On 05/07/2017 07:09 AM, suavesteve wrote: > I've come across this earlier and had created a local patch, > > Tested with both v3.0.6 and v3.1.4 (Latest Stable release)
Thanks Philip/suavesteve for posting that! It's strange to me that the libmodbus folks went from using a struct timeval for their timeout to passing the fields individually, but sure enough they did. According to http://libmodbus.org, 3.0 is stable and 3.1 is still listed as unstable, even though it's been around for years and has had several stable-looking "releases". Some feedback on your patch: You pass timeout.tv_sec for both the seconds and the microseconds, I think you meant to pass timeout.tv_usec for the third argument. Details on that here: http://libmodbus.org/docs/v3.1.4/modbus_set_response_timeout.html Instead of taking the pointer of timeout and dereferencing it ("(uint32_t) (&timeout)->tv_sec"), it's simpler and clearer to just use the field directly ("(uint32_t)timeout.tv_sec"). There's a bug in the #if you use for version detection, think about (for example) a hypothetical future version 4.0. Use LIBMODBUS_VERSION_HEX instead of LIBMODBUS_VERSION_MAJOR and LIBMODBUS_VERSION_MINOR. Instead of this: + #if LIBMODBUS_VERSION_MAJOR >= 3 && LIBMODBUS_VERSION_MINOR >= 1 Use this (untested): + #if LIBMODBUS_VERSION_HEX >= 0x03010000 If you make those changes, and sign off on your commit (see http://linuxcnc.org/docs/devel/html/code/contributing-to-linuxcnc.html#_signed_off_by_policy), then I'll gladly apply your patch in LinuxCNC 2.7. -- Sebastian Kuzminsky ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
