rpms/setroubleshoot/devel setroubleshoot.spec,1.59,1.60

[fedora-cvs-commits at redhat.com]

Author: jdennis

Update of /cvs/dist/rpms/setroubleshoot/devel
In directory cvs.devel.redhat.com:/tmp/cvs-serv21316

Modified Files:
	setroubleshoot.spec 
Log Message:
update to 1.6



Index: setroubleshoot.spec
===================================================================
RCS file: /cvs/dist/rpms/setroubleshoot/devel/setroubleshoot.spec,v
retrieving revision 1.59
retrieving revision 1.60
diff -u -r1.59 -r1.60
--- setroubleshoot.spec	20 Nov 2006 21:23:33 -0000	1.59
+++ setroubleshoot.spec	20 Nov 2006 21:25:25 -0000	1.60
@@ -105,6 +105,102 @@
 %{_datadir}/dbus-1/services/sealert.service
 
 %changelog
+* Mon Nov 20 2006 John Dennis <jdennis at redhat.com> - 1.6-1
+	* logfile scanning finally seems to work connected to browser
+	
+	* Additional Information section of report now includes line
+	number information (if alert was generated from logfile)
+	
+	* replace database update_callback() with notify interface, a more
+	generic solution more easily shared between components
+
+	* object implementing rpc method is now explicitly attached via
+	connect_rpc_interface() instead of walking the MRO chain with
+	magic exclusions. explicitly connecting is more flexible and
+	robust (no getting the wrong object by mistake)
+	
+	* fix handling of return args in local rpc case
+
+	* fix signal connections between audit and logfile
+
+	* split databae and database_properties for audit and logfile
+
+	* fix initial connection state
+
+	* fix lookup_local_id 
+
+	* add more support for updating single rows in browser list view
+	rather than reloading everything
+
+	* correctly handle moving selection after a delete, lots of little
+	selection fixes.
+
+	* improve handling of ListStore model, nuke the kludgy pre-extended rows
+	* signatures_updated() now works
+	
+	* fix bug #214218, sealert aborts if lang is not english, also
+	improve error handling, the actual error was not being trapped and
+	instead a subsequent error induced as a consequence was being
+	reported, which was a red herring, also fix redundant variable
+	which was used to hold the broswer window/widget instance.
+
+	* modify how changes to the database are propagated, database
+	object now takes an update callback. Now only the database is
+	responsible for reporting changes, previously it had been the
+	caller who modified the database and then had to also know what
+	type of update to signal. Modify the 'signatures_updated' family
+	of functions and signals to pass a upated type (add,delete,modify)
+	and an item identifier, currently the local id.
+
+	* signals to display the browser window now take an optional
+	argument to indicate what it should be displaying. This was added
+	for the case where the browser was visiting a logfile but the user
+	clicked on the alert notification icon, the browser should in this
+	instance always display the current alert, the user can use the
+	view menu to switch back to the logfile view after having viewed
+	the current alert.
+
+	* add timeout's to status messages
+
+	* tighten up the concept of 'visiting' in the browser, both in
+	data structure and methods, as well as status messages.
+
+	* make sure when the 'mark_seen' event triggers in the future it
+	is bound the database that generated it, not the currently viewed
+	database. 
+
+	* fix some debug messages which were not inside 'if debug'
+
+	* move the signatures_updated signal to the database family of
+	objects. 
+
+	* modify the RPC call mechanism so that it is possible to call a
+	local object through the RPC API without it connecting to a remote
+	object. Thus users of the RPC API can be ignorant if the object
+	they are bound to is local or remote.
+
+	* split the RPC interface for the server into server specific
+	entry points and a new database interface. The new database
+	interface supports the browser binding to either a remote database
+	or a local instance created for logfile scanning.
+
+	* fix how timestamps are assigned, we used to just timestamp an
+	alert when it arrived from the audit system but now with logfile
+	scanning timestamps are embedded in the logfile messages so in
+	order to display the correct time an event occurred we needed to
+	pass the timestamp read in the logfile through the pipeline.
+
+	* create a thread in the browser for logfile scanning, perform the
+	analysis in the thread with the browser GUI bound to the threads
+	operation, display the scanning progress in the progress bar.
+
+	* introduce the notion in the broswer of "visiting" a database, at
+	the moment one is either vising the audit database or the last
+	scanned logfile, added "View" menu and menu items to view the
+	audit database or the logfile. When visiting have the browser
+	connect to different signals. The visit mechanism still needs
+	some work.
+
 * Wed Nov 8 2006 Dan Walsh <dwalsh at redhat.com> - 1.5-1
 - Speed up startup of service