Don't edit 廣韻.csv directly.

That file is built from the source files with a script , any changes will be overwritten (so to avoid implicit differences between versions of data, which often makes maintaince difficult, as we had done several times before).

If you want to make fixes to the data, you should incorporate them into the build process, either in a source file or in the build script.

As for the incorrect 釋義補充 data, we are aware of this problem, and are trying to bring up a systematic solution. Ideas are welcome!

Thanks for the clarification. I'll close the pr.