[AGILE-251] Migrate Backlogs to Pragmatic DnD by myabc · Pull Request #23218 · opf/openproject

myabc · 2026-05-14T20:30:27Z

Ticket

https://community.openproject.org/wp/AGILE-251
https://community.openproject.org/wp/AGILE-284 (optimistic reordering, folded in from #23643)
https://community.openproject.org/wp/AGILE-272
https://community.openproject.org/wp/AGILE-323

Resolves BUGs: https://community.openproject.org/wp/AGILE-252, https://community.openproject.org/wp/AGILE-258

What are you trying to accomplish?

Migrate Backlogs drag and drop to Pragmatic Drag and Drop while preserving the existing inbox / sprint / backlog-bucket move workflows.

The legacy generic (jQuery/Dragula) wiring is replaced by a small Stimulus suite on @atlaskit/pragmatic-drag-and-drop. Move semantics and server endpoints are unchanged — only the client interaction layer. Backlogs cards now render as ViewComponents (server-rendered markup), and the root controller repairs DnD wiring after Turbo morphs (a turbo:morph-element heal re-registers lists and items) so DnD survives stream updates.

What approach did you choose and why?

The DnD layer is a reusable, Backlogs-agnostic sortable-lists Stimulus suite:

sortable-lists.controller.ts — list "root": owns the move request, drag monitor, and auto-scroll; scopes each drop to its root (accepted type matches, source + resolved target contained) so multiple roots don't steal drops.
item.controller.ts — per-card draggable + drop target with closest-edge indicators and a sanitized drag preview.
drag-and-drop.ts — the Pragmatic payload contract plus pure, unit-testable drop-intent resolution.
list-dom.ts — DnD-agnostic list/row DOM contract: selectors, item/row resolution (incl. truncation markers), optimistic reorder with snapshot/restore.
turbo/pragmatic-dnd-morph-attributes.ts — a turbo:before-morph-attribute guard preserving client-owned DnD state attributes (drop indicators, selection) across idiomorph morphs. Pragmatic's internal registration marker is deliberately not preserved: registration is repaired by the root controller's turbo:morph-element heal instead, avoiding marker/registry divergence.
list-refresh.controller.ts — non-DnD: reloads the frame when a work package changes outside a drag (split-pane / Angular edits emit HAL events, not a Turbo stream), hardened against a subscribe-after-disconnect race.

Reordering is optimistic (AGILE-284, from #23643): the row moves on drop and persists in the background; the turbo-stream response reconciles on success, and a failure rolls back with a prefers-reduced-motion-aware FLIP animation.

Server moves still flow through Backlogs::WorkPackages::UpdateService, which normalizes list_id before validating list_type (sprint / backlog bucket / inbox) so an empty inbox id and an integer id are handled consistently.

Keyboard handling is intentionally minimal: the card is the single focus target and the action menu is the accessible move path. Prototyped keyboard DnD and the drag-specific ARIA-role layer were removed rather than shipped half-complete; full accessible DnD is tracked under follow-up work.

Working checklist

Addressed from review (2026-07-05 triage)

Scope backlogs drag styling under .op-border-box-list so the global data-dragging marker no longer restyles admin/settings dragula lists (review #5)
Guard optimistic rollback against a captured parent detached by a mid-flight morph (review #9)
Clear the drop indicator when an item disconnects mid-drag (review #7)
Document the list-level "drop container" and drop-position-value in frontend/CONTEXT.md (review #15)
Hoist the sortable-lists PDD mocks off the shared Vitest registry (review #13)
Route list_type through Backlogs::Target instead of hand-spelled literals (review #14)

Review follow-ups (deferred; 2026-07-05 triage)

