I've been trying some of my code that was developed on Boxturtle on
C-turtle, and I noticed something today: I have a message that
contains a field of type geometry_msgs/Vector3Stamped. I have bagfiles
recorded from when my code was built on boxturtle that show the
header.seq field within the Vector3Stamped incrementing on each
publish, though my code does not explicitly set this field.

With the same code built under C-turtle, the Vector3Stamped header.seq
fields are always zero. Note that the seq field within the 'toplevel'
header field *does* increment as before, and as before this field is
also not touched explicitly by my code.

The .msg file is like so:
Header header # <-- the seq field in here increments in boxturtle and cturtle
geometry_msgs/Vector3Stamped latest_waypoint # <-- the header.seq
field in here only incremented in boxturtle

It's not a hard fix to change my code to increment the 'inner'
header.seq on my own (and in fact I might just change downstream
behaviour to look at the 'outer' one anyway) but I'm curious if this
was an intended change. Also, when seq does get incremented
automatically, how does this happen? When the message is published?

Thanks,
Pat