Complete adblock

3 files changed, 81 insertions, 8 deletions

diff --git a/Makefile b/Makefile
index 09e396e..2c9a12f 100644
--- a/Makefile
+++ b/Makefile
@@ -3,7 +3,7 @@ CFLAGS += -Wall -Wextra -O3
 zea: zea.c
 	$(CC) $(CFLAGS) $(LDFLAGS) \
 		-o $@ $< \
-		`pkg-config --cflags --libs gtk+-2.0 webkit-1.0`
+		`pkg-config --cflags --libs gtk+-2.0 glib-2.0 webkit-1.0`
 
 clean:
 	rm -f zea
diff --git a/README b/README
index 056fe1f..d6bff5d 100644
--- a/README
+++ b/README
@@ -1,9 +1,9 @@
 zea - zeig's einfach an
 =======================
 
-zea is a minimalistic web browser using Gtk2 and WebKit. "Zeig's einfach
-an" is german and translates roughly to "just show it" (the damn web
-page).
+zea is a minimalistic web browser using Gtk2, GLib and WebKit. "Zeig's
+einfach an" is german and translates roughly to "just show it" (the damn
+web page).
 
 Features:
 
@@ -13,11 +13,11 @@ Features:
 	- Pluggability into suckless' tabbed
 	- vi-like scrolling (modified by CTRL)
 	- Searching the current page for a word
+	- Adblock
 	- Support for Flash and Java
 
 Planned features:
 
-	- Adblock
 	- Keyword based searching (opening "wi foo" will search wikipedia)
 
 
@@ -33,6 +33,18 @@ Each new tab will then show your bookmarks and is scaled by a factor of
 0.8.
 
 
+Adblock
+=======
+
+zea has built-in adblock functionality. In each line of
+
+	~/.config/zea/adblock.black
+
+you can store a regular expression. These expressions match
+case-insensitive and partially, i.e. ".*foo.*" is the same as ".*FOO.*"
+and you can use anchors like "^https?://...".
+
+
 Literature
 ==========
 
@@ -40,3 +52,9 @@ API references:
 
 	- http://webkitgtk.org/reference/webkitgtk/stable/index.html
 	- https://developer.gnome.org/gtk2/stable/index.html
+	- https://developer.gnome.org/glib/stable/index.html
+
+Regular expressions supported by GRegex, you can use these in your
+adblock patterns:
+
+	- https://developer.gnome.org/glib/stable/glib-regex-syntax.html
diff --git a/zea.c b/zea.c
index 9fc7d54..d10d892 100644
--- a/zea.c
+++ b/zea.c
@@ -17,6 +17,7 @@ static gboolean zea_do_download(WebKitWebView *, WebKitDownload *, gpointer);
 static gboolean zea_download_request(WebKitWebView *, WebKitWebFrame *,
                                      WebKitNetworkRequest *, gchar *,
                                      WebKitWebPolicyDecision *, gpointer);
+static void zea_load_adblock(void);
 static void zea_load_status_changed(GObject *obj, GParamSpec *pspec,
                                     gpointer data);
 static gboolean zea_location_key(GtkWidget *, GdkEvent *, gpointer);
@@ -39,6 +40,7 @@ static gdouble global_zoom = 1.0;
 static gchar *search_text = NULL;
 static gchar *first_uri = NULL;
 static gboolean show_all_requests = FALSE;
+static GSList *adblock_patterns = NULL;
 
 
 struct Client
@@ -57,17 +59,30 @@ zea_adblock(WebKitWebView *web_view, WebKitWebFrame *frame,
             WebKitWebResource *resource, WebKitNetworkRequest *request,
             WebKitNetworkResponse *response, gpointer data)
 {
+	GSList *it = adblock_patterns;
+	const gchar *uri;
+
 	(void)web_view;
 	(void)frame;
 	(void)resource;
 	(void)response;
 	(void)data;
 
+	uri = webkit_network_request_get_uri(request);
 	if (show_all_requests)
-		fprintf(stderr, "-> %s\n", webkit_network_request_get_uri(request));
+		fprintf(stderr, "-> %s\n", uri);
 
-	/* XXX Changing the URI here using webkit_network_request_set_uri()
-	 * effectively blocks the request. */
+	while (it)
+	{
+		if (g_regex_match((GRegex *)(it->data), uri, 0, NULL))
+		{
+			webkit_network_request_set_uri(request, "about:blank");
+			if (show_all_requests)
+				fprintf(stderr, "\tBLOCKED!\n");
+			return;
+		}
+		it = g_slist_next(it);
+	}
 }
 
 void
@@ -327,6 +342,44 @@ zea_uri_changed(GObject *obj, GParamSpec *pspec, gpointer data)
 }
 
 void
+zea_load_adblock(void)
+{
+	GRegex *re = NULL;
+	GError *err = NULL;
+	GIOChannel *channel = NULL;
+	gchar *path = NULL;
+	gchar *buf = NULL;
+	gsize length, term;
+
+	path = g_strdup_printf("%s/zea/adblock.black", g_get_user_config_dir());
+	channel = g_io_channel_new_file(path, "r", &err);
+	if (channel != NULL)
+	{
+		while (g_io_channel_read_line(channel, &buf, &length, &term, &err)
+		       == G_IO_STATUS_NORMAL)
+		{
+			g_strstrip(buf);
+			re = g_regex_new(buf,
+			                 G_REGEX_CASELESS | G_REGEX_OPTIMIZE,
+			                 G_REGEX_MATCH_PARTIAL, &err);
+			if (err != NULL)
+			{
+				fprintf(stderr, "zea: Could not compile regex: %s\n", buf);
+				g_error_free(err);
+				err = NULL;
+			}
+			adblock_patterns = g_slist_append(adblock_patterns, re);
+
+			g_free(buf);
+		}
+
+		if (err != NULL)
+			g_free(err);
+	}
+	g_free(path);
+}
+
+void
 zea_load_status_changed(GObject *obj, GParamSpec *pspec, gpointer data)
 {
 	struct Client *c = (struct Client *)data;
@@ -470,6 +523,8 @@ main(int argc, char **argv)
 		exit(EXIT_FAILURE);
 	}
 
+	zea_load_adblock();
+
 	first_uri = g_strdup(argv[optind]);
 	for (i = optind; i < argc; i++)
 		zea_new_client(argv[i]);