Corrected #6 : the problem was caused with diagnostics call (WHO = 1001, 1004 or 1013) for which responses are quite long...
The regex used to filter could then find a part (within the command) which looked like light updates.
ex: #100103014000##
where the end of it (14000##) is similar to a light/switch change, which it is not...
Changed to ensure the regexp is tested at the start of the string and not in the middle of it.
Corrected #6 : the problem was caused with diagnostics call (WHO = 1001, 1004 or 1013) for which responses are quite long... The regex used to filter could then find a part (within the command) which looked like light updates. ex: #100103014000## where the end of it (14000##) is similar to a light/switch change, which it is not... Changed to ensure the regexp is tested at the start of the string and not in the middle of it.