Hold the moving guard until the reconciling frame reload lands, not on PUT resolve — a second drag races the about-to-morph DOM (review #2; sortable-lists.controller.ts) → https://community.openproject.org/wp/AGILE-338
Scope the morph guard to owner-controller elements — it currently blocks every aria-current removal inside the backlogs frame (review #8a; pragmatic-dnd-morph-attributes.ts) → https://community.openproject.org/wp/AGILE-336 (server hydrates aria-current, controller re-syncs after morphs; drag-gating rejected — the guard exists for non-drag list-refresh morphs)
Restore the truncation marker ("Show N more") as a drop target — a direct drop no-ops (same-list) or top-inserts (cross-list) instead of filing after the collapsed block (review #10; sortable-lists/drag-and-drop.ts) → https://community.openproject.org/wp/AGILE-337 (product input needed on target semantics)
Wrap the two-phase list move (UpdateService commit + post-commit move_after) in one transaction — a failure between the phases can leave a stale/nil position (review #12; work_packages/update_service.rb) — done in stacked PR [AGILE-322] Backend support for optimistic reordering #24157
Move to a filtered-out sprint/bucket silently hides the card (dialogs offer unfiltered targets, the frame re-render honours filters) — extended the invisibility check to the sprint/bucket filters; dialogs deliberately keep offering every destination (review #4; work_packages_controller.rb)
Silence the vitest "vi.mock not at top level" deprecation in the sortable-lists specs (top-level vi.mock + aliased core-stimulus/test-helpers; vitest-v4 hoister × vite-alias, non-fatal today)
Make the backlogs DnD pickup helper robust without retry — pick_up_and_release_work_package is now capped via retry_block, but the stale node it retries on (likely a mid-drag list-refresh morph detaching the grabbed row) should be waited out or re-queried inside the drag so no retry is needed (spec/support/pages/backlog.rb) → root cause (mid-drag list refresh) tracked in https://community.openproject.org/wp/AGILE-336

Review follow-ups (deferred; 2026-07-09 review)

Phase the busy overlay in via CSS (opacity) and debounce it so it only appears when the move request takes longer (discussion; sortable-lists/list.controller.ts)
Replace the manual cross-controller selection state handling with Stimulus dispatch events (discussion; backlogs/work-package.controller.ts)
Extract the backlogs.controller.ts split into list-refresh / work-package controllers into a separate PR — parked for now (discussion)

Actionable issues from Codex review (2026-07-10)

Restrict sticky item targets to the rows region so moving over a list header or empty space honors the configured dropPosition instead of reusing the last-hovered row (sortable-lists/item.controller.ts).
Invalidate optimistic rollback after a concurrent morph removes or repositions the dragged row while its source <ul> remains connected; restore only if the row is still at its optimistic placement (sortable-lists/list-dom.ts).
~~Defer or roll back eager data-selected updates until Turbo accepts the split-pane visit~~ Rejected — eager selection on card activation is intentional: data-selected is ephemeral client state (and becomes multi-select later), while aria-current tracks the URL-synced work package; the two are distinct axes. The review's failure model is also incorrect: the promoted frame visit fires turbo:before-visit from fetchResponseLoaded — after the frame has already fetched and rendered — so a declined unsaved-changes prompt leaves the new pane visible, matching the eager selection; only the URL and aria-current stay behind, by design. (Supersedes review #6.)

Actionable issues from Codex + Copilot review (2026-07-13 triage)

Resolve same-list container drops to the configured dropPosition instead of the blanket no-op guard, and short-circuit genuine no-op drops before the move request (sortable-lists/drag-and-drop.ts, sortable-lists.controller.ts)
Flash the existing "moved outside this filter" notice when a menu move targets a filter-hidden sprint/bucket (work_packages_controller.rb#work_package_invisible_after_move?)
Bound the bare retry in the four DnD drag spec helpers via retry_block (spec/support/pages/backlog.rb) — parked; revisit alongside [AGILE-322] Backend support for optimistic reordering #24157 (its frame_reload derivation touches the same helpers). Chrome 150 also surfaces mid-morph stale nodes as Selenium::WebDriver::Error::UnknownError ("Node with given id does not belong to the document"), which no current rescue list catches — fold into the same pass.
Guard the morph heal to children of the owning root — hardening; nested roots are theoretical today (sortable-lists.controller.ts)

Archived

Follow-up work

Add accessibility live announcements for sortable-list drag start, candidate position change, drop, and cancellation using Primer live-region-element
Configure data for external consumers
Implement a clickable? option on the BorderBox list work-package item (discussion)

