xref: /trunk/main/vcl/unx/gtk/a11y/TODO (revision cdf0e10c)
1 cws 'atkbridge'
2 #Issue number: i#47890#
3 Submitted by: mmeeks
4 
5 Hacked up prototype of atk bridge
6 
7 
8 Serious problems
9     + Threading/locking:
10 	+ incoming CORBA calls & the GDK lock
11 		+ how are these being processed & on what thread ?
12 		+ are we holding the GDK_THREADS lock ?
13 		+ can we even do that ?
14 		+ is it really necessary to be thread safe ?
15 		+ how does this work in combination with the (unsafe) GAIL code ?
16 		+ what should incoming CORBA calls be doing ?
17 			+ esp. since we can't tell if they're coming from
18 			  in-proc or not either [ though this is unlikely ]
19 
20 
21 Test:
22 	+ in-line text editing, does the TEXT_CHANGED signal get it right,
23 	+ why not copy/paste/delete etc. ?
24 		+ check vs. writer & other bits ...
25 	+ AtkSelection
26 	+ AtkHyper*
27 
28 * At-poke
29 	+ implement non-gui mode - for to-console event logging
30 	+ logging
31 		+ more detail from remaining events
32 	+ add a Tree navigation thing instead (?)
33 		+ poke a sub-child (?)
34 		+ embed a tree widget inside the tree view ?
35 	+ AtkHyperText testing (?)
36 
37 
38 Known bugs:
39 	+ AtkText
40 		+ selection interface - multiple selections ?
41 		+ word boundary issues
42 			+ copy AccessibleTextImpl.java's getAfterIndex eg.
43 	+ the 'getFoo' methods need to use UNO_QUERY_THROW &
44 	  throw an exception to avoid null pointer dereferences.
45 	+ AtkAttributeSet (etc.)
46 	+ AtkEditableText
47 	+ finish/test AtkTable
48 	+ HyperLink 'link_activated', HyperText 'link_selected' (?)
49 	+ tooltips create new toplevels with broken roles.
50