Bug 2040381 - Show error state in Access Connector urlbar button and panel when proxy is unavailable

[jporter-dev]

Description

Bugzilla: Bug-2040381

This is part 2 of 3 for this bug. When the Access Connector proxy is unavailable and Firefox shows the accessConnectorFailure error page (implemented in #1033), the urlbar button now reflects the error state. The icon swaps to an "unavailable" variant displayed in a critical color, the tooltip and label update, and the Access Connector panel displays dedicated error content.

---

Dependencies / Related Issues

• Depends on: Bug 2026242 - Improve access connector UX #683
• Part 2 of: Bug 2040381 - Show custom error page when Access Connector is unavailable #1033

---

Screenshots

Access Connector in active state

Access Connector in error state

---

Testing

• Added tests
• Manual testing performed

Steps to verify changes:

1. Enable the AccessConnector policy with a proxy server that is unreachable (ex: change Port by 1)
2. Navigate to a site covered by the inclusion patterns
3. Observe the Access Connector urlbar button and panel in an error state

Expected result:

The Access Connector error button renders with an error icon, and the panel title and message show error text.

[jporter-dev]

This is blocked by both the UX PR and the upstream IPPProtection.getPrincipalRule patch, but it could be reviewed now in preparation of those both landing. This PR is based on #683 and I've added a temporary commit with the relevant changes from the upstream patch to get things working.

Looking for a review despite this being blocked just to get it to good state ahead of time 😄

Upstream blocking PRs have all landed!

[1rneh]

I started reviewing the Bug 2040381 - * commits only but noticed that they also include some changes from Bug-2026242 (or elsewhere) that by now have been merged to enterprise-main.

Could you rebase the PR to simplify the review?

[lissyx]

Looks like it risks getting stale over future

[lissyx]

Make sure this is ok by l10n people?

[jporter-dev]

I'm not sure what you're referencing - this adds new strings for the error state and uses the same conventions as the "active" state in the file. Maybe this had something else in it before I fixed the rebase?

[1rneh]

I reviewed the commits referencing "Bug 2040381":

Can we split this into two PRs - one for the UI and one for the telemetry? If this were to land upstream, the telemetry would need its own bug, since a data steward's review happens on that bug. Once it exists, you'd reference it under bugs and link the data-review response comment under data_reviews in the event config. We'll leave data_reviews as TBD for now since we don't perform data reviews at the moment, but it would help if bugs already listed a dedicated bug.

Bonus: I have only two minor comments on the UI changes, so this can already land.

[jporter-dev]

Can we split this into two PRs - one for the UI and one for the telemetry? If this were to land upstream, the telemetry would need its own bug, since a data steward's review happens on that bug. Once it exists, you'd reference it under bugs and link the data-review response comment under data_reviews in the event config. We'll leave data_reviews as TBD for now since we don't perform data reviews at the moment, but it would help if bugs already listed a dedicated bug.

Thank you for pointing this out, splitting telemetry off makes sense with this context. I've removed the telemetry changes from this PR and created #1121 for adding it.