Merge checklist

Added/updated tests
Added/updated documentation in Lookbook (patterns, previews, etc)
Tested major browsers (basic tests)
- macOS — Chrome, Safari, Firefox
- Windows — Chrome, Edge, Firefox
- Linux — Chrome, Firefox
- Android — Chrome (Firefox not supported)
- iOS/iPadOS — Safari, Chrome (Webkit)

github-actions · 2026-05-14T21:05:20Z

Deploying openproject with ⚡ PullPreview

Field	Value
Latest commit	`19a531b`
Job	`deploy`
Status	🧹 Destroying preview...
Preview URL	Destroying

View logs

Copilot

Pull request overview

This PR migrates Backlogs drag-and-drop from the legacy approach to Atlassian’s Pragmatic Drag and Drop, moving Stimulus wiring onto the inner work package card and introducing new frontend helpers/controllers to compute drop targets and ordering.

Changes:

Introduces Pragmatic DnD-based Stimulus controllers/helpers for Backlogs item dragging and list drop targeting.
Moves Backlogs-specific data attributes (story + draggable item wiring) from BorderBox rows to the inner work package card, and updates Backlogs views/components accordingly.
Adds Atlaskit Pragmatic DnD dependencies and updates Ruby/JS specs to reflect the new DOM + data attributes.

Reviewed changes

Copilot reviewed 24 out of 25 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
spec/components/open_project/common/work_package_card_component_spec.rb	Adds coverage for forwarding system arguments to the root card element.
modules/backlogs/spec/support/pages/backlog.rb	Updates test selectors to target the new draggable item data attribute on the card.
modules/backlogs/spec/requests/backlogs/backlog_spec.rb	Adds assertions for Backlogs controller wiring and list targets in rendered backlog HTML.
modules/backlogs/spec/components/backlogs/work_package_card_list_item_component_spec.rb	Adjusts expectations to verify Backlogs Stimulus wiring is on the card, not the row.
modules/backlogs/spec/components/backlogs/work_package_card_list_component_spec.rb	Updates list/component specs for new target attributes and card-level data wiring.
modules/backlogs/spec/components/backlogs/story_points_component_spec.rb	Adds expectation around wrapper positioning for SR-only label placement.
modules/backlogs/spec/components/backlogs/sprint_component_spec.rb	Updates sprint DnD expectations to use Backlogs list targets + card-level draggable attributes.
modules/backlogs/spec/components/backlogs/inbox_component_spec.rb	Updates inbox drop-target assertions to new Backlogs list target attributes.
modules/backlogs/spec/components/backlogs/bucket_component_spec.rb	Updates bucket drop-target + card wiring assertions for new DnD approach.
modules/backlogs/app/views/backlogs/backlog/show.html.erb	Removes legacy content_controller wiring for Backlogs.
modules/backlogs/app/views/backlogs/backlog/_backlog_list.html.erb	Replaces legacy generic-drag-and-drop wiring with `data-controller="backlogs"` root.
modules/backlogs/app/components/backlogs/work_package_card_list_item_component.rb	Moves Backlogs Stimulus wiring/data to the inner card; adds draggable card args.
modules/backlogs/app/components/backlogs/work_package_card_list_component.rb	Changes list target data attributes to `data-backlogs-target="list"` + id.
modules/backlogs/app/components/backlogs/work_package_card_component.rb	Forwards system arguments to the shared/common work package card.
modules/backlogs/app/components/backlogs/inbox_component.html.erb	Updates inbox list container data attributes to match Backlogs controller targets.
frontend/src/stimulus/controllers/dynamic/backlogs/item.controller.ts	Adds a Stimulus controller to make individual cards draggable + item drop targets.
frontend/src/stimulus/controllers/dynamic/backlogs/item.controller.spec.ts	Adds unit tests for item controller state rendering + preventUnhandled integration.
frontend/src/stimulus/controllers/dynamic/backlogs/drag-and-drop.ts	Adds helper utilities for resolving targets and computing `prev_id` for move requests.
frontend/src/stimulus/controllers/dynamic/backlogs/drag-and-drop.spec.ts	Adds unit tests for drag-and-drop helper logic (target resolution, ordering).
frontend/src/stimulus/controllers/dynamic/backlogs.controller.ts	Replaces HAL-events refresh behavior with Pragmatic DnD monitoring + move requests.
frontend/src/global_styles/primer/_overrides.sass	Updates styles for new drag/drop indicators and Box-card dragging visuals.
frontend/package.json	Adds Atlaskit Pragmatic DnD dependencies.
frontend/package-lock.json	Locks Atlaskit Pragmatic DnD dependencies and transitive packages.
app/components/open_project/common/work_package_card_component.rb	Allows forwarding system arguments; ensures `Box-card` class on root; adds `card_arguments`.
app/components/open_project/common/work_package_card_component.html.erb	Uses `card_arguments` to apply forwarded system arguments to the root element.

