1 2004-07-05 Martin Quinson
2 - YEAH! GRAS/SG and GRAS/RL are both able to run the ping example !
4 - Plug a whole bunch of memleaks
5 - each process now have to call gras_{init,exit}. One day, their log
6 settings will be separated
7 - Continue the code factorisation between SG, RL and common in Transport.
9 2004-07-04 Martin Quinson
11 - Redistribution between SG and RL.
12 We wanna have to accept in SG, so move accepted related parts of RL in
13 the common part. (more precisely, the dynar of all known sockets is no
14 more a static in transport.c, but part of the process_data)
17 - Bug fix: Do call gras_process_init from gras_init (wasnt called in RL).
19 2004-07-03 Martin Quinson
20 - Create a new log channel tbx containing dict, set, log, dynar (to shut
21 them all up in one shot)
23 - Fix the ugly case of reference to dynamic array.
24 - New (semi-public) function gras_datadesc_size to allow the messaging
25 layer to malloc the needed space for the buffer.
27 - gras_socket_close now expect the socket to close (and not its address to
28 put NULL in it after it). This is because the socket passed to handlers
29 is one of their argument (=> not writable).
31 - propagate the interface cleanup from last week in datadesc, ie remove a
32 superfluous level of indirection. User pass adress of variable
33 containing data (both when sending and receiving), and not of a variable
34 being a pointer to the data. Let's say that I like it better ;)
35 The price for that is constructs like "int msg=*(int*)payload" in
36 handlers, but it's a fine price, IMHO.
38 - Let it work in RL (yuhu)
40 2004-06-21 Martin Quinson
42 - porting SG plugin and SG select to new standards (works almost).
43 - plug memleaks and fix bugs around.
46 - cleanup the prototype of data recv and force users to specify when they
47 want to handle references to objects. Test case working even for cycles.
48 - plug memleaks. Valgrind is perfectly ok with this.
50 2004-06-12 Martin Quinson
52 - cleanup the separation between plugin and main code in plugin creation
54 2004-06-11 Martin Quinson
56 - Reput hook for raw sockets, needed for BW experiments
57 - kill a few lines of dead code
58 [Data description] Interface cleanup
59 - gras_datadesc_by_name returns the searched type or NULL.
60 That way, no variable is needed to use a type desc once, which makes
62 - gras_datadesc_declare_[struct|union]_append_name is removed. The last
63 two parameters were strings (field name, type name), leading to
65 [Dicos] Interface cleanup
66 - gras_dico_retrieve -> gras_dico_get ; gras_dico_insert -> gras_dico_set
67 This is consistant with the dynar API.
69 2004-04-21 Martin Quinson
71 - Porting to new standards.
74 There is no bag anymore, no need to take extra provision to mask the
76 Better splitup of functions between files create/exchange/convert.
77 This is still a bit artificial since convert and receive are so
78 interleaved, but anyway.
80 - add a queued message list to procdata (the ones not matching criteria
82 - factorize some more code between SG and RL wrt procdata
84 - use gras_exit in example to track memleaks
85 - get rid of gs_example now that GS is properly integrated into gras
86 - update run_test to integrate the lastest tests (datadesc)
88 - rename WARNINGn macros to WARNn since it prooved error-prone
90 2004-04-19 Martin Quinson
92 - register init/exit functions within gras module mecanism
93 - send/receive function.
94 Convertion is not implemented, but short-cutted if not needed.
95 struct/array elements are sent one by one (instead of block-wise), but
96 nobody really cares (yet). Get a prototype before optimizing.
97 - tests (using a file socket) for DD send/receive on:
98 - base types: int, float
99 - array: fixed size, string (ie ref to dynamic string)
100 - structure: homogeneous, heterogeneous
101 - chained list, graph with cycle
102 Believe it or not, valgrind is not too unhappy with the results. The
103 cycle happily segfaults, but the others are ok. And I'm sick of pointers
107 - Bugfix when using a filename explicitely (instead of '-')
109 2004-04-09 Martin Quinson
111 - factorize more code between RL and SG in socket creation
112 - Complete the implementation and tests of:
114 o file (only in RL, and mainly for debugging)
116 I lost 3 days to design a portable address resolver, and then decided
117 that the prototype mainly have to run on my box.
118 Addressing portability too early may be like optimizing too early :-/
120 - use gras_init in the Tests instead of the crappy parse_log_opt
121 (the latter function is removed)
122 [Conditional execution]
123 - New functions: gras_if_RL/gras_if_SG (basic support for this)
124 [Code reorganisation]
125 - Get rid of libgrasutils.a since it makes more trouble than it solves.
126 Build examples against the RL library, since there is no way to disable
127 its creation for now.