1 - Probably no point in a version number in the daemon name
2 (s/dbus-daemon-1/dbus-daemon/)
4 - How we will handle DCOP needs sorting out. Among other things, we
5 need to check that service and service-ownership semantics map to DCOP
8 - Activation needs some careful additional thinking-through.
10 - Property list feature on message bus (list of properties associated
11 with a connection). May also include message matching rules
12 that involve the properties of the source or destination
15 - Automatic service activation, should probably be done through a message flag.
17 - Disconnecting the remote end on invalid UTF-8 is probably not a good
18 idea. The definitiion of "valid" is slightly fuzzy. I think it might
19 be better to just silently "fix" the UTF-8, or perhaps return an error.
21 Owen says we should only validate the UTF-8 on dbus_message_get_string()
22 (changing get_string to have an error return, and allowing a type error
25 - The convenience functions in dbus-bus.h should perhaps have
26 the signatures that they would have if they were autogenerated
27 stubs. e.g. the acquire service function. We should also evaluate
28 which of these functions to include, in light of the fact that
29 GLib/Qt native stubs will probably also exist.
31 - assorted _-prefixed symbols in libdbus aren't actually used by
32 libdbus, only by the message bus. These bloat up the library
33 size. Not sure how to fix, really.
35 - build and install the Doxygen manual in Makefile when --enable-docs
37 - if you send the same message to multiple connections, the serial number
38 will only be right for one of them. Probably need to just write() the serial
39 number, rather than putting it in the DBusMessage, or something.
41 - perhaps the bus driver should have properties that reflect attributes
42 of the session, such as hostname, architecture, operating system,
43 etc. Could be useful for code that wants to special-case behavior
44 for a particular host or class of hosts, for example.
46 - currently the security policy stuff for messages to/from
47 the bus driver is kind of strange; basically it's hardcoded that
48 you can always talk to the driver, but the default config file
49 has rules for it anyway, or something. it's conceptually
52 - when making a method call, if the call serial were globally unique,
53 we could forward the call serial along with any method calls made
54 as a result of the first method call, and allow reentrancy that was
55 strictly part of the call stack of said method call. But I don't
56 really see how to do this without making the user pass around the
57 call serial to all method calls all the time, or disallowing
60 - the invalid messages in the test suite are all useless because
61 they are invalid for the wrong reasons due to protocol changes.
62 (Consider extending test suite to validate that they are
63 invalid for right reason, e.g. an "INVALID_ERROR Foo" line
66 - I don't want to introduce DBusObject, but refcounting and object
67 data could still be factored out into an internal "base class"
70 - modify the auth protocol to also support other initial-handshake
71 type of information such as protocol version
73 - document the auth protocol as a set of states and transitions, and
74 then reimplement it in those terms
76 - dbus_gproxy or dbus_g_proxy?
78 - add dbus_message_has_path(), maybe has_member/interface
80 - re_align_field_recurse() in dbus-message.c is broken because it
81 crashes on some types of header field values. security problem.
83 - modify the wire protocol to keep the args signature separate
84 from the args themselves. Make the name of TYPE_CUSTOM part
85 of the type signature, rather than part of the value.
86 Then you have the full typecheck in a single string.
87 See http://freedesktop.org/pipermail/dbus/2004-June/001169.html
89 - dbus_message_iter_init_array_iterator has "iter" and "iterator"
90 in the same function name
92 - the GLib bindings varargs take DBUS_TYPE_WHATEVER and
93 return stuff allocated with dbus_malloc(); should this
94 be made more "G" at some expense in code duplication?
95 You also still have to use some D-BUS functions such as
96 dbus_message_get_args() which takes a DBusError.
97 Probably we need to either fully encapsulate and hide
98 dbus/dbus.h, or encapsulate it slightly less e.g. no
99 GError. Or maybe it's as simple as "never return dbus_malloc()
100 memory" and just fully encapsulate the get_args() type of
103 - need to define bus behavior if you send a message to
104 yourself; is it an error, or allowed? If allowed,
105 we need to have a test for it in the test suite.
107 - array lengths should probably be returned as size_t rather than int
108 (though they are kind of a pita to pass in as size_t with the
109 varargs, so maybe not - what does glib do with g_object_get()?)
111 - recursive dispatch, see dbus_connection_dispatch()
113 - Perhaps the auth protocol should be able to negotiate a protocol
114 version to the least-common-denominator between client and server?
115 Though in practice ever using this feature would be pretty tough,
116 since protocol probably modifies the API. But we could have it there
119 - STRING_OR_NIL is wrong, doesn't work in C++ etc. ; should not have done that.
120 Use empty string or special string values or separate functions/signals
123 - For recursive types, one approach is that "structs" are done as parens,
124 so e.g. s(ii) is a string and struct { int; int; } etc. Type codes
125 then all have to be done as strings not single ints.
126 We could also put the type signature for the message body in a header field.
127 An "any" type has the type string included in the value.