Migrate user_profile to the new resource-layer methods by rtibblesbot · Pull Request #15183 · learningequality/kolibri

rtibblesbot · 2026-08-12T15:05:32Z

Summary

Reads become retrieve/list, the two profile writes become update, the profile page's user read becomes useRetrieve.
The facilities read points at the shared fetchFacilities_v2, added by Migrate kolibri-common to the new resource methods #15163.

References

Closes Migrate user_profile to the new Resource-layer methods #15064. Recipe Migrate all plugins and shared packages to the new Resource-layer methods #15057, methods Add new simplified Resource-layer methods (additive, non-breaking) #15056.

Reviewer guidance

Code review

For a developer reading the diff.

SelectFacility.vue:125 — NetworkLocationResource is shared outside this plugin, so the recipe's shared-_v2 branch applies: the call site moves, the legacy standalone stays for the static/dynamic resources.
actions.js:13 — update gets no baseline; ProfileEditPage.getUpdates() (:181) already diffs against the loaded user. The save is still a single-field PATCH /api/auth/facilityuser/<id>/.
ProfilePage.spec.js:31 — the module automock is dropped: it stubs useRetrieve too, so the spec spies on the real retrieve.
Reads that omitted force (useCurrentUser, ProfileEditPage.vue:173) now always hit the network, and a MergeFacility poll tick coalesces onto an in-flight GET rather than queueing behind it.
useCurrentUser no longer leaves isLoading true after a failed fetch. clear_v2 resolves response.data, which all three call sites discard.
The merge task's completion path (retrieve, clear_v2) needs one user on two devices, so MergeFacility.spec.js is its only cover.
/profile/#/edit reports one KDS KSelect placeholder-contrast axe AA violation, present on develop too; no markup or style line is in this diff.

QA steps

For a tester installing the build artifacts in a VM. No UI change — every screen below should behave as on develop. Rows 3 and 6 need a second user in the facility; rows 5 and 6 need a second VM on the network running any released Kolibri.

Where	What to check
Profile, signed in as a learner	full name, username, facility and role all show, and a reload shows the same values
Profile > Edit — change full name, username, birth year, gender, save	snackbar shows, page returns to Profile, new values on both the profile and the reopened edit form
Profile > Edit — save with the second user's username, then with a blank full name	each error lands on its own field and focus moves there; a corrected resubmit saves
Profile > Change password — enter the old and the new password	modal closes, and signing out and back in with the new password works
Profile > Change facility — device list	the second VM is found, each device lists its facilities, no facility appears twice, and a device you then power off is dropped rather than wedging the modal
Profile > Change facility — confirmation screen, run once as the only user in the source facility and once with a second user in it	the "last user on this device" wording and the merge-or-new-account choice differ between the runs

Page	Screenshot
Profile, after a full-name save
Change facility, device list

AI usage

Claude Code applied the shared migration recipe and rewrote the two specs whose mocks named the replaced methods. Verified with Jest, prek, and manual QA in a browser against a local server.

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

How was this generated?

Ran pre-flight CI checks (lint, format, tests) and verified all pass
Rebased onto the target branch and resolved any conflicts
Reorganized commit history into clean, logical commits
Audited the diff to ensure only issue-relevant files are changed
Built PR body from the repository's PR template with evidence blocks

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

How was this generated?

Ran pre-flight CI checks (lint, format, tests) and verified all pass
Rebased onto the target branch and resolved any conflicts
Reorganized commit history into clean, logical commits
Audited the diff to ensure only issue-relevant files are changed
Built PR body from the repository's PR template with evidence blocks

	🟡 Waiting for feedback Last updated: 2026-08-25 00:29 UTC

github-actions · 2026-08-12T15:20:35Z

Build Artifacts

Asset type	Download link
PEX file	kolibri-0.20.0a2.dev167_gb8cc8f3f7.pex
Windows Installer (EXE)	kolibri-setup-0.20.0a2.dev167+gb8cc8f3f7-unsigned.exe
Debian Package	kolibri_0.20.0a2.dev167+gb8cc8f3f7-0ubuntu1_all.deb
Debian Package	kolibri-server_0.20.0~a2.dev167+gb8cc8f3f7-0ubuntu1_all.deb
Mac Installer (DMG)	kolibri-0.20.0a2.dev167+gb8cc8f3f7.dmg
Android Package (APK)	kolibri-0.20.0a2.dev167+gb8cc8f3f7-debug.apk
Raspberry Pi Image	kolibri-pi-image-0.20.0a2.dev167+gb8cc8f3f7.zip
TAR file	kolibri-0.20.0a2.dev167+gb8cc8f3f7.tar.gz
WHL file	kolibri-0.20.0a2.dev167+gb8cc8f3f7-py3-none-any.whl

