Review main Merged

89373[BUGFIX] Improve GUI-managed map/list

Andreas Kienast CR +2 V +2 Hard +168 −126 updated 09/12/25, 12:13

Commit message

The handling and UI of the GUI-managed map/list in the Install Tool's
Configurations module is improved:

* the markup is not rendered in tables anymore, allowing to use the full
  available width without hacks
* the arbitrary "empty" row has been replaced with a more explicit
  button that focuses the input field, avoiding UI magic
* when a new row is added while the last one has empty inputs,
  the first empty one is focussed
* the row being inserted is stored in a `<template>` tag, eliminating
  the need to do additional DOM manipulations after insert
* dropped "internal" CSS animations and Bootstrap utility classes
* re-styled the configurable "key/value" header for `map` types
* added missing "description" to `element-list/map` types.

Review

  • Code review CR +2 Done
    • Andreas Kienast +2
    • Benjamin Franzke +1
    • Garvin Hicking +1
  • Verified V +2 Done
    • Andreas Kienast +2
    • Benjamin Franzke +1
    • Garvin Hicking +1
    • core-ci +1 CI

Issues 3

Files 5

Review log 2734

Show 69 earlier messages
  1. Andreas Kienast uploaded patch set 1 05/13/25, 14:57

  2. core-ci voted on patch set 1 V +1 05/13/25, 15:07 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/78851

  3. Andreas Kienast uploaded patch set 2 V +1 05/14/25, 10:46

  4. core-ci voted on patch set 2 V +1 05/14/25, 10:52 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/78901

  5. Andreas Kienast uploaded patch set 3 V +1 05/14/25, 14:34

  6. core-ci voted on patch set 3 V +1 05/14/25, 14:40 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/78916

  7. Garvin Hicking commented on patch set 3 CR +1 V +1 05/15/25, 10:32

    Thanks a lot for the rework. Tested in FF and Chrome, works for me and is a good improvement in usability in my opinion.

    local-configuration.ts:135

    Good question, I am unsure. Currently the code puts the sequential number of an array key as the empty key (so the label "required key" is actually misleading). I would argue an empty key for a map does not make sense; whenever we will use maps we'll likely require a key. So we should decide if we: a) Allow empty keys and auto-fix them (keep current logic). Rename "(required key)" to "(required key, auto-enumerated when empty)" or something similar. I don't know a good short description right now. We might even remove "(required key)" completely then? b) Actually enforce non-empty keys (filter out values with an empty key, effectively removing configuration values) c) Actually allow empty keys, remove "(required key)" and do not auto-enumerate. This would probably be the more complex solution. Personally I'd vote for a.) or just keep the todo in place for now.
  8. Andreas Kienast commented on patch set 3 05/15/25, 10:52

    local-configuration.ts:135 Reply

    The comment is about **values**, not keys. However, we may use `required` and change how the data is submitted to the server (`requestSubmit()`), but that needs more UI/UX considerations, e.g. in case a required field is hidden in a collapsed section.
  9. Garvin Hicking commented on patch set 3 05/16/25, 07:10

    local-configuration.ts:135 Reply

    Oh, sorry for my confusion (even though the comment could stay for "keys"). As for values, I'm not sure if our default config should allow empty values? At least for the options I'm currently thinking of, empty values would make no sense. Maybe we should only address that once it becomes a necessity? Since it's only internal "API" and only for Core config options, I'd say this should be non-blocking?
  1. Georg Ringer voted on patch set 3 CR +1 V +1 05/23/25, 06:05

  2. Benjamin Franzke commented on patch set 3 06/02/25, 06:28

    SubSection.html:64

    The case when item.value is empty looks odd, as it only renders a button, which looks like a mistake.

    SubSection.html:67

    We don't have a usage in core yet, but I think it's now even harder to visually understand that we're configuring key-value pairs here. Previously there was a Key/Value table header, that seems to be gone, so that it's really just two input fields next to each other(?)

    SubSection.html:126

    There is too less gap between rows, causing input-focus border to overlap other rows/this button.
  3. Garvin Hicking commented on patch set 3 08/13/25, 20:17

    SubSection.html:67 Reply

    Good catch. Yes my idea was to have configurable headers with a key/value fallback if not specified. I think we should implement that here too.
  4. Garvin Hicking commented on patch set 3 08/13/25, 20:20

    local-configuration.ts:135 Reply

    I would love to continue this patch if there's time and interest - then I could move along this patch which otherwise would have some quirks: https://review.typo3.org/c/Packages/TYPO3.CMS/+/87582
  5. Simon Praetorius commented on patch set 3 08/21/25, 14:39

