Make sure the player is in the vehicle before running rosPubMsg

[tckarenchiang]

This is to fix #25

[gavanderhoorn]

This will avoid the nil related error, but should self.rosPubMsg also not be reset to false in this case?

Otherwise I believe ModROS:update(..) will try to still publish data, right?

Here:

FS19_modROS/modROS.lua

Lines 104 to 119 in cf452a6

	function ModROS:update(dt)
	-- self.dt = self.dt + dt
	-- create TFMessage object
	self.tf_msg = tf2_msgs_TFMessage:init()
	if self.doPubMsg then
	-- avoid writing to the pipe if it isn't actually open
	if self.file_pipe then
	self:publish_sim_time_func()
	self:publish_veh_func()
	self:publish_laser_scan_func()
	self:publish_imu_func()
	self:publish_tf()
	end
	end

[gavanderhoorn]

And in that case we should probably also guard the opening of the pipe file, otherwise that will be attempted multiple times.

[tckarenchiang]

But 0f5e401 also avoids publishing data if the player is not in a vehicle

[gavanderhoorn]

But why even continue with rosPubMsg(..) if it's clear the player is not in a vehicle?

By adding more conditionals everywhere it looks like we're just complicating things which could be kept simple.

[tckarenchiang]

because initially I wanted to control the inactive vehicles as well, as mentioned in #3. but yes, let's fix the current issue for now

[gavanderhoorn]

So are you planning to change anything, or should we merge this version?

[tckarenchiang]

yes, we should just merge this version for now. Even though it's not a proper fix, it does solve the issue I believe.

[gavanderhoorn]

For future development it might be best to decouple "being in a vehicle" from initialising the publishing infrastructure.

That would seem to be necessary for #3 anyway.

[gavanderhoorn]

Ok, let's address the immediate issue.

[gavanderhoorn]

@tckarenchiang: this will need a rebase I believe.

[tckarenchiang]

@tckarenchiang: this will need a rebase I believe.

[tckarenchiang]

sorry I accidentally clicked close with comment