gui: Make text import assistant accept only one quote character.
[pspp] / src / ui / gui / psppire-delimited-text.c
index f95cb594a5183c6a4a5974b2a3323b02f6051cba..59b423010ec10a118da91823d2705d4e6fc05383 100644 (file)
@@ -32,82 +32,61 @@ enum
     PROP_0,
     PROP_CHILD,
     PROP_DELIMITERS,
+    PROP_QUOTE,
     PROP_FIRST_LINE
   };
 
-struct enclosure
-{
-  gunichar opening;
-  gunichar closing;
-};
-
-static const struct enclosure enclosures[3] =
-  {
-    {'(',   ')'},
-    {'"',   '"'},
-    {'\'',  '\''}
-  };
-
 static void
 count_delims (PsppireDelimitedText *tf)
 {
-  if (tf->child)
+  if (tf->child == NULL)
+    return;
+
+  tf->max_delimiters = 0;
+  GtkTreeIter iter;
+  gboolean valid;
+  for (valid = gtk_tree_model_get_iter_first (tf->child, &iter);
+       valid;
+       valid = gtk_tree_model_iter_next (tf->child, &iter))
     {
-      tf->max_delimiters = 0;
-      GtkTreeIter iter;
-      gboolean valid;
-      for (valid = gtk_tree_model_get_iter_first (tf->child, &iter);
-          valid;
-          valid = gtk_tree_model_iter_next (tf->child, &iter))
-       {
-         gint enc = -1;
-         // FIXME: Box these lines to avoid constant allocation/deallocation
-         gchar *foo = 0;
-         gtk_tree_model_get (tf->child, &iter, 1, &foo, -1);
+      gunichar quote = -1;
+      // FIXME: Box these lines to avoid constant allocation/deallocation
+      gchar *line = NULL;
+      gtk_tree_model_get (tf->child, &iter, 1, &line, -1);
+      {
+       char *p;
+       gint count = 0;
+       for (p = line; ; p = g_utf8_find_next_char (p, NULL))
          {
-           char *line = foo;
-           gint count = 0;
-           while (*line)
+           const gunichar c = g_utf8_get_char (p);
+           if (c == 0)
+             break;
+
+            if (c == quote)
+              quote = -1;
+            else if (tf->quote && c == tf->quote)
+              quote = c;
+
+           if (quote == -1)
              {
-               const gunichar c = *line; //FIXME: Not multibyte safe!
-               if (enc == -1)
+               GSList *del;
+               for (del = tf->delimiters; del; del = g_slist_next (del))
                  {
-                   gint i;
-                   for (i = 0; i < 3; ++i)
-                     {
-                       if (c == enclosures[i].opening)
-                         {
-                           enc = i;
-                           break;
-                         }
-                     }
+                   if (c == GPOINTER_TO_INT (del->data))
+                     count++;
                  }
-               else if (c == enclosures[enc].closing)
-                 {
-                   enc = -1;
-                 }
-               if (enc == -1)
-                 {
-                   GSList *del;
-                   for (del = tf->delimiters; del; del = g_slist_next (del))
-                     {
-                       if (c == GPOINTER_TO_INT (del->data))
-                         count++;
-                     }
-                 }
-               line++;
              }
-           tf->max_delimiters = MAX (tf->max_delimiters, count);
          }
-         g_free (foo);
-       }
+       tf->max_delimiters = MAX (tf->max_delimiters, count);
+      }
+      g_free (line);
     }
 }
 
 static void
 cache_invalidate (PsppireDelimitedText *tf)
 {
-  memset (tf->cache_starts, 0, 512);
+  memset (tf->cache_starts, 0, sizeof tf->cache_starts);
   if (tf->const_cache.string)
     {
       ss_dealloc (&tf->const_cache);
@@ -137,6 +116,9 @@ psppire_delimited_text_set_property (GObject         *object,
       g_slist_free (tf->delimiters);
       tf->delimiters =  g_slist_copy (g_value_get_pointer (value));
       break;
+    case PROP_QUOTE:
+      tf->quote = g_value_get_uint (value);
+      break;
     default:
       G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
       break;
@@ -162,16 +144,15 @@ psppire_delimited_text_get_property (GObject         *object,
     case PROP_DELIMITERS:
       g_value_set_pointer (value, text_file->delimiters);
       break;
+    case PROP_QUOTE:
+      g_value_set_uint (value, text_file->quote);
+      break;
     default:
       G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
       break;
     };
 }
 
-
-static void psppire_delimited_text_init            (PsppireDelimitedText      *text_file);
-static void psppire_delimited_text_class_init      (PsppireDelimitedTextClass *class);
-
 static void psppire_delimited_text_finalize        (GObject           *object);
 static void psppire_delimited_text_dispose        (GObject           *object);
 
@@ -369,7 +350,7 @@ split_row_into_fields (PsppireDelimitedText *file, gint n)
       return;
     }
 
