Open tmplt opened 2 years ago
All echoes that aren't CSV will have to be done to stderr
.
Some thoughts:
should we just drop the messy frontend implementation is favor of simply talking CSV? Doing so would increase modularity and decrease complexity and code that need to be supported. Then some choice post-processing scripts could be maintained as a starting point.
cargo itm-scope [trace --recovery-flavor [ rtic | zephyr | ... ] | replay ... ] [--out trace.csv]
?
Some
--export-csv
flag should be added to that the resolved trace can easily be piped into whatever post-processing program one can come up with. The CSV will have to be exhaustive.