Daniel Mendler

See https://github.com/minad/consult/issues/636#issuecomment-1229481271 where I also complained about this. But I am too lazy to debate with the maintainers. They are stubborn anyway. :-P

> I really have no opinion on this matter. All I know is that nothing changes upstream, all users will encounter at least one warning whenever an embark (or vertico,...

> Reading the [mailing list thread](https://debbugs.gnu.org/cgi/bugreport.cgi?bug=26624) linked to by @tarsius in https://github.com/oantolin/embark/pull/546 has made me change my mind. @oantolin May I ask which argument in the thread convinced you? For...

I have difficulties to understand what you are trying to achieve here, but it reminds me of consult-dir and consult-history. If you run consult-dir it will replace the current input...

Thanks for the detailed responseand the clarifications. I understand that there is a general pattern and that it may be worth to pursue this in the general setting of Embark....

To address the completion command reuse aspect - I could imagine a facility where you run an alternative completion command and then instead of invoking the continuation of completing-read you...

Thinking about this more - one thing I would explore more is why "conventional" approaches don't work for you or for this special use case. Is there not a way...

@karthink I like the idea of your `minibuffer-replace-input` command, which propagates input upwards. I think this could fit Embark with a few changes. 1. In the recursive minibuffer `minibuffer-replace-input` should...

Hmm, the way I described the workflow in my previous comment is just the reverse workflow of embark-act. Instead of Citar->act->attach you do attach->Citar->replace-input. Given that I don't think such...

I just investigated this issue in the context of my Corfu completion package which exclusively builds upon Capfs (`completion-at-point-functions`). While the modern Capf API is okay, there seems to be...