-  memset (file->cache_starts, 0, 512);
+  memset (file->cache_starts, 0, sizeof file->cache_starts);
   /* Cache miss */
   if (file->const_cache.string)
     {
@@ -380,35 +361,30 @@ split_row_into_fields (PsppireDelimitedText *file, gint n)
   struct substring cs = file->const_cache;
   int field = 0;
   file->cache_starts[0] = cs.string;
-  gint enc = -1;
+  gunichar quote = -1;
   for (;
        UINT32_MAX != ss_first_mb (cs);
        ss_get_mb (&cs))
     {
       ucs4_t character = ss_first_mb (cs);
       gboolean char_is_quote = FALSE;
-      if (enc == -1)
-       {
-         gint i;
-         for (i = 0; i < 3; ++i)
-           {
-             if (character == enclosures[i].opening)
-               {
-                 enc = i;
-                 char_is_quote = TRUE;
-                 file->cache_starts[field] += ss_first_mblen (cs);
-                 break;
-               }
-           }
-       }
-      else if (character == enclosures[enc].closing)
+      if (quote == -1)
+        {
+          if (file->quote && character == file->quote)
+            {
+              quote = character;
+              char_is_quote = TRUE;
+              file->cache_starts[field] += ss_first_mblen (cs);
+            }
+        }
+      else if (character == quote)
        {
          char_is_quote = TRUE;
          nullify_char (cs);
-         enc = -1;
+         quote = -1;
        }
 
-      if (enc == -1 && char_is_quote == FALSE)
+      if (quote == -1 && char_is_quote == FALSE)
        {
          GSList *del;
          for (del = file->delimiters; del; del = g_slist_next (del))
@@ -485,44 +461,9 @@ __tree_model_init (GtkTreeModelIface *iface)
   iface->iter_parent     = __iter_parent;
 }
 
-
-GType
-psppire_delimited_text_get_type (void)
-{
-  static GType text_file_type = 0;
-
-  if (!text_file_type)
-    {
-      static const GTypeInfo text_file_info =
-       {
-         sizeof (PsppireDelimitedTextClass),
-         NULL,         /* base_init */
-         NULL,         /* base_finalize */
-         (GClassInitFunc) psppire_delimited_text_class_init,
-         NULL,         /* class_finalize */
-         NULL,         /* class_data */
-         sizeof (PsppireDelimitedText),
-         0,
-         (GInstanceInitFunc) psppire_delimited_text_init,
-       };
-
-      static const GInterfaceInfo tree_model_info = {
-       (GInterfaceInitFunc) __tree_model_init,
-       NULL,
-       NULL
-      };
-
-      text_file_type = g_type_register_static (G_TYPE_OBJECT,
-                                              "PsppireDelimitedText",
-                                              &text_file_info, 0);
-
-      g_type_add_interface_static (text_file_type, GTK_TYPE_TREE_MODEL,
-                                  &tree_model_info);
-    }
-
-  return text_file_type;
-}
-
+G_DEFINE_TYPE_WITH_CODE (PsppireDelimitedText, psppire_delimited_text, G_TYPE_OBJECT,
+                        G_IMPLEMENT_INTERFACE (GTK_TYPE_TREE_MODEL,
+                                               __tree_model_init))
 
 static void
 psppire_delimited_text_class_init (PsppireDelimitedTextClass *class)
@@ -530,7 +471,7 @@ psppire_delimited_text_class_init (PsppireDelimitedTextClass *class)
   GObjectClass *object_class;
 
   parent_class = g_type_class_peek_parent (class);
-  object_class = (GObjectClass*) class;
+  object_class = G_OBJECT_CLASS (class);
 
   GParamSpec *first_line_spec =
     g_param_spec_int ("first-line",
@@ -545,6 +486,13 @@ psppire_delimited_text_class_init (PsppireDelimitedTextClass *class)
                          P_("A GSList of gunichars which delimit the fields."),
                          G_PARAM_READWRITE);
 
+  GParamSpec *quote_spec =
+    g_param_spec_unichar ("quote",
+                         "Quote Character",
+                         P_("A character that quotes the field, or 0 to disable quoting."),
+                         0,
+                         G_PARAM_READWRITE);
+
   GParamSpec *child_spec =
     g_param_spec_object ("child",
                         "Child Model",
@@ -563,6 +511,10 @@ psppire_delimited_text_class_init (PsppireDelimitedTextClass *class)
                                    PROP_DELIMITERS,
                                    delimiters_spec);
 
+  g_object_class_install_property (object_class,
+                                   PROP_QUOTE,
+                                   quote_spec);
+
   g_object_class_install_property (object_class,
                                    PROP_FIRST_LINE,
                                    first_line_spec);
@@ -582,24 +534,24 @@ psppire_delimited_text_init (PsppireDelimitedText *text_file)
   text_file->const_cache.string = NULL;
   text_file->const_cache.length = 0;
   text_file->cache_row = -1;
-  memset (text_file->cache_starts, 0, 512);
+  memset (text_file->cache_starts, 0, sizeof text_file->cache_starts);
 
   text_file->max_delimiters = 0;
 
+  text_file->quote = 0;
+
   text_file->dispose_has_run = FALSE;
   text_file->stamp = g_random_int ();
 }
 
 
-GtkTreeModel *
+PsppireDelimitedText *
 psppire_delimited_text_new (GtkTreeModel *child)
 {
-  PsppireDelimitedText *retval =
+  return
     g_object_new (PSPPIRE_TYPE_DELIMITED_TEXT,
                  "child", child,
                  NULL);
-
-  return GTK_TREE_MODEL (retval);
 }
 
 static void