Smoke test screenshot

Download screenshot

rtibblesbot · 2026-08-12T18:37:05Z

Build EXE file failed on the chocolatey community feed, not the diff: choco install wget got Failed to fetch results from V2 feed → Unable to find package 'Wget', so make webview2 then died on wget: command not found. #15184 hit the same feed error on innosetup 45 minutes later and went green on its next push. The feed answers 200 again; the job needs a re-run, which this account has no rights for.

rtibblesbot · 2026-08-12T18:56:04Z

Resolved without intervention: a fresh run on the same SHA (31629087140) built the EXE green. No re-run needed.

rtibblesbot · 2026-08-17T17:56:48Z

Added a QA steps section to the body — same request as on #15182, applied across the family's open PRs.

AlexVelezLl

Just couple of small details :).

AlexVelezLl · 2026-08-17T17:54:10Z

  /**
   * @name fetchModel
   * @function
   * @returns {NetworkLocation}
   */

  /**
   * @name fetchCollection
   * @function
   * @returns {NetworkLocation[]}
   */


Lets also remove these comments, please.

Gone — all six. grep -rn "@name fetch\(Model\|Collection\)" over kolibri/ and packages/ matches only this file: two stubs in each of the three resources. Same deletion as #15163's hunk.

AlexVelezLl · 2026-08-17T17:55:17Z

 function updateConnectionStatus(id) {
  const url = urls['kolibri:core:networklocation_update_connection_status'](id);
  return this.client({ url, method: 'post' }).then(response => {
    return response.data;
  });
 }

 /**
 * Fetch the facilities advertised by a remote network location.
 * @param {string} id - network location id
 * @returns {{facilities: [{id: string}]}} the facilities listing from the remote device
 */
 function fetchFacilities(id) {
  return this.client({
    url: urls['kolibri:core:networklocation_facilities_detail'](id),
  }).then(response => {
    return response.data;
  });
 }


Could we turn this into the NetworkLocationResource object and use this.request instead?

Done — fetchFacilities is now a NetworkLocationResource method over this.request; the standalone and the _v2 sibling I had added are both gone.

The standalone resolved networklocation_facilities_detail for all three resources, so the static and dynamic registrations pointed at the networklocation endpoint regardless. request cannot do that, and nothing called them — dropped.

grep -rn "NetworkLocationResource\.fetchFacilities": one caller, SelectDeviceModalGroup/api.js:46, on NetworkLocationResource and unchanged. SelectFacility.vue:125 is back to fetchFacilities, so that file no longer differs from develop.

updateConnectionStatus is the same shape in this file, but it is Migrate kolibri-common to the new resource methods #15163's to rewrite — left alone so two branches do not both move it.

Landed in develop via #15163 while this sat — fetchFacilities_v2 and updateConnectionStatus_v2 on NetworkLocationResource over this.request, and the @name stubs gone. That covers this thread and the two siblings on this file.

The rebase takes develop's copy whole, so NetworkLocationResource.js is out of this PR's diff; user_profile now calls fetchFacilities_v2 at SelectFacility.vue:125.

rtibblesbot · 2026-08-17T19:08:47Z

Red again on ac943425f, same chocolatey outage as 12 Aug, not self-resolved this time: choco install wget got a 503 from the community V2 feed, exited 0 with Chocolatey installed 0/0 packages, and make webview2 then died on wget: command not found (job). The feed answers 200 now; Build EXE file needs a re-run, which this account cannot trigger. Every other check on this SHA is green, including Frontend tests and All file linting.

AlexVelezLl · 2026-08-17T20:58:32Z

@@ -32,69 +32,27 @@ function updateConnectionStatus(id) {
  });
 }


Should we move this one, too?

