[evolution] EHTMLEditorView - Fix appending a text to the link

[Tomas Popela <tpopela src gnome org>]

commit f7cbd1865d8479911417fbafde422558eb44476f
Author: Tomas Popela <tpopela redhat com>
Date:   Mon Jun 15 13:19:09 2015 +0200

    EHTMLEditorView - Fix appending a text to the link
    
    Don't try to append the text to the link if it starts with punctuation
    or contains spaces. Also fix one leak.

 e-util/e-html-editor-view.c |   37 +++++++++++++++++++++++++------------
 1 files changed, 25 insertions(+), 12 deletions(-)
---
diff --git a/e-util/e-html-editor-view.c b/e-util/e-html-editor-view.c
index bfcc88b..9ae398c 100644
--- a/e-util/e-html-editor-view.c
+++ b/e-util/e-html-editor-view.c
@@ -39,6 +39,8 @@
 #define URL_PROTOCOLS "news|telnet|nntp|file|https?|s?ftp||webcal|localhost"
 #define URL_PATTERN "((((" URL_PROTOCOLS ")\\:\\/\\/)|(www\\.|ftp\\.))[^\\s\\/\\$\\.\\?#].[^\\s]*)"
 #define URL_PATTERN_SPACE URL_PATTERN "\\s"
+/* Taken from camel-url-scanner.c */
+#define URL_INVALID_TRAILING_CHARS ",.:;?!-|}])\""
 
 /* http://www.w3.org/TR/html5/forms.html#valid-e-mail-address */
 #define E_MAIL_PATTERN \
@@ -1537,13 +1539,12 @@ html_editor_view_check_magic_links (EHTMLEditorView *view,
        urls = g_match_info_fetch_all (match_info);
 
        if (urls) {
-               gchar *final_url, *url_end_raw;
+               const gchar *end_of_match = NULL;
+               gchar *final_url, *url_end_raw, *url_text;
                glong url_start, url_end, url_length;
                WebKitDOMDocument *document;
-               WebKitDOMNode *url_text_node_clone;
                WebKitDOMText *url_text_node;
                WebKitDOMElement *anchor;
-               const gchar* url_text;
 
                document = webkit_web_view_get_dom_document (WEBKIT_WEB_VIEW (view));
 
@@ -1556,8 +1557,18 @@ html_editor_view_check_magic_links (EHTMLEditorView *view,
                 * that we get from g_match_info_fetch_pos are not UTF-8 aware */
                url_end_raw = g_strndup(node_text, end_pos_url);
                url_end = g_utf8_strlen (url_end_raw, -1);
-
                url_length = g_utf8_strlen (urls[0], -1);
+
+               end_of_match = url_end_raw + end_pos_url - (include_space ? 3 : 2);
+               /* URLs are extremely unlikely to end with any punctuation, so
+                * strip any trailing punctuation off from link and put it after
+                * the link. Do the same for any closing double-quotes as well. */
+               while (end_of_match && end_of_match != url_end_raw && strchr (URL_INVALID_TRAILING_CHARS, 
*end_of_match)) {
+                       url_length--;
+                       url_end--;
+                       end_of_match--;
+               }
+
                url_start = url_end - url_length;
 
                webkit_dom_text_split_text (
@@ -1565,12 +1576,11 @@ html_editor_view_check_magic_links (EHTMLEditorView *view,
                        include_space ? url_end - 1 : url_end,
                        NULL);
 
-               url_text_node = webkit_dom_text_split_text (
+               webkit_dom_text_split_text (
                        WEBKIT_DOM_TEXT (node), url_start, NULL);
-               url_text_node_clone = webkit_dom_node_clone_node (
-                       WEBKIT_DOM_NODE (url_text_node), TRUE);
-               url_text = webkit_dom_text_get_whole_text (
-                       WEBKIT_DOM_TEXT (url_text_node_clone));
+               url_text_node = WEBKIT_DOM_TEXT (webkit_dom_node_get_next_sibling (node));
+               url_text = webkit_dom_character_data_get_data (
+                       WEBKIT_DOM_CHARACTER_DATA (url_text_node));
 
                if (g_str_has_prefix (url_text, "www."))
                        final_url = g_strconcat ("http://"; , url_text, NULL);
@@ -1603,6 +1613,7 @@ html_editor_view_check_magic_links (EHTMLEditorView *view,
 
                g_free (url_end_raw);
                g_free (final_url);
+               g_free (url_text);
        } else {
                gboolean appending_to_link = FALSE;
                gchar *href, *text, *url, *text_to_append = NULL;
@@ -1618,7 +1629,8 @@ html_editor_view_check_magic_links (EHTMLEditorView *view,
                if (WEBKIT_DOM_IS_HTML_ANCHOR_ELEMENT (prev_sibling)) {
                        text_to_append = webkit_dom_node_get_text_content (node);
                        if (text_to_append && *text_to_append &&
-                           !g_unichar_isspace (g_utf8_get_char (text_to_append))) {
+                           !strstr (text_to_append, " ") &&
+                           !strchr (URL_INVALID_TRAILING_CHARS, *text_to_append)) {
 
                                appending_to_link = TRUE;
                                parent = WEBKIT_DOM_ELEMENT (prev_sibling);
@@ -2863,7 +2875,8 @@ body_input_event_cb (WebKitDOMElement *element,
 
                        text = webkit_dom_node_get_text_content (node);
 
-                       if (g_strcmp0 (text, "") != 0 && !g_unichar_isspace (g_utf8_get_char (text))) {
+                       if (text && *text && !strstr (text, " ") &&
+                           !strchr (URL_INVALID_TRAILING_CHARS, *text)) {
                                WebKitDOMNode *prev_sibling;
 
                                prev_sibling = webkit_dom_node_get_previous_sibling (node);
@@ -5800,7 +5813,7 @@ create_anchor_for_link (const GMatchInfo *info,
        /* URLs are extremely unlikely to end with any punctuation, so
         * strip any trailing punctuation off from link and put it after
         * the link. Do the same for any closing double-quotes as well. */
-       while (end_of_match && end_of_match != match && strchr (",.:;?!-|}])\"", *end_of_match)) {
+       while (end_of_match && end_of_match != match && strchr (URL_INVALID_TRAILING_CHARS, *end_of_match)) {
                truncate_from_end++;
                end_of_match--;
        }