Make the gnumeric reader behave the same as the odsreader

[pspp] / src / data / gnumeric-reader.c

diff --git a/src/data/gnumeric-reader.c b/src/data/gnumeric-reader.c
index 5e4b6978ef55ebc7c3239b438f8d564b1816ac05..9193cfb061c4949be89b8467877174bcbf1a0264 100644 (file)
--- a/src/data/gnumeric-reader.c
+++ b/src/data/gnumeric-reader.c
@@ -28,12 +28,10 @@
 
 #include "spreadsheet-reader.h"
 
-#include "c-xvasprintf.h"
-
 #if !GNM_SUPPORT
 
 struct casereader *
-gnumeric_open_reader (struct spreadsheet_read_info *gri, struct spreadsheet_read_options *opts, struct dictionary **dict)
+gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictionary **dict)
 {
   msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "Gnumeric");
 
@@ -65,6 +63,7 @@ static void gnm_file_casereader_destroy (struct casereader *, void *);
 
 static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
 
+
 static const struct casereader_class gnm_file_casereader_class =
   {
     gnm_file_casereader_read,
@@ -89,7 +88,8 @@ enum reader_state
 
 struct sheet_detail
 {
-  xmlChar *name;
+  /* The name of the sheet (utf8 encoding) */
+  char *name;
 
   int start_col;
   int stop_col;
@@ -104,6 +104,7 @@ struct sheet_detail
 struct gnumeric_reader
 {
   struct spreadsheet spreadsheet;
+  int ref_cnt;
 
   /* The libxml reader for this instance */
   xmlTextReaderPtr xtr;
@@ -134,19 +135,44 @@ struct gnumeric_reader
 };
 
 
+void
+gnumeric_destroy (struct spreadsheet *s)
+{
+  struct gnumeric_reader *r = (struct gnumeric_reader *) s;
+
+
+#if 0
+  if (0 == --r->ref_cnt)
+    {
+      int i;
+
+      for (i = 0; i < s->n_sheets; ++i)
+       {
+         xmlFree (r->sheets[i].name);
+       }
+    
+      free (r->sheets);
+
+      free (r);
+    }
+#endif
+}
+
+
 const char *
 gnumeric_get_sheet_name (struct spreadsheet *s, int n)
 {
   struct gnumeric_reader *gr = (struct gnumeric_reader *) s;
   assert (n < s->n_sheets);
 
-  return gr->sheets[n].name; // Kludge: Assumes the encoding is utf8
+  return gr->sheets[n].name; 
 }
 
 
 static void process_node (struct gnumeric_reader *r);
 
 
+
 char *
 gnumeric_get_sheet_range (struct spreadsheet *s, int n)
 {
@@ -176,9 +202,11 @@ static void
 gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
 {
   struct gnumeric_reader *r = r_;
+
   if ( r == NULL)
        return ;
 
+#if 0
   if ( r->xtr)
     xmlFreeTextReader (r->xtr);
   r->xtr = NULL;
@@ -188,19 +216,11 @@ gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
 
   caseproto_unref (r->proto);
 
-#if 0
-  for (i = 0; i < r->spreadsheet.n_sheets; ++i)
-    {
-      xmlFree (r->sheets[i].name);
-    }
-    
-  free (r->sheets);
-
-
-  free (r);
+  gnumeric_destroy (&r->spreadsheet);
 #endif
 }
 
+
 static void
 process_node (struct gnumeric_reader *r)
 {
@@ -243,7 +263,7 @@ process_node (struct gnumeric_reader *r)
        }
       else if (XML_READER_TYPE_TEXT == r->node_type)
        {
-         r->sheets [r->spreadsheet.n_sheets - 1].name = xmlTextReaderValue (r->xtr);
+         r->sheets [r->spreadsheet.n_sheets - 1].name = CHAR_CAST (char *, xmlTextReaderValue (r->xtr));
        }
       break;
 
@@ -425,20 +445,14 @@ struct var_spec
 };
 
 
-void 
-gnumeric_destroy (struct spreadsheet *s)
-{
-  gnm_file_casereader_destroy (NULL, s);
-}
-
-
 static void
 gnumeric_error_handler (void *ctx, const char *mesg,
                        UNUSED xmlParserSeverities sev, xmlTextReaderLocatorPtr loc)
 {
   struct gnumeric_reader *r = ctx;
        
