On Thu, Apr 08, 2021 at 06:46:34PM +0000, Joseph Myers wrote:
> 
> I don't think it's a good idea to depend on whether the exception is 
> raised directly in a detectors file.  (I prefer the style where all the 
> substantive code used by detectors etc. is in external modules and as 
> little as possible is in the .py files in the instance.)

Me too

So I think we could
- Remove the special handling for KeyError, IndexError now
- Deprecate the use of ValueError for uses where Reject should be used
- In a future version remove the special handling for ValueError, too

I'm not very happy with additional options for these.
And I think Johns Use-Case (invalid value entered by the user) can be
met by using Reject for these cases, too (the user does *not* see any
difference between Reject and ValueError). And all too often a
ValueError is unintentional in a detector. It could even disclose
information to the user that she should not see at this point.

What do you think?

Ralf
-- 
Dr. Ralf Schlatterbeck                  Tel:   +43/2243/26465-16
Open Source Consulting                  www:   www.runtux.com
Reichergasse 131, A-3411 Weidling       email: office@runtux.com