Debian Codegraph @

docs/random/ds/0.9-suggested-changes

raw
1
2
3 API:
4
5 - MAKE A DOCUMENT THAT LISTS ALL API CHANGES AND HOW TO DEAL WITH THEM
6
7 - replace object/structure set/get macros with actual functions.
8
9 - reorganize headers (split app headers vs plugin headers maybe)
10 while at it also make sure to only allow including of the main headers like
11 glib, so we can freely shuffle stuff around
12
13 - make GstPadLinkReturn internal (to either plugins+core or just core)
14 and return gboolean to apps.
15
16 - rewrite GstIndex (it's gross)
17
18 - gst_init() et al. need to work correctly when called multiple times
19 and from libraries, etc.
20
21 - gst_pad_get_pad_template_caps -> gst_pad_get_template_caps()
22
23 - make templates mandatory for pads.
24
25 - force pads to use a pad template from the element they belong to.
26
27 - s/PadTemplate/PadFactory/ to describe better what it is and be more in line
28 with the rest of the API?
29
30 - add ways for autopluggers to manage negotiation of their children.
31 This would allow autopluggers to no longer need helper identities.
32 (Though the main reason for identities atm is to be able to connect without
33 plugging - Company)
34
35 - read/write locks on buffers
36
37 - be able to send events to unlinked pads (bug #114442)
38
39 - caps should have a flag on fields to indicate that the field is optional
40
41 - fakesrc handoff should be changed to return a GstData * or carry a
42 GstData **, wo the user can provide own buffers and events.
43
44 - don't install gstmarshal.h
45
46 - make sure GstClockTime is used wherever it should (e.g. gstplay)
47
48 - make gst_bin_get_list virtual so subclasses can override it (e.g.
49 gst-play which contains an internal thread not inside the bin that is
50 gstplay)
51 - make gst_bin_get_by_interface search on the list provided by this get_list
52 so that you can search a GstPlay for an element implementing the interface
53 (right now gstplay has added a gst_play_get_by_interface)
54
55 - make multichannel properties on audio caps manditory
56
57 - talk to vektor about various wierdness with NTSC video (combinations of
58 24 fps pulled down to 60 and straight video, etc) and make sure our
59 video stream descriptions make sense.
60
61 - do an audit to remove GtkObject-isms from gtk-1.2
62
63 caps:
64
65 (Company:)
66
67 "An idea I had once was to force registration of mimetypes and then
68 only allow caps with a registered mimetype - like it's done now with caps.
69 You could then add functions to those registrations like
70 - gchar *get_human_readable_string (caps);
71 should be kinda selfexplanatory. Would return stuff like "16bit signed
72 integer audio", "MPEG2 video" or "video format in RGB colorspace" (for
73 unfixed RGB)
74 - GstCaps *default_fixate (const GstCaps *caps);
75 A default fixation function, so we don't fixate to minimum integer
76 anymore but have the ability to fixate to 320x240 for video or 44100 for
77 audio rates
78 - gboolean validate (const GstCaps *caps);
79 check if the given caps are ok for this mimetype - debugging only function
80 I'd _really_ like to have - especially during gst_register.
81
82 But that's 0.10 material."
83
84 (Thaytan)
85 - Add rowstride as a property so that video frames don't have to have
86 rowstride == width. For example, gdkpixbufdec would prefer to pad
87 rowstrides to multiples of 4 bytes.
88
89 negotiation:
90
91 - autopluggers would be easier to write if there was a core method
92 to do what plugidentities do.
93
94 - there should be a way to hook into negotiation and be able to do something if
95 negotiation fails. Autopluggers would need that for exchanging elements when
96 they can't link.
97
98 bugs with interesting info:
99
100 XML descriptions of plugin information:
101 http://bugzilla.gnome.org/show_bug.cgi?id=114216
102
103
104 object hierarchy:
105 - a method for elements to know when downstream elements are ignoring
106 the data stream. This would allow automatic shutoff of pipelines
107 whose processing is being dumped.
108
109 - object properties:
110 - all bitrates should be bits per second, since they're specified
111 as integers
112 - use device for nodes (not location; dvdreadsrc, dvdnavsrc, cdparanoia).
113 Also figure out if we want to continue separating uri/location.
114
115 - categorization of plugins:
116 - categories were intended from the start as a list of keywords, instead
117 of an implicit tree.
118 - this makes a lot more sense and probably fixes our constant discussion
119 about categorizing
120 - make this explicit by having a list of keywords instead of a fixed string
121 with slashes
122 - fix the editor/el browser to display this nicely