-  msg (MW, _("There was a problem whilst reading the Gnumeric file `%s' (near line %d): `%s'"),
+  msg (MW, _("There was a problem whilst reading the %s file `%s' (near line %d): `%s'"),
+       "Gnumeric",
        r->spreadsheet.file_name,
        xmlTextReaderLocatorLineNumber (loc),
        mesg);
@@ -493,6 +507,7 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro
   r->row = r->col = -1;
   r->state = STATE_PRE_INIT;
   r->xtr = xtr;
+  r->ref_cnt++;
 
   /* Advance to the start of the workbook.
      This gives us some confidence that we are actually dealing with a gnumeric
@@ -504,6 +519,7 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro
       process_node (r);
     }
 
+
   if ( ret != 1)
     {
       /* Does not seem to be a gnumeric file */
@@ -514,6 +530,22 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro
 
   r->spreadsheet.type = SPREADSHEET_GNUMERIC;
 
+  if (show_errors)
+    {
+      const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
+      xmlCharEncoding xce = xmlParseCharEncoding (CHAR_CAST (const char *, enc));
+
+      if ( XML_CHAR_ENCODING_UTF8 != xce)
+       {
+         /* I have been told that ALL gnumeric files are UTF8 encoded.  If that is correct, this 
+            can never happen. */
+         msg (MW, _("The gnumeric file `%s' is encoded as %s instead of the usual UTF-8 encoding. "
+                    "Any non-ascii characters will be incorrectly imported."),
+              r->spreadsheet.file_name,
+              enc);
+       }
+    }
+
   return r;
 }
 
@@ -529,9 +561,9 @@ gnumeric_probe (const char *filename, bool report_errors)
 
 struct casereader *
 gnumeric_make_reader (struct spreadsheet *spreadsheet,
-                     const struct spreadsheet_read_info *gri, 
                      const struct spreadsheet_read_options *opts)
 {
+  int x = 0;
   struct gnumeric_reader *r = NULL;
   unsigned long int vstart = 0;
   int ret;
@@ -545,6 +577,8 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
   if (r->row != -1)
     r = gnumeric_reopen (r, NULL, true);
 
+
+
   if ( opts->cell_range )
     {
       if ( ! convert_cell_ref (opts->cell_range,
@@ -591,7 +625,7 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
       n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
     }
 
-  if ( gri->read_names )
+  if ( opts->read_names )
     {
       r->start_row++;
       n_cases --;
@@ -635,7 +669,7 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
 
          if ( r->row < r->start_row)
            {
-             if ( gri->read_names )
+             if ( opts->read_names )
                {
                  var_spec [idx].name = xstrdup (text);
                }
@@ -645,8 +679,8 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
              var_spec [idx].first_value = xmlStrdup (value);
 
              if (-1 ==  var_spec [idx].width )
-               var_spec [idx].width = (gri->asw == -1) ?
-                 ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : gri->asw;
+               var_spec [idx].width = (opts->asw == -1) ?
+                 ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : opts->asw;
            }
 
          free (value);
@@ -706,13 +740,15 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
   r->first_case = case_create (r->proto);
   case_set_missing (r->first_case);
 
-  int x = 0;
+
   for ( i = 0 ; i < n_var_specs ; ++i )
     {
+      const struct variable *var;
+
       if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
        continue;
 
-      const struct variable *var = dict_get_var (r->dict, x++);
+      var = dict_get_var (r->dict, x++);
 
       convert_xml_string_to_value (r->first_case, var,
                                   var_spec[i].first_value);
@@ -727,17 +763,6 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet,
   free (var_spec);
   
 
-#if 0  
-  if (opts->cell_range == NULL)
-    {
-      opts->cell_range = c_xasprintf ("%c%d:%c%ld", 
-                                      r->start_col + 'A',
-                                      r->start_row,
-                                      r->stop_col + 'A' + caseproto_get_n_widths (r->proto),
-                                      r->start_row + n_cases);
-    }
-#endif
-  
   return casereader_create_sequential
     (NULL,
      r->proto,