Tomoya Fujita

as far as i know, we do not have it currently. i think that we only have two options to represent the time. (both are not human readable though.) https://github.com/ros2/rcutils/blob/c1623e7db620bf051cb539e3d4eb7029bb554bb2/src/logging.c#L701-L702...

I am okay with either environment variable or adding tokens. > If I were the user, We Sony (including you) are actually user 😄

> There are already a bunch of them, they are not well documented (though I'm working on that), and adding more just makes it harder to explain. agree.

@wjwwood @ivanpauno CC: @eboasson would you give us some feedback what/how you feel about this feature extension? pointing you, because i usually have communication with you these layers, such as...

@wjwwood @dirk-thomas friendly ping, but not urgent.

@ivanpauno really appreciate for checking on this and advice that is helpful. > I would really wait for the design approval before having a full implementation. i was thinking that...

> Adding an optional feature sounds fine to me, but you're proposing making use of cft in some built-in use cases (parameter events and actions). It does not have to...

@gbiggs thanks for sharing you thoughts! i do agree on that.

> It could be implemented in rcl instead. i was thinking about that, it will be better.

we are doing some test about efficiency of ContentFitleredTopic based on RTI Connext with scenarios. - Confirm filtering happens either on writer or reader side. (Network transmission if that happens...