (!var_is_valid_name (name, false) || *name == '#' || *name == '$')
684 error (r, _("Invalid variable name `%s' in position %d."), name, i);
685 str_uppercase (name);
687 if (width < 0 || width > 255)
688 error (r, _("Bad width %d for variable %s."), width, name);
690 v = dict_create_var (dict, name, width);
694 for (i = 1; i < 100000; i++)
696 char try_name[VAR_NAME_LEN + 1];
697 sprintf (try_name, "%.*s_%d", VAR_NAME_LEN - 6, name, i);
698 v = dict_create_var (dict, try_name, width);
703 error (r, _("Duplicate variable name %s in position %d."), name, i);
704 warning (r, _("Duplicate variable name %s in position %d renamed "
705 "to %s."), name, i, var_get_name (v));
708 print = convert_format (r, &fmt[0], v, &report_error);
709 write = convert_format (r, &fmt[3], v, &report_error);
710 var_set_print_format (v, &print);
711 var_set_write_format (v, &write);
713 /* Range missing values. */
714 mv_init (&miss, width);
717 double x = read_float (r);
718 double y = read_float (r);
719 mv_add_range (&miss, x, y);
721 else if (match (r, 'A'))
722 mv_add_range (&miss, read_float (r), HIGHEST);
723 else if (match (r, '9'))
724 mv_add_range (&miss, LOWEST, read_float (r));
726 /* Single missing values. */
727 while (match (r, '8'))
729 int mv_width = MIN (width, 8);
732 parse_value (r, mv_width, &value);
733 value_resize (&value, mv_width, width);
734 mv_add_value (&miss, &value);
735 value_destroy (&value, width);
738 var_set_missing_values (v, &miss);
744 read_string (r, label);
745 var_set_label (v, label);
749 if (weight_name != NULL)
751 struct variable *weight_var = dict_lookup_var (dict, weight_name);
752 if (weight_var == NULL)
753 error (r, _("Weighting variable %s not present in dictionary."),
756 dict_set_weight (dict, weight_var);
760 /* Parse a value of with WIDTH into value V. */
762 parse_value (struct pfm_reader *r, int width, union value *v)
764 value_init (v, width);
768 size_t n_bytes = read_bytes (r, buf);
769 value_copy_buf_rpad (v, width, buf, n_bytes, ' ');
772 v->f = read_float (r);
775 /* Parse a value label record and return success. */
777 read_value_label (struct pfm_reader *r, struct dictionary *dict)
789 v = pool_nalloc (r->pool, nv, sizeof *v);
790 for (i = 0; i < nv; i++)
793 read_string (r, name);
795 v[i] = dict_lookup_var (dict, name);
797 error (r, _("Unknown variable %s while parsing value labels."), name);
799 if (var_get_type (v[0]) != var_get_type (v[i]))
800 error (r, _("Cannot assign value labels to %s and %s, which "
801 "have different variable types."),
802 var_get_name (v[0]), var_get_name (v[i]));
805 n_labels = read_int (r);
806 for (i = 0; i < n_labels; i++)
812 parse_value (r, var_get_width (v[0]), &val);
813 read_string (r, label);
815 /* Assign the value label to each variable. */
816 for (j = 0; j < nv; j++)
817 var_replace_value_label (v[j], &val, label);
819 value_destroy (&val, var_get_width (v[0]));
823 /* Reads a set of documents from portable file R into DICT. */
825 read_documents (struct pfm_reader *r, struct dictionary *dict)
830 line_cnt = read_int (r);
831 for (i = 0; i < line_cnt; i++)
834 read_string (r, line);
835 dict_add_document_line (dict, line);
839 /* Reads and returns one case from portable file R. Returns a
840 null pointer on failure. */
841 static struct ccase *
842 por_file_casereader_read (struct casereader *reader, void *r_)
844 struct pfm_reader *r = r_;
845 struct ccase *volatile c;
848 c = case_create (r->proto);
849 setjmp (r->bail_out);
852 casereader_force_error (reader);
857 /* Check for end of file. */
864 for (i = 0; i < r->var_cnt; i++)
866 int width = caseproto_get_width (r->proto, i);
869 case_data_rw_idx (c, i)->f = read_float (r);
873 size_t n_bytes = read_bytes (r, buf);
874 u8_buf_copy_rpad (case_str_rw_idx (c, i), width, buf, n_bytes, ' ');
881 /* Returns true if FILE is an SPSS portable file,
884 pfm_detect (FILE *file)
886 unsigned char header[464];
888 int cooked_cnt, raw_cnt, line_len;
891 cooked_cnt = raw_cnt = 0;
893 while (cooked_cnt < sizeof header)
896 if (c == EOF || raw_cnt++ > 512)
900 while (line_len < 80 && cooked_cnt < sizeof header)
902 header[cooked_cnt++] = ' ';
909 header[cooked_cnt++] = c;
914 memset (trans, 0, 256);
915 for (i = 64; i < 256; i++)
917 unsigned char c = header[i + 200];
919 trans[c] = portable_to_local[i];
922 for (i = 0; i < 8; i++)
923 if (trans[header[i + 456]] != "SPSSPORT"[i])
929 static const struct casereader_class por_file_casereader_class =
931 por_file_casereader_read,
932 por_file_casereader_destroy,