[gaim-migrate @ 14581]

Thu, 01 Dec 2005 00:36:19 +0000

author
Richard Laager <rlaager@pidgin.im>
date
Thu, 01 Dec 2005 00:36:19 +0000
changeset 12277
6e45a609022c
parent 12276
f16c914b6d58
child 12278
21bcbadf956d

[gaim-migrate @ 14581]
Readding the error checking that I took out in the last commit. I thought about this more today and just confirmed.... It ensures correct behavior if the UI doesn't implement this part of the notify API (and it's not required to). Bartosz Oler agreed that I was right when presented with this case. I'm sorry for the indecision on this.

src/protocols/gg/gg.c file | annotate | diff | comparison | revisions
--- a/src/protocols/gg/gg.c	Wed Nov 30 22:52:44 2005 +0000
+++ b/src/protocols/gg/gg.c	Thu Dec 01 00:36:19 2005 +0000
@@ -1038,6 +1038,16 @@
 				(GaimNotifyCloseCallback)ggp_sr_close_cb,
 				gaim_connection_get_account(gc));
 
+		if (h == NULL) {
+			gaim_debug_error("gg", "ggp_pubdir_reply_handler: "
+					 "Unable to display the search results.\n");
+			gaim_notify_error(gc, NULL,
+					  _("Unable to display the search results."),
+					  NULL);
+			ggp_sr_close_cb(gaim_connection_get_account(gc));
+			return;
+		}
+
 		info->searchresults_window = h;
 	} else {
 		gaim_notify_searchresults_new_rows(gc, results,

mercurial