Open bdarcus opened 1 year ago
I have wanted this on occasion but not enough to implement it! I think if we want to do it, we have to do what vertico-repeat
does and keep track of the input in post-command-hook
.
If it's not too complicated, it does seem like it'd be a nice usability enhancement.
If it's not too complicated, it does seem like it'd be a nice usability enhancement.
It shouldn't be too complicated and I might take a stab at this soonish. But this has got me thinking if vertico can provide a simple multiple selection command based on vertico-repeart
. It has all the pieces in place for that.
Not sure how that would differ from the earlier CRM prototype (which last I checked is still in that report on the crm
branch), but he abandoned that because there was one or two limitations he couldn't resolve, the details of which I can't remember clearly, but which I think related to dynamic tables.
Not sure, but this new 29 stuff may have some relevance?
This commit make citar-open-files
cannot open pdf file on a select candidate.
The step I do: M-x citar-open-files, move to one candidate with arrow key instead of input any char to filter candidate, then enter to open file. The echo area display No associated files for nil
, but I do have pdf file in chosen candidate.
I downgrade to commit a9fae9c , it works well.
Thanks @Jousimies!
@aikrahguzar - do you have time to take a look?
@aikrahguzar - do you have time to take a look?
Not any time soon I am afraid. I don't understand very well how citar-open-files
works now. If someone can give a recipe for a misbehavior of citar--select-multiple
I might be able to do something.
I don't understand very well how
citar-open-files
works now.
It just presents a list of paths for llbrary files associated with the key(s).
If someone can give a recipe for a misbehavior of
citar--select-multiple
I might be able to do something.
@Jousimies - can you do this?
I haven't yet been able to reproduce it.
FYI, @Jousimies, I reverted this for now.
I'm reopening this, because I reverted it without every figuring out a fix for the initial commit.
I'm not exactly sure how to re-implement this, since neither of us were able to reproduce the bug reported by @Jousimies.
I do consider this to be one obvious missing UX feature that I really want to add back though.
I'm not exactly sure how to re-implement this, since neither of us were able to reproduce the bug reported by @Jousimies.
My dim memory of what happened was that the code used a feature of completing-read
that was new in Emacs 29. I think the easiest thing to do is to use the initial-input
arg of the completing-read. It is technically deprecated but I don't think it is going away.
From mastodon.
I think we'd want to allow a modifier key on
TAB
that would preserve the input, though I'm unsure how to do that ATM. Any ideas @aikrahguzar?