Re: Announcement: NoCatSplash 0.93pre2
"Casey T. Deccio" <[email protected]>
| Newsgroups | gmane.network.nocat |
|---|---|
| Organization | Sandia National Laboratories |
| Message-ID | <[email protected]> |
On Thu, 2005-03-03 at 01:52 -0700, Schuyler Erle wrote: > This is a pre-release only because it hasn't been thoroughly tested > yet, and Casey reported a memory leak issue that I want to get fixed. > If you're running NoCatSplash, please try out this pre-release and > post your findings to the list. (Casey, would you be so kind as to > submit a patch against the latest pre-release?) > I've attached a diff of the patch (it includes some corrected contact info and the name of another individual who helped me track down and fix the issue with the original patch to http.c). Basically, all this patch does is add a call to peer_free from inside check_peer_expire, after it is determined that a peer has expired and it has been "denied" from the firewall rules. Since check_peer_expire is called from g_hash_table_foreach_remove, the hash entry will be removed when the function returns true, but the memory won't be properly deallocated. If or when this is configured to use glib-2.0, this issue (and others) may be simplified by creating a hash table with a GDestroyNotify function using g_hash_table_new_full ( http://developer.gnome.org/doc/API/2.0/glib/glib-Hash-Tables.html#g-hash-table-new-full ). Casey _______________________________________________ NoCat mailing list [email protected] http://lists.nocat.net/mailman/listinfo/nocat