diff src/main.c @ 10245:c143a3fac58d

[gaim-migrate @ 11385] Binary relocation, step one. I had a fairly long commit message, and cvs ate it and said Read from remote host cvs.sourceforge.net: Connection reset by peer I'm displeased. This is just method one, method two to follow tomorrow after I add a way to register a fallback and try to push it upstream. That way I don't have to put method 2 inside prefix.c. As for as core/ui split goes, they can either each have their own copy after the divorce is final, or the UI can use the core's. It'll work either way, since #1 finds location of the caller, and #2 doesn't work on libraries anyway. That's one advantage I forgot to mention, btw, that #1 will let a library find its own location. So, I'm sure something isn't quite right and someone will want to fix it. So they can fix it while i implement phase two. Which won't take long at all, but the autopackage guy is in bed, and I should be too. committer: Tailor Script <tailor@pidgin.im>
author Tim Ringenbach <marv@pidgin.im>
date Tue, 23 Nov 2004 05:53:59 +0000
parents 952fb08c60b0
children 5ce83aad6869
line wrap: on
line diff
--- a/src/main.c	Tue Nov 23 02:49:09 2004 +0000
+++ b/src/main.c	Tue Nov 23 05:53:59 2004 +0000
@@ -821,7 +821,7 @@
 		abort();
 	}
 
-	plugin_search_paths[0] = LIBDIR;
+	plugin_search_paths[0] = g_strdup(LIBDIR);
 	plugin_search_paths[1] = gaim_user_dir();
 	plugin_search_paths[2] = g_build_filename(gaim_user_dir(), "plugins", NULL);
 
@@ -829,6 +829,7 @@
 								  sizeof(*plugin_search_paths),
 								  plugin_search_paths);
 
+	g_free(plugin_search_paths[0]);
 	g_free(plugin_search_paths[2]);
 
 	gaim_plugins_probe(NULL);