Instead of adding a Bedrock specific column to htop's source requiring someone maintain a fork, perhaps attempt to upstream a feature to htop to support custom columns that can call out to some user-set shell command to populate. Maybe set some envvar, e.g. $HTOPPID, as the PID for the corresponding shell command to process. You could include example use cases such as
Code:
dpkg -S $(readlink /proc/$HTOPPID/exe) | cut -d: -f1
Code:
pacman -Qo $(readlink /proc/$HTOPPID/exe) | awk '{print$(NF-1)}
etc to show which package corresponds to the given process for various distros to sell the feature without appealing to the relatively unknown Bedrock. There are plenty of other use cases, such as finding a corresponding Docker image, number of open file descriptors, et al.
The Bedrock community can then set `brl which $HTOPPID` without having to maintain a fork.
Other-Software-Feature-Dreaming: htop, and what
