[evolution] Bug 759564 - Pasting a plain text paragraph starting with an URL will linkify the entire paragraph

[Tomas Popela <tpopela src gnome org>]

commit 2227a0b5b4eb5a1cf0f245f3ebaa395a8619fc04
Author: Tomas Popela <tpopela redhat com>
Date:   Thu Jan 7 14:13:12 2016 +0100

    Bug 759564 - Pasting a plain text paragraph starting with an URL will linkify the entire paragraph
    
    Correct the regex that is used for links to return the right group when
    non-breaking space is presented in the link. Also fix the create_anchor_for_link
    that was not adopted for recent regex changes.

 e-util/e-html-editor-view.c |   37 ++++++++++++++++++++++++++++---------
 1 files changed, 28 insertions(+), 9 deletions(-)
---
diff --git a/e-util/e-html-editor-view.c b/e-util/e-html-editor-view.c
index bab9a7b..71b56ff 100644
--- a/e-util/e-html-editor-view.c
+++ b/e-util/e-html-editor-view.c
@@ -36,9 +36,9 @@
        ((obj), E_TYPE_HTML_EDITOR_VIEW, EHTMLEditorViewPrivate))
 
 /* stephenhay from https://mathiasbynens.be/demo/url-regex */
-#define URL_PROTOCOLS "news|telnet|nntp|file|https?|s?ftp||webcal|localhost|ssh"
+#define URL_PROTOCOLS "news|telnet|nntp|file|https?|s?ftp|webcal|localhost|ssh"
 #define URL_PATTERN_BASE "(?=((?:(?:(?:" URL_PROTOCOLS 
")\\:\\/\\/)|(?:www\\.|ftp\\.))[^\\s\\/\\$\\.\\?#].[^\\s]*)"
-#define URL_PATTERN_NO_NBSP ")((?!&nbsp;).)*"
+#define URL_PATTERN_NO_NBSP ")((?:(?!&nbsp;).)*)"
 #define URL_PATTERN URL_PATTERN_BASE URL_PATTERN_NO_NBSP
 #define URL_PATTERN_SPACE URL_PATTERN_BASE "\\s$" URL_PATTERN_NO_NBSP
 /* Taken from camel-url-scanner.c */
@@ -6912,17 +6912,32 @@ create_anchor_for_link (const GMatchInfo *info,
                         GString *res,
                         gpointer data)
 {
+       gboolean link_surrounded, with_nbsp = FALSE;
        gint offset = 0, truncate_from_end = 0;
+       gint match_start, match_end;
+       gchar *match_with_nbsp, *match_without_nbsp;
        const gchar *end_of_match = NULL;
-       gchar *match;
-       gboolean link_surrounded;
+       const gchar *match, *match_extra_characters;
 
-       match = g_match_info_fetch (info, 0);
+       match_with_nbsp = g_match_info_fetch (info, 1);
+       /* E-mail addresses will be here. */
+       match_without_nbsp = g_match_info_fetch (info, 0);
+
+       if (!match_with_nbsp || (strstr (match_with_nbsp, "&nbsp;") && !g_str_has_prefix (match_with_nbsp, 
"&nbsp;"))) {
+               match = match_without_nbsp;
+               match_extra_characters = match_with_nbsp;
+               g_match_info_fetch_pos (info, 0, &match_start, &match_end);
+               with_nbsp = TRUE;
+       } else {
+               match = match_with_nbsp;
+               match_extra_characters = match_without_nbsp;
+               g_match_info_fetch_pos (info, 1, &match_start, &match_end);
+       }
 
        if (g_str_has_prefix (match, "&nbsp;"))
                offset += 6;
 
-       end_of_match = match + strlen (match) - 1;
+       end_of_match = match + match_end - match_start - 1;
        /* Taken from camel-url-scanner.c */
        /* URLs are extremely unlikely to end with any punctuation, so
         * strip any trailing punctuation off from link and put it after
@@ -6966,7 +6981,11 @@ create_anchor_for_link (const GMatchInfo *info,
        if (truncate_from_end > 0)
                g_string_append (res, end_of_match);
 
-       g_free (match);
+       if (!with_nbsp && match_extra_characters)
+               g_string_append (res, match_extra_characters + (match_end - match_start));
+
+       g_free (match_with_nbsp);
+       g_free (match_without_nbsp);
 
        return FALSE;
 }
@@ -7281,7 +7300,7 @@ parse_html_into_blocks (EHTMLEditorView *view,
                                        rest_to_insert,
                                        -1,
                                        0,
-                                       0,
+                                       G_REGEX_MATCH_NOTEMPTY,
                                        create_anchor_for_link,
                                        NULL,
                                        NULL);
@@ -7460,7 +7479,7 @@ parse_html_into_blocks (EHTMLEditorView *view,
                                rest_to_insert,
                                -1,
                                0,
-                               0,
+                               G_REGEX_MATCH_NOTEMPTY,
                                create_anchor_for_link,
                                NULL,
                                NULL);