Files not reviewed (1)

frontend/package-lock.json: Language not supported

During functional testing the second card could not be dropped below the previously placed one, while the existing regression spec only pinned a drop above it. Adds an after: option to drag_work_package to cover the lower edge, which passes against the current branch.

The resting blankslate copy and the drag-over overlay copy had slightly drifted apart. Single sentences lose their full stops, the drop label is dropping the redundant container mention and goes singular ("this item", to be revisited once multi-drag lands), and the inbox description merely claims work packages outside a sprint or bucket land there.

Follows the core convention from spec/rails_helper.rb: support directories are meant for glob-required helpers, and _spec.rb files inside them risk being required into every suite and run twice. Specs covering test infrastructure live under spec/support_spec instead.

github-actions · 2026-07-10T11:09:29Z

Warning

Flaky specs

rspec ./spec/features/types/subtypes_index_spec.rb[1:5]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./spec/features/types/subtypes_index_spec.rb[1:5]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

Restores the rows after a failed request only for as long as they remain exactly where the optimistic reorder had placed them. A concurrent morph that removed or repositioned them reflects the fresher server state that a rollback to the stale pre-move snapshot would silently undo.

Keeps row drop targets sticky only when pointer moves within the region. Over the list header or empty spaces, the sticky target releases itself. Otherwise the row would silently land next to its last-hovered item.

github-actions · 2026-07-10T17:15:12Z

Warning

Flaky specs

rspec ./spec/features/types/subtypes_index_spec.rb[1:5]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./spec/features/types/subtypes_index_spec.rb[1:5]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

Adapts the add-existing-work-package flow from PR #23403 to this branch's (list_type, list_id) wire format: Target.parse/to_s give way to Target.from_list/to_list_params, and Target.for_work_package now encodes a work package's current location while Target.for keeps dev's container-based signature.

A type can only be absent on an unpersisted work package, but such a one raised a `NoMethodError` from `to_fs`, even corrupting RSpec failure messages that render work packages. Degrades to the standard-type layout instead, dropping the caption's colon too.

github-actions · 2026-07-11T13:08:20Z

Warning

Flaky specs

rspec ./spec/features/types/subtypes_index_spec.rb[1:5]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./spec/features/types/subtypes_index_spec.rb[1:5]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

github-actions · 2026-07-11T13:37:18Z

Warning

Flaky specs

rspec ./spec/features/types/subtypes_index_spec.rb[1:5]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./spec/features/types/subtypes_index_spec.rb[1:5]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

Copilot

Pull request overview

Copilot reviewed 104 out of 105 changed files in this pull request and generated 4 comments.

Files not reviewed (1)

frontend/package-lock.json: Generated file

github-actions · 2026-07-13T13:49:04Z

Warning

Flaky specs

rspec ./modules/backlogs/spec/features/inbox_column_spec.rb[1:6:1]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./modules/backlogs/spec/features/inbox_column_spec.rb[1:6:1]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

A Turbo morph desyncs the drag-and-drop wiring in two ways: rows the morph replaces never receive the root reference, because Stimulus outlet-connected callbacks do not re-fire for them, so canDrag and canDrop refuse every interaction; and an element can keep Pragmatic's preserved drop-target marker attribute without its WeakMap registry entry, which silently dead-ends the drop-target search and kills all rows rendered underneath it. Both states persist until a full page reload. Re-hands the root and re-registers list, item and scrollable children from controller state once per morph batch.

The morph heal re-registers every child after a morph, restoring the marker attribute together with the registry entry it stands for, so preserving the attribute is redundant. Worse, preservation was the one mechanism that let the marker outlive its registration — the stranded state at which Pragmatic silently aborts its drop-target search. Letting the morph strip it keeps controller state the single source of truth.

