-
Idea suggested by Hannu Vuolasaho
-
- return only useful data client side - available in TCP when a gateway to RTU is used - need to add isolated handling of indication/confirmation messages
-
- more coherent - namespace - opaque and smaller context - usual wording
-
A new API will be committed to remove the slave in TCP communication.
-
Original patch by Sisyph (eric-paul).
-
Page 23 of Modbus_Messaging_Implementation_Guide_V1_0b
-
The function was used only one time.
-
The ID used at init time will be the device ID of the caller and the server ID in request functions is the target to reach.
-
The library is now simpler to include (only one <modbus.h>) To avoid confusion MB_VERSION defines have been renamed to LIBMODBUS_VERSION.
-
This reverts commit df0cf7927249954f15e0aa85b02d21b990fa9af9. Use <modbus.h> in tests.
-
This time the change is definitive :)
-
I don't want to put a burden on build system support anymore
-
Requests not to send SIGPIPE on errors on stream oriented sockets when the other end breaks the connection.
-
The previous implementation dereferenced a type-puned pointer, which is illegal according to ANSI C. Some newer versions of GCC will complain about this. This implementation uses "memcpy" to avoid this problem.
-
- the return value is used to pass the message length - remove the hack on exception check in modbus receive - update tests
-
It occurs on first occurence of slave timeout (found with Valgrind 3.3.0). Reported by Henrik Munktell.
-
- simpler wscript_build files - reuse VERSION - add path to config.h
-
Reported by Henrik Munktell.
-
- adjusts various comments on return values - updated tests - added entry to MIGRATION
-
Reported and tested by Yishin Li. Real fix #333455 - IPTOS_LOWDELAY not supported on cygwin Thank to Yishin Li for the better problem description.
-
- check IPTOS_LOWDELAY define in configure - generate a config.h - include the generated config.h in modbus.c
-
- new function modbus_flush - new names and values for error defines - finer recovery on error - merge TOO_MANY_DATA and INVALID_DATA - stop unit-test-master at the first error - FLUSH_OR_RECONNECT_ON_ERROR -> FLUSH_OR_CONNECT_ON_ERROR - more precise tests in unit-test-master
-
- same values (2) - man of socket() says PF_INET (packet family) - AF_INET (address family) for addresses
