sys-file-reader: Fully verify multiple response set names.

[pspp] / src / data / sys-file-reader.c

diff --git a/src/data/sys-file-reader.c b/src/data/sys-file-reader.c
index 54788deb6e217bb918d8873077e2cb305ffbea77..dbc61345a85869eda4c0d1eb7f0f0aa04fe4c9ce 100644 (file)
--- a/src/data/sys-file-reader.c
+++ b/src/data/sys-file-reader.c
@@ -203,6 +203,7 @@ struct sfm_reader
     size_t sfm_var_cnt;         /* Number of variables. */
     int case_cnt;               /* Number of cases */
     const char *encoding;       /* String encoding. */
+    bool written_by_readstat; /* From https://github.com/WizardMac/ReadStat? */
 
     /* Decompression. */
     enum any_compression compression;
@@ -861,7 +862,7 @@ sfm_decode (struct any_reader *r_, const char *encoding,
      amount that the header claims.  SPSS version 13 gets this
      wrong when very long strings are involved, so don't warn in
      that case. */
-  if (r->header.nominal_case_size != -1
+  if (r->header.nominal_case_size > 0
       && r->header.nominal_case_size != r->n_vars
       && r->info.version_major != 13)
     sys_warn (r, -1, _("File header claims %d variable positions but "
@@ -967,6 +968,8 @@ read_header (struct sfm_reader *r, struct any_read_info *info,
   if (!read_string (r, header->magic, sizeof header->magic)
       || !read_string (r, header->eye_catcher, sizeof header->eye_catcher))
     return false;
+  r->written_by_readstat = strstr (header->eye_catcher,
+                                   "https://github.com/WizardMac/ReadStat");
 
   if (!strcmp (ASCII_MAGIC, header->magic)
       || !strcmp (EBCDIC_MAGIC, header->magic))
@@ -1828,10 +1831,9 @@ decode_mrsets (struct sfm_reader *r, struct dictionary *dict)
       size_t i;
 
       name = recode_string ("UTF-8", r->encoding, s->name, -1);
-      if (name[0] != '$')
+      if (!mrset_is_valid_name (name, dict_get_encoding (dict), false))
         {
-          sys_warn (r, -1, _("Multiple response set name `%s' does not begin "
-                             "with `$'."),
+          sys_warn (r, -1, _("Invalid multiple response set name `%s'."),
                     name);
           free (name);
           continue;
@@ -2228,7 +2230,15 @@ parse_value_labels (struct sfm_reader *r, struct dictionary *dict,
 
           if (!var_add_value_label (var, &value, utf8_labels[j]))
             {
-              if (var_is_numeric (var))
+              if (r->written_by_readstat)
+                {
+                  /* Ignore the problem.  ReadStat is buggy and emits value
+                     labels whose values are longer than string variables'
+                     widths, that are identical in the actual width of the
+                     variable, e.g. both values "ABC123" and "ABC456" for a
+                     string variable with width 3. */
+                }
+              else if (var_is_numeric (var))
                 sys_warn (r, record->pos,
                           _("Duplicate value label for %g on %s."),
                           value.f, var_get_name (var));
@@ -2975,7 +2985,7 @@ open_text_record (struct sfm_reader *r,
 }
 
 /* Closes TEXT, frees its storage, and issues a final warning
-   about suppressed warnings if necesary. */
+   about suppressed warnings if necessary. */
 static void
 close_text_record (struct sfm_reader *r, struct text_record *text)
 {