The service also raises this failure when `list_id` is combined with the `direction` param, but the message only mentioned `list_type`.

The unified `list_type`/`list_id` wire format dropped the fail-fast on a missing target that `params.expect(:target_id)` previously enforced.

A `backlog_bucket_id` referencing no record cleared contract validation, only to crash on the foreign key constraint when saving. Sprints already have an equivalent guard through the assignable-sprints check.

github-actions · 2026-07-13T17:21:29Z

Warning

Flaky specs

rspec ./spec/features/work_packages/table/switch_types_spec.rb[1:1:1]

🤖 Ask Copilot to investigate

Copy the prompt below into a new comment on this PR to delegate the investigation to GitHub Copilot. It will look into the flakiness and open a separate pull request with you as reviewer.

@copilot The following spec(s) are flaky in CI (first seen on PR #23218, linked for reference only):

- `rspec ./spec/features/work_packages/table/switch_types_spec.rb[1:1:1]`

Treat this as a standalone task, unrelated to PR #23218. Create a new branch from origin/dev and open a new pull request targeting dev — do not stack it on PR #23218 or reuse that branch.

Follow the playbook in docs/development/testing/handling-flaky-tests/README.md to find the root cause and fix the underlying race — do not skip, delete, or weaken the spec to make it pass; disabling is a last resort per the playbook, and only with a bug ticket. Verify the fix by running the spec(s) repeatedly (e.g. `script/bulk_run_rspec --run-count 10`).

If you cannot reproduce the flake or are not confident in a fix after reasonable investigation, do not fabricate a change or skip the spec to force CI green. Instead, leave the pull request in draft and document what you tried, the suspected cause, and any leads in its description, then assign @myabc to take over.

Once the fix is verified, title the PR after the spec(s) it fixes, and use the PR description to explain the root cause, how the change resolves it, and the before/after results. Label the PR `flaky-spec`, assign @myabc, and request a review from @myabc.
On every commit, set @myabc as the sole co-author with a `Co-authored-by:` trailer (use their GitHub no-reply email so it links to their account), so it is traceable who dispatched the fix.

A list-only drop back onto the source list was blanket-nulled while the list still rendered its drop outline, so the affordance lied. On start-position lists (the inbox) it also suppressed a legitimate move-to-top. Resolves such drops like any other list-only drop; the next commit suppresses drops that land at the source's current position before any request fires.

An item dropped at its own current position -- on an adjacent edge or back onto its source list -- still fired the full PUT, server reorder and reconciling frame reload for a DOM no-op. Compares the post- reorder placement against the pre-drop snapshot and returns before the request when nothing moved.

Stimulus outlets are document-scoped, so a broad outlet selector can match another root's lists and items; the heal then hands them the wrong root reference, last write wins. Backlogs pins its selectors under the frame id, so this is hardening for future multi-root pages rather than a live bug.

The move dialogs offer every sprint and bucket while the move URL keeps the page's active filters, so moving to a filter-hidden destination made the card silently vanish from the reloaded frame. Extends the invisibility check to the sprint/bucket filters and names the destination sprint in the existing flash message.

ulferts

What a ride but you reached the goal 🎉

myabc force-pushed the implementation/74970-backlogs-pragmatic-dnd branch from 8ac3e26 to 9994516 Compare May 14, 2026 21:02

myabc added feature DO NOT MERGE javascript Pull requests that update Javascript code pullpreview labels May 14, 2026

myabc added this to the 17.5.x milestone May 14, 2026

myabc force-pushed the implementation/74970-backlogs-pragmatic-dnd branch from 9994516 to 4a84694 Compare May 14, 2026 21:36

myabc requested a review from Copilot May 14, 2026 21:36

Copilot started reviewing on behalf of myabc May 14, 2026 21:37 View session

Copilot AI reviewed May 14, 2026

View reviewed changes

Comment thread modules/backlogs/app/components/backlogs/work_package_card_list_item_component.rb Outdated

