Debian Codegraph @

TODO

raw
1 ----------------------
2 --- IMPORTANT TODO ---
3 ----------------------
4
5 --------------------
6 --- NEEDED TESTS ---
7 --------------------
8 - test:
9 - service method lookups
10 - out-of-order fields in messages (ie if the number isn't ascending)
11 - gaps in numbers: check that the number of ranges is correct
12 - default values
13 - message unpack alloc failures when allocating new slab
14 - message unpack alloc failures when allocating unknown field buffers
15 - packed message corruption.
16 - meta-todo: get a list of all the unpack errors together to check off
17
18 ---------------------
19 --- DOCUMENTATION ---
20 ---------------------
21 Document:
22 - services
23 - check over documentation again
24
25 --------------------------
26 --- LOW PRIORITY STUFF ---
27 --------------------------
28 - support Group (whatever it is)
29 - proper support for extensions
30 - slot for ranges in descriptor
31 - extends is implemented as c-style function
32 whose name is built from the package, the base message type-name
33 and the member. which takes the base message and returns the
34 value, if it is found in "unknown_values".
35 boolean package__extension_member_name__get(Message *message,
36 type *out);
37 void package__extension_member_name__set_raw(type in,
38 ProtobufCUnknownValue *to_init);
39
40 ------------------------------------
41 --- EXTREMELY LOW PRIORITY STUFF ---
42 ------------------------------------
43 - stop using qsort in the code generator: find some c++ish way to do it
44
45 ----------------------------------------------
46 --- ISSUES WE ARE PROBABLY GOING TO IGNORE ---
47 ----------------------------------------------
48 - strings may not contain NULs
49
50 -------------------------
51 --- IDEAS TO CONSIDER ---
52 -------------------------
53
54 - optimization: structures without repeated members could skip
55 the ScannedMember phase
56
57 - optimization: a way to ignore unknown-fields when unpacking
58
59 - optimization: certain functions are not well setup for WORDSIZE==64;
60 especially the int64 routines are inefficient that way.
61 The best might be an internal #define WORDSIZE (sizeof(long)*8)"
62 except w/ a real constant there, one that the preprocessor can use.
63 I think the functions in protobuf-c.c are already tagged.
64
65 - lifetime functions for messages:
66 message__new()
67 return a new message using an allocator with standard allocation policy
68 message__unpack_onto(...)
69 unpack onto an initialized message
70 message__clear(...)
71 clears all allocations, does not free the message itself
72 message__free(...)
73 free the message.
74 [yeah, right: after typing it out, i see it's way too complicated]
75
76 - switching to pure C.
77 - Rewrite the code-generator in C, including the parser.
78 - This would have the huge advantage that we could use ".proto" files
79 directly, instead of having to invoke the compilers.
80 - keep in a separate c file for static linking optimziation purposes
81 - need alignment tests
82 - the CAVEATS should discuss our structure-packing assumptions