Done — updateConnectionStatus is now a NetworkLocationResource method over this.request({ method: 'POST', action: 'update_connection_status', routeParams: id }), same URL and verb as the hardcoded urls[...] call. Its only caller (SelectDeviceModalGroup/api.js:75) is unchanged.

Swept the branch for the same class — legacy this.client({ url: urls[...] }) custom methods on any resource this PR touches: one file matched (git diff --name-only upstream/develop...HEAD | xargs grep -l "this.client("), and it had exactly this one function left. urls is now an unused import there and is dropped.

Static/dynamic registrations dropped too: grep -rn "StaticNetworkLocationResource\|DynamicNetworkLocationResource" shows only create/delete calls in api.js, so neither updateConnectionStatus was ever called.

No spec here for the same reason as fetchFacilities — #15163 adds NetworkLocationResource.spec.js covering both, and a partial copy would be an add/add conflict. Verified locally with a throwaway spec: URL .../networklocation_update_connection_status/dev-1, method POST, resolves response.data.

radinamatic · 2026-08-20T14:33:28Z

Removed the Needs QA label so the PR can be rebased to include changes from #15217 that will unblock the pending QA workflows.

rtibblesbot · 2026-08-20T14:39:18Z

Rebased onto develop at eca38c5, which contains #15217. All five commits replayed without conflict — #15217 is backend-only (kolibri/core/auth, discovery, serializers.py), and this branch touches no Python.

Re-verified after the rebase: pnpm test-jest --testPathPatterns user_profile 73 passed / 11 suites; SelectDevice|NetworkLocation|SyncSchedule (the callers of the two methods moved onto NetworkLocationResource) 335 passed / 3 suites; prek run --from-ref upstream/develop --to-ref HEAD clean.

rtibblesbot · 2026-08-24T19:46:12Z

Holding off on changes while this PR is in 🧪 Needs QA — the branch will not move while it is being tested. Questions still get answered.

Change the board status when QA is done and I'll pick up any outstanding review feedback, CI failures and conflicts on the next cycle.

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

How was this generated?

Ran pre-flight CI checks (lint, format, tests) and verified all pass
Rebased onto the target branch and resolved any conflicts
Reorganized commit history into clean, logical commits
Audited the diff to ensure only issue-relevant files are changed
Built PR body from the repository's PR template with evidence blocks

AlexVelezLl · 2026-08-24T19:46:33Z

Hey @rtibblesbot! Please rebase.

`fetchFacilities_v2` landed on `NetworkLocationResource` with learningequality#15163. The legacy standalone stays in place: the resource is shared outside this plugin, so learningequality#15057 moves the call site only. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

The read and the writes move together: `fetchModel` is what marked the model synced, which is what made `saveModel(..., { exists: true })` PATCH a diff rather than the whole cached object. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

`force: true` has no successor — the cache it bypassed is gone. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

The spec factory drops `clear` rather than keeping it, so a missed call site throws instead of passing silently. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

rtibblesbot · 2026-08-24T19:52:37Z

Rebased onto develop at a703bd9 — one conflict, in NetworkLocationResource.js.

#15163 merged there in the meantime with the same this.request bodies you asked for, as fetchFacilities_v2 / updateConnectionStatus_v2, and with the @name stubs dropped — all three of your requests on that file. Took develop's copy whole: renaming _v2 away now would break SelectDeviceModalGroup/api.js and NetworkLocationResource.spec.js, which #15163 points at those names. The first commit is now just SelectFacility.vue:125 moving to fetchFacilities_v2, and the diff is entirely inside user_profile.

pnpm test-jest --testPathPatterns user_profile: 11 suites / 73 tests, unchanged from before the rebase. NetworkLocationResource + SelectDevice specs green. prek run --from-ref upstream/develop --to-ref HEAD passes.

radinamatic · 2026-08-25T00:15:14Z

@rtibblesbot Once this is ready for the manual testing (labeled Needs QA, could you please separate clearly the QA guidelines for code review (if necessary) from the instructions for manual QA? As of now they seem to be mixed... 🤔

rtibblesbot · 2026-08-25T00:25:32Z

Split, @radinamatic — Reviewer guidance now has ### Code review and ### QA steps.

