summaryrefslogtreecommitdiff
path: root/TODO
blob: bd15d44c69038a83cb6fe8f8f2a4d2b740f25bc1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
GNOME 1.4 release (must freeze VERY soon)
===

The release with 1.4 is not going to be super-useful for large-scale
installation administration; it's mostly just going to be a
process-transparent way to store settings, limited to single
workstation config files.

No more source-incompatible API changes are planned for 1.4 at this
time.

API additions
==

* Implement batch gets

Other
==

* Maintain documentation

* Envisioneering

Maybe 1.4
===

* Implement dump/slurp functionality (define XML DTD to represent 
  modifications to the database; augment gconftool to be able to 
  write out the current state of the database in this format, 
  and also apply the changes given in the format)

* Make it so that once the first notification of a change in a GConfChangeSet
  is delivered, the other values will be retrieved by gconf_get() and 
  gconf_client_get(), which means a way to invalidate GConfClient
  cached stuff, and doing the setting of all values in the changeset
  before the notifications.

* Allow various currently-hardcoded items to be set from environment variables
  or a config file ("home" directory to use, timeout lengths, etc. are 
  some candidates).

* "Laptop mode" where GConf avoids touching the disk much

* Implement server-side search (Kind of hard to actually implement 
  on the server, at least in any sort of fast way, and 
  all other gconf-using apps will block while the server is searching,
  without some tricks to let the main loop run sometimes, so, dunno.)

* Implement a way to get the GConfMetaInfo

Future
===

* Berkeley DB backend (note: consider issues surrounding various incompatible 
  versions of DB and historical problems with the upgrade path, cf. 
  RPM and gnome-mime-db)

* Performance tuning

* Document locking issues for backends (backends should perform
  their own locking, handle concurrency, etc.)

* Document which database GConf will write to given multiple 
  writeable databases (i.e. the first one it can write to,
  at the moment, maybe eventually the "database map" will 
  specify which it writes to)

* Implement a way for backends to notify gconfd of changes they detect

* Implement a "database map"; this would be a tree structure (similar
  in implementation to GConfListeners). Rather than storing listeners
  at the tree nodes, store a list of databases in order, and 
  readability/writability of each database. Create a config file
  (perhaps in the GMarkup XML subset from glib 2.0) for configuring
  the database map. Figure out whether this can entirely replace
  the readable/writable methods from the backend vtable.
  It likely replaces the gconf/path configuration file. (Essentially
  the idea is a database path per key/directory, instead of a 
  global database path, giving administrators more flexibility.)

  Also, aliases for paths, and way for apps to install a suggested
  default database alias ("per_display" "per_homedir" etc.)
  (See gconf-list post)

  Details to be figured out.

* GUI admin tool, and GUI user tool (are these the same?)

* Thread support for scalability; may require ORBit thread safety?
  Or a protocol with oneway CORBA methods (client requests a value,
  gconfd calls back when it has the value)

* Fix non-default GConfEngines: this means propagating change
  notifications from them to other engines with the same 
  databases. Or maybe instead we should use the mechanism
  used when the same database is in two gconfds (backend 
  notifies us of changes).

  Suspect that all notification has to come from the backend,
  this is the only way to get sane behavior if _some_ notification
  comes from the backend. Hmm.

* Use a real DTD and a nicer structure for the XML backend format

* The design of the client-server architecture is horked, overcomplicating 
  GConf.idl; the client should remember all its state (listeners, etc.), 
  and when the server disappears (we lose the connection to it), the 
  client resends its state; thus eliminating the saved state file 
  and making things more robust.