Page MenuHomePhabricator

Remove "Add Existing Panel" from dashboard arrange view
AbandonedPublic

Authored by chad on Mar 22 2017, 2:08 AM.
Tags
None
Referenced Files
Unknown Object (File)
Thu, Nov 28, 3:20 AM
Unknown Object (File)
Sat, Nov 23, 9:47 PM
Unknown Object (File)
Fri, Nov 22, 7:08 PM
Unknown Object (File)
Oct 28 2024, 12:46 AM
Unknown Object (File)
Oct 16 2024, 2:44 AM
Unknown Object (File)
Oct 9 2024, 7:41 AM
Unknown Object (File)
Oct 1 2024, 12:02 AM
Unknown Object (File)
Sep 23 2024, 7:26 PM
Subscribers

Details

Reviewers
epriestley
Maniphest Tasks
T10390: Dashboards v2 (UX)
Summary

Ref T10390. Going to remove this button from arrange view and add "Install Panel" to individual panel views. If someone is building a custom panel there, it's slightly nicer anyways to just directly install it. Also simplifies the arrange page.

Test Plan

Try to install to an existing dashboard, success. Add same panel again, get error. Archive all dashboards, get notice.

Diff Detail

Repository
rP Phabricator
Branch
master
Lint
Lint Passed
Unit
Tests Passed
Build Status
Buildable 16077
Build 21327: Run Core Tests
Build 21326: arc lint + arc unit

Event Timeline

I think we should remove this workflow entirely, not just move how you access it.

If we have any ability to add an existing panel to a dashboard anywhere in the product, we can not make the policy simplification for T8033.

Specifically, we get into trouble when users do this:

  1. Create a "Visible To: Public" dashboard.
  2. Create a standalone "Visible To: epriestley" panel.
  3. Put the panel on the dashboard.

By completely preventing step (3), we can avoid the problem in T8033. But if step (3) is possible, we're stuck with the possibility that dashboards and panels may have different policies.

This revision now requires changes to proceed.Mar 30 2017, 3:24 PM

The issue to me is tab panels. We have no means of creating them ad-hoc and any workflow for that seems... very complex.

In the long run, I think we're better off doing the legwork for tab panels (which I agree is pretty substantial, but seems tractable) than allowing the mixed-policy state (which I think will be confusing forever).

Up to you, I don't have any more work planned on Dashboards. Everything left on v2 is either minor or too substantial for me to build (or both).