+- a way to get the network proximity (needed by Pastry, at least)
+
- pull method of source diffusion in graspe-slave
- Use a xbt_set for gras_procdata_t->libdata instead of a dict
###
- tcp->incoming_socks
- sock specific tcp (buffsize) inutile
+ sock specific tcp (buffsize) useless
###
### Soon
- gras_datadesc_import_nws?
-- rawsock
- Check that messages don't go on raw socks
- Rename to meassock or whatever to show that they're not intended for
- out of band communication, but for socket operation measurements.
-
- Implement gras_datadesc_cpy to speedup things in the simulator
For now, we mimick closely the RL when on simulator, which is not needed.
(this was easier to do).
gras_datadesc_cpy needs to provide the size of the corresponding messages, so
that we can report it into the simulator.
-- callback on reception ?? (to put functions pointer back in place, etc)
-- category "ignored" should be dropped, since it's not portable (what's its
- size on remote site?). But function's pointer may benefit from it.
- We could change it to an attribute just as the "cycle" one. That way, it
- would get malloced, but not transfered.
+- when a send failed because the socket was closed on the other side,
+ try to reopen it seamlessly. Needs exceptions or another way to
+ differentiate between the several system_error.
+- cache accepted sockets and close the old ones after a while.
+ Depends on the previous item.
+
###
### A bit later
###
+- timeout the send/recv too
- Adaptative timeout
- datadesc_set_cste: give the value by default when receiving.
It's not transfered anymore, which is good for functions pointer.
- virtualize (linux, solaris, SG) & conditions
- binary representation: any type, SNWF (Sender Native Wire Format)
- modules (log control, manage, token ring, bw)
- - cleanups, documentation
[autoconf]
- Check in autoconf that no datatype is bigger than 64, or dynar_map will
- get into trouble...
Check the gcc version on powerpc. We disabled -floop-optimize on powerpc,
but versions above 3.4.0 should be ok.
- The ucontext usability test is too light. It returns success on IRIX, but
- shouldn't since ucontext are broken for us there.
[portability layer]
- Dynar of type, such as ref of type
Mallocators
[Messaging]
Hijack message from a given category to another for a while (to mask
initializations, and more)
Allow each process in simulation to have its own setting
+- a init/exit mecanism for logging appender
+- more logging appenders
[dict]
speed up the cursors, for example using the contexts when available
Error handling in cbps
Regression tests of cbps
-[testsuite]
- use logging, not printf
-
*********
* GRAS1 * Integrer grassouillet a gras; multiplexage XML; module de comm
*********
Check short ***
Check struct { struct { int a } b; }
- Factorise code in union/struct field adding
+ Factorize code in union/struct field adding
[Other]
- Allow [homogeneous] dynar and dico to be sent
+ Allow [homogeneous] dico to be sent
Make GRAS thread safe by mutexing what needs to be