#106668CTRL+K does not toggle Backend Search Modal
Description
Hi,
Currently CTRL+K only opens the search modal, the next CTRL+K focuses the browser address-bar search input.
One can exit the modal by ESC .
Feature request: As a user I would like to toggle the search modal by CTRL+K. CraftCMS has it implemented that way.
Reviews 0
No change resolves or names this issue in its commit message, and no comment links one.
History 16
-
Garvin Hicking changed category Tags ux-decision 05/08/25, 10:09
-
Benni Mack changed Sprint Focus On Location Sprint 08/07/25, 11:58
-
Fabian Cloïn commented 08/07/25, 19:59
-
Garvin Hicking changed status
NewNeeds Feedback 08/07/25, 22:30 -
Ingo Fabbri commented 08/08/25, 05:52
Fabian Cloïn wrote in #note-3:
> Hey there!
> I looked into this feature. If i take a look at the APG guidelines, i only find a mention about closing it with the `esc` key.
> https://www.w3.org/WAI/ARIA/apg/patterns/dialog-modal/. A quick google for closing modals with the metaKey. (command or windows button) in combination with the `k` key doesn’t give any results. Personally i would suggest to keep it at just the `esc` key. Would you agree with that?I’d suggest preventing Ctrl/Cmd + K from triggering the browser’s default search while the modal is open.
-
Sven Liebert changed status
Needs FeedbackNew 08/24/25, 20:36 -
Garvin Hicking commented and changed status
NewRejected 08/25/25, 07:19Hi! Thanks for the feedback. :)
I checked back with Andreas Kienast as the author of our livesearch integration. We intentionally did not want to permanently (b)lock the browser shortcut here, and allow users to access the regular functionality. Thus, a toggle is not something we plan for at this time, even though other CMS may implement it this way. I hope you can understand that for this we’ll keep it this way.
-
Ingo Fabbri commented 08/25/25, 08:01
Thank you, I do understand. What I find a bit difficult, though, is the four months of decision time. Along the way my proposal was questioned, a re-proposal was made, and in the end it was simply declined. From my perspective, this could have been clarified within one or two weeks. I regret that the decision-making process in TYPO3 sometimes takes so long.
-
Garvin Hicking commented 08/25/25, 08:45
I understand. This is also frustrating to me. Sadly this is how OpenSource works – I for example am not paid to do triage, and need to coordinate this in my free time. You can see how many issues we have open.
I spend about 20 hours a week for free and still can only move so much. Others also do work where they have to prioritize.
I am very sorry about this bad experience and can relate. But unless we find more people interested in “working the issues” it will have a similar outcome.
Some tickets even take years to receive feedback. Decisions are exceptionally hard – a single person cannot easily just decide, but have to involve others, and they also have other things to prioritize.
If you have ideas how we could solve this, we could discuss this on talk.typo3.org for example.
-
Ingo Fabbri commented 08/25/25, 10:53
Garvin, please don’t take any offense — that wasn’t my intention. It was merely a small suggestion, and it just surprised me that it took four months to be rejected.
Over the years I’ve reported many issues or added thoughts to existing Forge tickets (though my former account was accidentally deleted). I’ve also contributed very small patches that often led to lengthy discussions in Gerrit — sometimes just because my commit message wasn’t ideal or the change wasn’t immediately understood by core mergers. In a few cases, patches remained stale for years until someone else eventually picked them up, adjusted them slightly, and merged them without further issues.
My impression is that in the TYPO3 world it often helps a lot if you personally know the right people — then things like new features can move quickly, as happened with the dashboard. Without such connections, however, even important patches may get ignored, dropped, or turn into discussions that can discourage potential contributors.
-
Garvin Hicking commented 08/25/25, 16:54
Yes, I’ve seen issues and patches you were involved in. People can easily feel offended if they are not “seen” or understood, both for reporters but alsi the volunteer core developers. It’s a microcosm after all, like any tight-knit OpenSource project.
Even big ones are often run by very few peopme who often know each other. Due to that understanding and knowledge, it’s often easier to talk to each other than with contributors that do that more sporadic or on very focussed issues. As far as I have gotten to know core developers, all of them care deeply for the product and take reports very seriously. They are however also under constant pressure to meet their own and other expectations. Thats often a battle you can’t win, and drains energy.
What gives energy is people cheering and recognizing the effort and being understanding and positively enforcing. Luckily that happens too (otherwise we would have all ran away).
I for one collect issues I want to inspect in a mail queue, and I often know it will take some hours to inspect or talk to others. And then issues pile up and up, ans on one day a more pressing bug comes up, or I have issues from me and my colleagues/friends, or something I know I could fix with less time than other issues. So the pile of shame gets bigger and more forgotten. But never because I do not value issues and work, but just due to it being an uphill battle.
Anyhow. My point is – I think few FOSS projects of our scale work differently. We have issues that lay there for years, but also other issues we solve in a day. Sometimes its a matter of luck. I know it’s frustrating, but: we don’t do this by intention.
-
Ingo Fabbri commented 08/25/25, 17:20
Back to the topic:
The reason I initially opened this issue was that, if I recall correctly, I wasn’t able to return to the search modal input field with a second Ctrl+K (or any to me known keyboard keys) after I had landed in the address bar. Would that behavior be considered an issue worth addressing?
-
Garvin Hicking commented 08/25/25, 17:26
When you’re in the address bar, no JS of a site gets executed. You need to focus back to the website to do that. Some browsers do that when hitting ESC.
-
Ingo Fabbri commented 08/25/25, 17:33
Garvin Hicking wrote in #note-13:
> When you’re in the address bar, no JS of a site gets executed. You need to focus back to the website to do that. Some browsers do that when hitting ESC.yes, but not returning to input field. focusing anything undefined on site. another ctrl+k does focus address bar again. you are basically stuck in a loop. While the modal is still open…..
-
Garvin Hicking commented 08/25/25, 18:56
“Esc” to get back to the website focus and then “esc” again should close the modal?
-
Ingo Fabbri commented 08/26/25, 04:40
Garvin Hicking wrote in #note-15:
> “Esc” to get back to the website focus and then “esc” again should close the modal?I have no access to a device with keyboard right now. But as a mouse user I can perfectly put focus on the modal’s input field again. Something I could not do by keyboard. I will have a look at this next week and open a bug report issue when I think it is needed. This issue – intended as a enhancement proposal and not a bug report – is done. Thanks.
Hey there!
I looked into this feature. If i take a look at the APG guidelines, i only find a mention about closing it with the `esc` key.
https://www.w3.org/WAI/ARIA/apg/patterns/dialog-modal/. A quick google for closing modals with the metaKey. (command or windows button) in combination with the `k` key doesn’t give any results. Personally i would suggest to keep it at just the `esc` key. Would you agree with that?