I'm super happy with inspircd-docker. It's allowed me to get an IRC server running locally really quickly. I'm not super familiar with Docker, but I was able to trivially update some conf/*.conf files so that I could have a lot more local connections while I develop my software.

However, when I wanted an oper to be able to use the sajoin command I edited conf/opers.sh and that didn't work. I wasn't too surprise, because there was a warning in that file about not editing it after a build. After (way too much?) futzing around, I think I understand what is happening, but I can't tell if it's a deliberate design decision or just the fallout of incremental development.

AFAICT (and I may be wrong), because conf/inspircd.conf has lines like:

<include executable="conf/opers.sh">

that pulls opers.sh in from a directory that is not overwritten when people use the -v option of docker to create a bind mount. However, the *.conf files are overwritten.

Locally I've created a consistent_configuration branch that uses inspircd/conf instead of conf, e.g.,

<include executable="inspircd/conf/opers.sh">

and it works for me. However, I do not know why the code is the way it is and it's quite possible that my branch is worse for other use cases. However, if my upcoming PR is inappropriate, then it may be worth adding a little to README.md to explain the dichotomy between the .conf and the .sh files.