Comment thread frontend/src/global_styles/primer/_overrides.sass Outdated

myabc force-pushed the implementation/74970-backlogs-pragmatic-dnd branch 3 times, most recently from cca6f5e to 10fa291 Compare May 15, 2026 18:56

myabc commented May 16, 2026

View reviewed changes

Comment thread spec/support/authentication_helpers.rb

myabc commented May 16, 2026

View reviewed changes

Comment thread spec/support/pages/page.rb Outdated

myabc commented May 16, 2026

View reviewed changes

Comment thread spec/support/edit_fields/edit_field.rb Outdated

myabc mentioned this pull request May 16, 2026

[73968, 73089] Extend and improve design of WorkPackageCardComponent #23175

Merged

16 tasks

myabc force-pushed the implementation/74970-backlogs-pragmatic-dnd branch 5 times, most recently from 2ffd5e6 to ba69795 Compare May 17, 2026 01:28

myabc changed the title ~~Implementation/74970 backlogs pragmatic dnd~~ [#74970] Migrate Backlogs to Pragmatic DnD May 18, 2026

Base automatically changed from implementation/74684-extract-border-box-list-component to dev May 18, 2026 12:55

github-actions Bot removed the pullpreview label May 29, 2026

myabc force-pushed the implementation/74970-backlogs-pragmatic-dnd branch 3 times, most recently from 8c94399 to 3e53b87 Compare June 1, 2026 21:53

myabc requested a review from Copilot June 1, 2026 21:53

myabc added 4 commits July 10, 2026 11:16

UX/UI feedback round: Remove drop overlay border

7da9516

myabc added 2 commits July 10, 2026 17:51

Release sticky row targets outside the rows span

7c812f7

Keeps row drop targets sticky only when pointer moves within the region. Over the list header or empty spaces, the sticky target releases itself. Otherwise the row would silently land next to its last-hovered item.

myabc added 2 commits July 11, 2026 13:48

Copilot AI reviewed Jul 11, 2026

View reviewed changes

myabc mentioned this pull request Jul 12, 2026

make Backlogs Inbox dom_target compatible #24207

Open

UX/UI feedback round: Show drop target label muted

607aa38

myabc added 5 commits July 13, 2026 17:35

Reword ambiguous_target to mention list_id

1b475f8

The service also raises this failure when `list_id` is combined with the `direction` param, but the message only mentioned `list_type`.

Fail fast when add_existing lacks list_type

4f528ca

The unified `list_type`/`list_id` wire format dropped the fail-fast on a missing target that `params.expect(:target_id)` previously enforced.

Validate backlog bucket exists in contract

d0f21df

A `backlog_bucket_id` referencing no record cleared contract validation, only to crash on the foreign key constraint when saving. Sprints already have an equivalent guard through the assignable-sprints check.

myabc mentioned this pull request Jul 13, 2026

[DREAM-756] Migrate Angular DnD surfaces to the sortable-lists engine #24243

Draft

3 tasks

myabc mentioned this pull request Jul 13, 2026

[DREAM-755] Introduce sortable-lists engine and Angular directives on Pragmatic DnD #24229

Open

3 tasks

myabc added 4 commits July 13, 2026 19:32

ulferts approved these changes Jul 14, 2026

View reviewed changes

Merge branch 'dev' into implementation/74970-backlogs-pragmatic-dnd

19a531b

Uh oh!

Conversation

myabc commented May 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Ticket

What are you trying to accomplish?

What approach did you choose and why?

Working checklist

Addressed from review (2026-07-05 triage)

Review follow-ups (deferred; 2026-07-05 triage)

Review follow-ups (deferred; 2026-07-09 review)

Actionable issues from Codex review (2026-07-10)

Actionable issues from Codex + Copilot review (2026-07-13 triage)

Follow-up work

Merge checklist

Uh oh!

github-actions Bot commented May 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Deploying openproject with ⚡ PullPreview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented Jul 10, 2026

Uh oh!

github-actions Bot commented Jul 10, 2026

Uh oh!

github-actions Bot commented Jul 11, 2026

Uh oh!

github-actions Bot commented Jul 11, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented Jul 13, 2026

Uh oh!

github-actions Bot commented Jul 13, 2026

Uh oh!

ulferts left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

7 participants

myabc commented May 14, 2026 •

edited

Loading

github-actions Bot commented May 14, 2026 •

edited

Loading