summaryrefslogtreecommitdiff
path: root/TODO
blob: d258667db4b489acd46d12f9876390da34200316 (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
XML format
----------
- Document the format
- Add attributes to connect signals to their default handlers
  and wrappers to their vfuncs
- Add a way to specify default values
- Add a way to mark init functions like gtk_init
- For argument/return values etc, store only c:identifier in the
  <class/> definition. Requires a link to the references type in the 
  Type node.
- replace gidl/gdk-pixbuf.gidl with something generated from headers,
  verify all information is available, possible to write

Binary format
-------------
- Sort arrays so that we can use bsearch for lookup
- Maybe add hashes to lookup interfaces and functions in interfaces
- Add default values
- Make annotations more flexible (and typed)

Compiler
--------
- Cleanup, maybe turn the nodes in to objects to avoid the horrific
  type switches
- Add thorough error checking

Repository
----------
- Reconsider the current clunky API, maybe an iterator approach is better
- Add thorough error checking
- Use hashes
- Maybe allow populating repositories at runtime
- Think about a system-wide repository mapping namespace ids to
  libraries/metadata files
- Fix extra indirection for out parameters in invoke

Scanner
------
- Property.readable/writable, GLibProperty.blurb/nick
- Constants
- Annotations in external files

General
-------
- More tests
- Check how the format scales to something the size of GTK+

C API
-------
- type SYMBOL is kind of inconvenient; would be nicer to go straight to boxed, etc.
- no way to construct a boxed type
- no way to access struct fields (should have a way to go field -> GArgument)
- types namespaced GI, functions namespaced g_
- no way to get GType given a GIRegisteredTypeInfo