What moved out of the QA half: the GET/PATCH endpoint and payload assertions, the spec/automock notes and the behaviour deltas, all now under code review. The QA half assumes the build artifacts installed in a VM, so the "register localhost as its own peer" rig became a second VM on the network. All six flows kept, as a Where | What to check table — bullets did not fit the body budget.

Swept my other five open PRs for the same mix: four matched, three needed the fix.

PR	State
#15182, #15206	same unlabelled code-review bullets above a `### QA steps` block — split. #15206's two pingback rows need a seeded `PingbackNotification`, so they moved to the code-review half rather than staying as a step QA cannot run
#15219	its walkthrough asked for "zero rows in `content_channelmetadata`" — reworded to on-screen terms under `### QA steps`
#15207	already split
#15028, #14903	backend-only, no QA half to separate

@rtibblesbot's comments are generated by an LLM, and should be evaluated accordingly

github-actions Bot added APP: User Re: User app (sign-in, sign-up, user profile, etc.) DEV: frontend SIZE: medium labels Aug 12, 2026

rtibblesbot force-pushed the issue-15064-fba872 branch from 02adddd to 70dcce6 Compare August 12, 2026 15:23

rtibblesbot marked this pull request as ready for review August 12, 2026 18:34

rtibblesbot force-pushed the issue-15064-fba872 branch from 70dcce6 to 8df7091 Compare August 12, 2026 18:43

rtibbles assigned AlexVelezLl Aug 12, 2026

rtibblesbot mentioned this pull request Aug 17, 2026

Migrate setup_wizard to the new resource layer #15182

Open

AlexVelezLl requested changes Aug 17, 2026

View reviewed changes

rtibblesbot force-pushed the issue-15064-fba872 branch from 8df7091 to ac94342 Compare August 17, 2026 18:07

AlexVelezLl reviewed Aug 17, 2026

View reviewed changes

rtibblesbot force-pushed the issue-15064-fba872 branch 2 times, most recently from f9aea87 to 2c14075 Compare August 17, 2026 21:58

rtibblesbot force-pushed the issue-15064-fba872 branch from 2c14075 to 5f55fd2 Compare August 20, 2026 14:39

rtibblesbot and others added 5 commits August 24, 2026 12:50

Read the session's user through useRetrieve

6f2ee33

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Migrate the two source-facility user list reads

41d0abe

`force: true` has no successor — the cache it bypassed is gone. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Migrate the merge-task polling loop

07b312f

The spec factory drops `clear` rather than keeping it, so a missed call site throws instead of passing silently. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

rtibblesbot force-pushed the issue-15064-fba872 branch from 5f55fd2 to 07b312f Compare August 24, 2026 19:51

github-actions Bot removed the SIZE: medium label Aug 24, 2026

github-actions Bot added the SIZE: small label Aug 24, 2026

rtibblesbot mentioned this pull request Aug 25, 2026

Migrate packages/kolibri to the new resource-layer methods #15206

Open

		@@ -32,69 +32,27 @@ function updateConnectionStatus(id) {
		});
		}

Uh oh!

Uh oh!

Conversation

rtibblesbot commented Aug 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

References

Reviewer guidance

Code review

QA steps

AI usage

Claude Code applied the shared migration recipe and rewrote the two specs whose mocks named the replaced methods. Verified with Jest, prek, and manual QA in a browser against a local server.

🟡 Waiting for feedback

Uh oh!

github-actions Bot commented Aug 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rtibblesbot commented Aug 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rtibblesbot commented Aug 12, 2026

Uh oh!

rtibblesbot commented Aug 17, 2026

Uh oh!

AlexVelezLl left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

rtibblesbot commented Aug 17, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

radinamatic commented Aug 20, 2026

Uh oh!

rtibblesbot commented Aug 20, 2026

Uh oh!

rtibblesbot commented Aug 24, 2026

Uh oh!

AlexVelezLl commented Aug 24, 2026

Uh oh!

rtibblesbot commented Aug 24, 2026

Uh oh!

radinamatic commented Aug 25, 2026

Uh oh!

rtibblesbot commented Aug 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

rtibblesbot commented Aug 12, 2026 •

edited

Loading

github-actions Bot commented Aug 12, 2026 •

edited

Loading

rtibblesbot commented Aug 12, 2026 •

edited

Loading