Calling rcl_parse_arguments

[jimmy-mcelwain]

Is there a particular reason that we are calling rcl_parse_arguments directly in CommunicationExecutor.c? Why don't we just pass in the faux_argc and faux_argv to rclc_support_init_with_options? rcl_parse_arguments is (eventually) called there, and the remap rules still work properly. Is it just so that it doesn't enter the error loop forever if the user has invalid remap rules? Is there some other reason? If so, I understand, but I wasn't sure if this is an oversight or a intentional decision.

[gavanderhoorn]

The main reason -- but I'd have to look up my notes from way back when that was implemented -- most likely is we wanted to use the smallest part of RCL and RCLC which would get us remap functionality.

So keep the existing control flow, then add the minimum needed to implement remapping support.

We could refactor and call

motoros2/src/CommunicationExecutor.c

Lines 140 to 141 in 0f0a510

	rcl_ret_t ret = rclc_support_init_with_options(&g_microRosNodeInfo.support,
	0, NULL, &g_microRosNodeInfo.initOptions, &g_motoros2_Allocator);

later, after having created the faux argv and argc. Functionally, it wouldn't really change anything IIUC.

Priority of other issues/enhancement requests was just always higher.

PS: please always link to the code you're referring to. Makes discussing it easier :)