[evolution/wip/webkit2] Bug 745943 - [ESpellEntry] Do not rely on pango with word-split

[Milan Crha <mcrha src gnome org>]

commit 200689e7d5acc5443a746c31fe53e3327ed09dc9
Author: Milan Crha <mcrha redhat com>
Date:   Fri Aug 28 16:11:09 2015 +0200

    Bug 745943 - [ESpellEntry] Do not rely on pango with word-split

 e-util/e-spell-entry.c |  109 ++++++++++++++++++++++++++++++++++--------------
 1 files changed, 78 insertions(+), 31 deletions(-)
---
diff --git a/e-util/e-spell-entry.c b/e-util/e-spell-entry.c
index adebbd8..12d5a50 100644
--- a/e-util/e-spell-entry.c
+++ b/e-util/e-spell-entry.c
@@ -211,55 +211,72 @@ get_word_extents_from_position (ESpellEntry *entry,
 }
 
 static void
+spell_entry_store_word (gchar ***set,
+                       gint **starts,
+                       gint **ends,
+                       const gchar *text,
+                       gint n_word,
+                       gint n_strings,
+                       const gchar *word_start,
+                       const gchar *word_end)
+{
+       gint bytes;
+
+       g_return_if_fail (n_word >= 0);
+       g_return_if_fail (n_word < n_strings);
+
+       /* Copy sub-string */
+       bytes = (gint) (word_end - word_start);
+       (*set)[n_word] = g_new0 (gchar, bytes + 1);
+       (*starts)[n_word] = (gint) (word_start - text);
+       (*ends)[n_word] = (gint) (word_start - text + bytes);
+       memcpy ((*set)[n_word], word_start, bytes);
+}
+
+static void
 entry_strsplit_utf8 (GtkEntry *entry,
                      gchar ***set,
                      gint **starts,
                      gint **ends)
 {
-       PangoLayout *layout;
-       PangoLogAttr *log_attrs;
-       const gchar *text;
-       gint n_attrs, n_strings, i, j;
+       const gchar *text, *ptr, *word_start;
+       gint n_strings, n_word;
 
-       layout = gtk_entry_get_layout (GTK_ENTRY (entry));
        text = gtk_entry_get_text (GTK_ENTRY (entry));
-       pango_layout_get_log_attrs (layout, &log_attrs, &n_attrs);
+       g_return_if_fail (g_utf8_validate (text, -1, NULL));
 
        /* Find how many words we have */
        n_strings = 0;
-       for (i = 0; i < n_attrs; i++)
-               if (log_attrs[i].is_word_start)
+       word_start = NULL;
+       for (ptr = text; *ptr; ptr = g_utf8_next_char (ptr)) {
+               if (g_unichar_isspace (g_utf8_get_char (ptr))) {
+                       word_start = NULL;
+               } else if (!word_start) {
                        n_strings++;
+                       word_start = ptr;
+               }
+       }
 
        *set = g_new0 (gchar *, n_strings + 1);
-       *starts = g_new0 (gint, n_strings);
-       *ends = g_new0 (gint, n_strings);
+       *starts = g_new0 (gint, n_strings + 1);
+       *ends = g_new0 (gint, n_strings + 1);
 
        /* Copy out strings */
-       for (i = 0, j = 0; i < n_attrs; i++) {
-               if (log_attrs[i].is_word_start) {
-                       gint cend, bytes;
-                       gchar *start;
-
-                       /* Find the end of this string */
-                       cend = i;
-                       while (!(log_attrs[cend].is_word_end))
-                               cend++;
-
-                       /* Copy sub-string */
-                       start = g_utf8_offset_to_pointer (text, i);
-                       bytes = (gint) (g_utf8_offset_to_pointer (text, cend) - start);
-                       (*set)[j] = g_new0 (gchar, bytes + 1);
-                       (*starts)[j] = (gint) (start - text);
-                       (*ends)[j] = (gint) (start - text + bytes);
-                       g_utf8_strncpy ((*set)[j], start, cend - i);
-
-                       /* Move on to the next word */
-                       j++;
+       word_start = NULL;
+       n_word = -1;
+       for (ptr = text; *ptr; ptr = g_utf8_next_char (ptr)) {
+               if (g_unichar_isspace (g_utf8_get_char (ptr))) {
+                       if (word_start)
+                               spell_entry_store_word (set, starts, ends, text, n_word, n_strings, 
word_start, ptr);
+                       word_start = NULL;
+               } else if (!word_start) {
+                       n_word++;
+                       word_start = ptr;
                }
        }
 
-       g_free (log_attrs);
+       if (word_start)
+               spell_entry_store_word (set, starts, ends, text, n_word, n_strings, word_start, ptr);
 }
 
 static gchar *
@@ -292,6 +309,33 @@ spell_entry_get_chars_from_byte_pos (ESpellEntry *entry,
 }
 
 static void
+spell_entry_byte_pos_to_char_pos (ESpellEntry *entry,
+                                 gint byte_pos,
+                                 gint *out_char_pos)
+{
+       const gchar *text, *ptr;
+
+       g_return_if_fail (E_IS_SPELL_ENTRY (entry));
+       g_return_if_fail (out_char_pos != NULL);
+
+       *out_char_pos = 0;
+
+       if (byte_pos <= 0)
+               return;
+
+       text = gtk_entry_get_text (GTK_ENTRY (entry));
+       if (!text || !g_utf8_validate (text, -1, NULL))
+               return;
+
+       for (ptr = text; ptr && *ptr; ptr = g_utf8_next_char (ptr)) {
+               if (byte_pos <= ptr - text)
+                       break;
+
+               *out_char_pos = (*out_char_pos) + 1;
+       }
+}
+
+static void
 add_to_dictionary (GtkWidget *menuitem,
                    ESpellEntry *entry)
 {
@@ -372,6 +416,9 @@ replace_word (GtkWidget *menuitem,
        newword = gtk_label_get_text (
                GTK_LABEL (gtk_bin_get_child (GTK_BIN (menuitem))));
 
+       spell_entry_byte_pos_to_char_pos (entry, start, &start);
+       spell_entry_byte_pos_to_char_pos (entry, end, &end);
+
        cursor = gtk_editable_get_position (GTK_EDITABLE (entry));
        /* is the cursor at the end? If so, restore it there */
        if (g_utf8_strlen (gtk_entry_get_text (GTK_ENTRY (entry)), -1) == cursor)