Sorry, will not release today. Too many changes have occurred (for good reasons) and we want to ensure that 0.93.3 is rock-solid, so that we may take time to work on new features for 0.94...
We fixed all the memory leaks we were able to spot. Please test extensively. Raphael - (2004.01.18 - RAM) * Changed minimum for "node_connected_timeout" to 120 and the default to 180, since we must account for possible flow-control, both at the node and TCP level. * Try to account for PARQ active upload slots when looking whether we have enough file descriptors to become an UP, but lower the impact of banned fd since we can now steal them when we run out of available ones. * Lowered banned IP timeout to 6 hours, and mark IP as bad on connect timeout. * Fixed default returned value for add_host_to_cache() to FALSE. * (downloads_gui.c) Optimized to do without the costly count_node_children() routine. * Routine gui_update_download_clear() now only records that we need to figure out the sensitivity of "Clear Stopped". However the real change is only performed once per second via gui_update_download_clear_now(). * Must perform downloads_gui_init() before settings_gui_init() due to the use of property callbacks that can trigger GUI code, and because we now cache the results of lookup_widget() in downloads_gui.c. * Periodically call gui_update_download_clear_now() from download_timer(). - (2004.01.18 - Richard) * [Gtk1] Disable "expand all" and "collapse all" when no search is open. * [Gtk1] Fix possible crash if sorting on less then 50 results (Thanks Emile). - (2004.01.17 - RAM) * Allow 6 unanswered "alive pings" in a row instead of 4 for UP connections. * Fixed incorrect list_delete_link_track() implementation causing memory havoc. * (search_gui.c) Removed needless atom_str_get() calls. * Redirect big walloc() calls to malloc_track() under TRACK_MALLOC, instead of g_malloc(), to not lose the source of the leak. * (dmesh.c) Made sure we cleanup the banned mesh by IP at close time. * Added g_string_*() trapping when -DTRACK_MALLOC. * Fixed stupid (recent?) typo preventing proper recording of the user-agent for uploads. * (downloads.c) Fixed memory leak. - (2004.01.17 - Richard) * Cleaned up gtk-missing.h. * [Gtk1] Hopefully fixed sorting when any node in the results was expanded by iterating only over the top-level nodes when sorting. Refactored some of the code in the process and added a couple of FIXMEs for later releases. * [Gtk1] Hopefully fixed a few memory leaks in the download pane probably caused by calling g_list_free() not on the first element of a GList. * [Gtk1] In the search pane two more leaks reported by the leak detection code should be fixed. * [Gtk1] Fixed crash when "expand all" or "collapse all" was pressed and no search was open. Raphael ------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ Gtk-gnutella-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/gtk-gnutella-devel
