Ok so adding that space back did not break the rule again.
I reviewed the logs and found out that some other rules get executed twice. And that is unrelated to the suspected trailing space. That rule is not cron based but trigered by the sensor change and in the debug I can see that there was no change of the sensor state as the living temp is the same for both runs
2017-01-25 20:22:31.094 [DEBUG] [eclipse.smarthome.model.script.TERMO] - ----Termostat rule kicked in!---------
2017-01-25 20:22:31.105 [DEBUG] [eclipse.smarthome.model.script.TERMO] - ----Termostat rule kicked in!---------
2017-01-25 20:22:31.138 [DEBUG] [eclipse.smarthome.model.script.TERMO] - outside_cur_temp:2.43
2017-01-25 20:22:31.143 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living_setpoint:23.0
2017-01-25 20:22:31.149 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living_cur_temp:24.25
2017-01-25 20:22:31.149 [DEBUG] [eclipse.smarthome.model.script.TERMO] - outside_cur_temp:2.43
2017-01-25 20:22:31.156 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living_setpoint:23.0
2017-01-25 20:22:31.162 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living_cur_temp:24.25
2017-01-25 20:22:31.181 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living valve OFF
2017-01-25 20:22:31.185 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living valve OFF
2017-01-25 20:22:31.195 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living temp_deficit raw:-1.25
2017-01-25 20:22:31.197 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living temp_deficit raw:-1.25
2017-01-25 20:22:31.200 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living temp_deficit adjusted:0
2017-01-25 20:22:31.206 [DEBUG] [eclipse.smarthome.model.script.TERMO] - living based new_water_setpoint:23.0