    SubSection.html:60

    Can you add the field description here, similar to the other field types? I've created a parallel issue and patch for that because I wasn't aware of this one. Will abandon my patch now. https://review.typo3.org/c/Packages/TYPO3.CMS/+/90464
  6. Garvin Hicking commented on patch set 3 V 0 CR -1 08/21/25, 15:15

    on it
  7. Garvin Hicking uploaded patch set 4 CR +1 V +1 08/21/25, 19:18

  8. core-ci voted on patch set 4 V +1 08/21/25, 19:26 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/83709

  9. Garvin Hicking commented on patch set 4 CR +1 V +1 08/21/25, 19:28

    Reply

    Done
    I've adjusted the patch. Since core has no "map" config item yet, please use the changes to `DefaultConfiguration.php` and `DefaultConfigurationDescription.yaml` of https://review.typo3.org/c/Packages/TYPO3.CMS/+/87582. I think this is a good improvement, and having this in would be a dependency to continue on https://review.typo3.org/c/Packages/TYPO3.CMS/+/87582, which is why I'm voting on it, and we could still try to improve it further in follow-ups?

    local-configuration.ts:135 Reply

    I've renamed the keys/values accordingly, following my "a)" approach and I feel comfortable with that as a better solution for now.

    SubSection.html:60 Reply

    Thanks, this is now inserted.

    SubSection.html:64 Reply

    Could you elaborate on this, I'm not really sure what you mean?

    SubSection.html:67 Reply

    I've re-added the label. It is not perfectly aligned (due to a missing button in the row) but when it's centered it's not too obvious, so maybe that's sufficient enough to resolve this. Please unresolve if you have a better solution to it that could "calculate" the width of the button below as a placeholder element.

    SubSection.html:126 Reply

    I double the gap and would say this is better now.
  10. Benni Mack commented on patch set 4 09/01/25, 15:13

    SubSection.html:64 Reply

  11. Benjamin Franzke commented on patch set 4 09/10/25, 09:37

    SubSection.html:64 Reply

    I'm sorry! I missed the notifications for the clarification demand. Here's a screenshot of an empty list with patchset 3: https://bnf.dev/~ben/uploads/67f26/review-39373-lockRootPath-patchset3.png Which shows little to no context for the end user, that's what I was referring to as looking odd: The button looks misplaced as it is "alone in the right corner", like it should be more prominent, when it's the only option (yet). This improved with patchset 4 since a description and some "Value" (pseudo-)table header is shown: https://bnf.dev/~ben/uploads/befe7/review-39373-lockRootPath-patchset4.png I think the interface would be better for the button to be left aligned (to comply with buttons in EXT:form repeatable elements), but that's not a blocker right now. I'm resolving here.
  12. Benjamin Franzke commented on patch set 4 CR -1 09/10/25, 09:40

    on it
  13. Benjamin Franzke uploaded patch set 5 CR +1 V +1 09/10/25, 09:50

  14. Benjamin Franzke commented on patch set 5 09/10/25, 09:50

    SubSection.html:70

    I understand that this is used to align spacing with the *actual* form field below, but we must not add `.form-control` to non form-elements "only" to pick some common styles, the side effects are bad and cause a hardly readable font (light gray on white) in this case. Before I start trying to fix this div-stuff, is there a reason why we did not use `<table>` here? This looks a lot like a table to me.
  15. core-ci voted on patch set 5 V -1 09/10/25, 09:57 Core CI is not happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/84768

  16. Andreas Kienast commented on patch set 5 09/12/25, 06:33

    SubSection.html:70 Reply

    I switched to divs because I don't like tables for such kind of things, but that's just a personal opinion. Please go ahead with whatever solution is the best in this case.
  17. Simon Praetorius wrote 09/12/25, 06:56

    Removed cc Simon Praetorius.
  18. Benjamin Franzke uploaded patch set 6 CR -1 V -1 09/12/25, 08:27

  19. Benjamin Franzke commented on patch set 6 09/12/25, 08:27

    SubSection.html:70 Reply

    Ok, I don't want to revert it all to tables right now, so I adapted the CSS to be scoped to this case and avoid `.input-group` and `.form-group` on non form elements. I also replaced the hack to center the headers (to workaround the alignments) by an alignment hack (yea i know, it's still a hack, but I think it's ok).
  20. core-ci voted on patch set 6 V +1 09/12/25, 08:34 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/84889

  21. Benjamin Franzke uploaded patch set 7 CR +1 V +1 09/12/25, 08:36

  22. core-ci voted on patch set 7 V +1 09/12/25, 08:44 Core CI is happy: https://git.typo3.org/typo3/CI/cms/-/pipelines/84891

  23. Garvin Hicking commented on patch set 7 CR +1 V +1 09/12/25, 11:55

    Many thanks! I'm happy with it until it gets thrown out in v14 :-)
  24. Andreas Kienast voted on patch set 7 CR +2 V +2 09/12/25, 12:01

  25. Andreas Kienast merged patch set 8 09/12/25, 12:13