Skip to content

HSS publisher, do some work on ReactorErrors

Ian Jackson requested to merge Diziet/arti:reactor-error-3 into main

I've spent a while trying to abolish ReactorError and the comment in "HSS publisher errors: Declare that we want to abolish ReactorError" explains my feelings as a result.

I did manage to abolish one of its variants. The rest seemed distinctly nontrivial. Starting at the top of the call stack definitely wasn't working well. I think untangling/abolishing MissingKey (which also exists in FatalError with a different comment complaining that it's wrong) would probably help a fair bit.

If you can think of a better approach, please do say.

Merge request reports