I must say that while this makes some sense at the UI level to me, resignalling errors is a bad practice all around. Not only does it convolute the code, but errors should have, in all situations, I guess even in interactive ones, as much stack as they can. I think the correct fix here is to change the way that the error is *printed*. In the end that was the original motivation behind the bug report! In CL this is easy enough to do with the PRINT-OBJECTgeneric function. That print function would print the alternate message in this particular interactive context. I think Emacs has recently gained the means to do something similar to PRINT-OBJECT, has it not? I seem to remember Stefan Monnier telling me so. If it has not, and we do decide to go this route, I'd just nitpick that the parameter name should be INTERACTIVE or RESIGNAL-ERROR. So that unsuspecting programs calles won't expect the error to have the properties of the scan-error object if they pass non-nil in the parameter. João On Wed, Sep 23, 2020 at 3:34 PM Mattias Engdegård wrote: > 23 sep. 2020 kl. 15.40 skrev Lars Ingebrigtsen : > > > The parameter is NOERROR, but now it does signal an error. :-) > > > > So perhaps the parameter should be USER-ERROR/TERSE-ERROR or something > > and the doc strings adjusted? > > Good catch, thank you! The pushed change uses USER-ERROR. > > > > > -- João Távora