6 - the README, FAQ, and probably other places talk about 1.0 and ABI guarantees,
7 all these places need updating
9 - the spec should say "the protocol will not be changed incompatibly after Month DD, YYYY"
11 - the README documents the configure flags, should check this for being in sync with reality
13 Important for 1.0 GLib Bindings
16 - Test point-to-point mode
18 - Add support for getting sender
20 - format_version in the object info doesn't look like it's handled correctly. The creator
21 of the object info should specify some fixed number per struct version; the library
22 should handle only specific numbers it knows about. There's no assumption that all
23 numbers >= the given one are compatible. The idea is that new versions of the lib
24 can offer totally different object info structs, but old versions
27 Important for 1.0 Python bindings
32 - Fix removing of signals from the match tree
34 - Fix refcounting and userdata lifecycles
36 - Write a generic mainloop
41 - protocol version in each message is pretty silly
46 - revamp dbus-launch a bit,
47 see http://lists.freedesktop.org/archives/dbus/2006-October/005906.html
50 - clean up the creds issue on *BSD's in dbus/dbus-sysdeps-unix.c.
51 They should work as is but we need to rearange it to make it
52 clearer which method is being used. configure.in should
53 be fixed up to make that decition.
55 - _dbus_connection_unref_unlocked() is essentially always broken because
56 the connection finalizer calls non-unlocked functions. One fix is to make
57 the finalizer run with the lock held, but since it calls out to the app that may
58 be pretty broken. More likely all the uses of unref_unlocked are just wrong.
60 - if the GUID is obtained only during authentication, not in the address,
61 we could still share the connection
63 - Allow a dbus_g_proxy_to_string()/g_object_to_string() that
64 would convert the proxy to an "IOR" and dbus_g_proxy_from_string()
65 that would decode; using these, dbus-glib users could avoid
66 DBusConnection entirely. Of course the same applies to other kinds
67 of binding. This would use dbus_connection_open()'s connection-sharing
68 feature to avoid massive proliferation of connections.
70 - DBusWatchList/TimeoutList duplicate a lot of code, as do
71 protected_change_watch/protected_change_timeout in dbus-connection.c
72 and dbus-server.c. This could all be mopped up, cut-and-paste
73 fixed, code size reduced.
75 - change .service files to allow Names=list in addition to Name=string
77 - The message bus internal code still says "service" for
78 "name", "base service" for "unique name", "activate" for
79 "start"; would be nice to clean up.
81 - Property list feature on message bus (list of properties associated
82 with a connection). May also include message matching rules
83 that involve the properties of the source or destination
86 - Disconnecting the remote end on invalid UTF-8 is probably not a good
87 idea. The definition of "valid" is slightly fuzzy. I think it might
88 be better to just silently "fix" the UTF-8, or perhaps return an error.
90 - build and install the Doxygen manual in Makefile when --enable-docs
92 - if you send the same message to multiple connections, the serial number
93 will only be right for one of them. Probably need to just write() the serial
94 number, rather than putting it in the DBusMessage, or something.
96 - perhaps the bus driver should have properties that reflect attributes
97 of the session, such as hostname, architecture, operating system,
98 etc. Could be useful for code that wants to special-case behavior
99 for a particular host or class of hosts, for example.
101 - currently the security policy stuff for messages to/from
102 the bus driver is kind of strange; basically it's hardcoded that
103 you can always talk to the driver, but the default config file
104 has rules for it anyway, or something. it's conceptually
105 screwy at the moment.
107 - when making a method call, if the call serial were globally unique,
108 we could forward the call serial along with any method calls made
109 as a result of the first method call, and allow reentrancy that was
110 strictly part of the call stack of said method call. But I don't
111 really see how to do this without making the user pass around the
112 call serial to all method calls all the time, or disallowing
115 If done post 1.0 will probably be an optional/ugly-API type
118 - I don't want to introduce DBusObject, but refcounting and object
119 data could still be factored out into an internal "base class"
122 - Keep convenience wrappers in sync with bus methods
124 - document the auth protocol as a set of states and transitions, and
125 then reimplement it in those terms
127 - recursive dispatch, see dbus_connection_dispatch()
129 - do we need per-display activation; if so I'd like to do this by setting a
130 "display ID" property on screen 0, with a GUID, and keying activation by
131 said GUID. Otherwise you get all kinds of unrobust
132 string/hostname-based mess. per-screen is then done by appending screen number
133 to the display. If displays have a deterministic ID like this, you can
134 do per-display by simply including GUID in the service name.
136 - optimization and profiling!
138 - Match rules aren't in the spec (probably a lot of methods on the bus
141 - the "break loader" and valid/invalid message tests are all disabled;
142 they need to be fixed and re-enabled with the new message args stuff.
143 I think I want to drop the .message files thing and just have code
144 that generates messages, more like the tests for
145 dbus-marshal-recursive.c (this is mostly done now, just needs some
148 - just before 1.0, try a HAVE_INT64=0 build and be sure it runs
150 - Windows port needs recursive mutexes
155 - look into supporting the concept of a "connection" generically
156 (what does this TODO item mean?)
158 - test/name-test should be named test/with-bus or something like that