Saturday, June 27, 2026
9 changes · 19.0
Resolved issues and error corrections
This fixes a test reliability issue in the Mail app by ensuring a menu is closed before sending a message during an automated public chat check. It also corrects which conversation is updated between test runs, helping prevent false failures and making future investigations easier.
Original PR description
Attempt at fixing the following race condition. It's not clear what causes it, but these changes make the test more robust and might help future investigations. discuss_channel_public_tour opens the composer "More Actions" menu to attach files but feeds the hidden file input directly, so the menu is never closed and is still open when Send is clicked. Close it and wait for it to disappear before sending, to avoid clicking Send while the dropdown is dismissing. Also fix _open_group_page_as_user, which updated the last message body of self.channel instead of self.group between the two tour runs. https://runbot.odoo.com/odoo/error/243436
This fixes an unstable automated test for retrying message loading in Discuss. It reduces false test failures caused by timing issues, helping keep the mail app release process more reliable without changing user-facing behavior.
Original PR description
The "Retry loading more messages on failed load more messages" test drove load-more by scrolling (real IntersectionObserver) and failed the fetch synchronously, then clicked retry immediately. The observer could fire the older-fetch twice and leave a second fetch in flight at the retry click, which then no-op'd (fetchMoreMessages bails while a fetch is loading), leaving 30 messages instead of 60. This is a test-timing artifact: a real user retries long after any fetch has settled. Fail the load-more through a Deferred rejected only once the fetch is in flight, like jump_to_present.test.js. While it is pending, duplicate observer fires no-op, so no orphan fetch can race the retry. https://runbot.odoo.com/odoo/error/242113
The automated website testing tool now blocks an additional Odoo Fin test URL so it does not contact live production services. This keeps dashboard tests reliable and isolated, reducing the risk of false failures or unintended external calls.
Original PR description
This commit follows up on [1] by extending the Odoo Fin request mock to cover the new version 2 (v2) favorite institutions endpoint. Previously, a mock was introduced to prevent the Clickall tool from making real external HTTP requests to `production.odoofin.com` when displaying the accounting dashboard. This update ensures that the newly introduced v2 URL is also safely intercepted, keeping the automated tests fully isolated from production servers. runbot-234936 [1] : https://github.com/odoo/odoo/commit/c6451015f1b01c3e1defe4a576989fd4bfdf2cdb
The website editor now hides the text formatting toolbar when selected content is removed from a page. This prevents a stale toolbar from remaining visible, reducing confusion while editing website snippets.
Original PR description
Commit 8df8b4f41809de75f9ca220d47274cc3166f1cd5 moved the check for selection inside `[data-prevent-closing-overlay]` to the `closeToolbar` plugin. But when this function is called from `selection_leave_handlers` it should not do that check, as it was done for the focus or click. This commit skips the check based on the selection when `closeToolbar` is called by `selection_leave_handlers`. Steps to reproduce: - Open website builder - Drop "Text Image" snippet - Select some text in the snippet - Click on the trash icon in the sidebar to remove the snippet - Bug: the toolbar is still shown task-6308364
This update makes an automated website test for clickable carousel slides more dependable by using a more predictable way to confirm a link. It helps reduce false test failures, supporting smoother validation and releases without changing the customer-facing website behavior.
Original PR description
Make the step deterministic by committing the URL with edit and pressing Tab in the input instead of relying on a dropdown click. runbot-241089
Webhook sample payload previews now handle complex field values correctly instead of failing to load. This helps users configure webhook notifications more reliably, especially on records with advanced accounting-related data structures.
Original PR description
**Steps to Reproduce:** - Create a Server Action of type 'Webhook Notification'. - Select a model containing a field that returns a `frozendict`-based structure (e.g. `account.move` →…
**Steps to Reproduce:** - Create a Server Action of type 'Webhook Notification'. - Select a model containing a field that returns a `frozendict`-based structure (e.g. `account.move` → `needed_terms`). - Add the field to the webhook fields. - Open the webhook sample payload preview. **Issue:** - During sample payload generation: - The selected fields are read from a sample record. - A selected field returns a structure containing `frozendict` objects. - The payload is serialized using `json.dumps()`. - JSON serialization fails with: ```text TypeError: keys must be str, int, float, bool or None, not frozendict ``` - The webhook sample payload computation crashes and the preview cannot be displayed. **Root Cause:** - The webhook sample payload may contain `frozendict` objects returned by selected fields. - The serializer used for payload generation does not handle such mapping-like objects, causing `json.dumps()` to fail. **Solution:** - Use a serializer that converts mapping-like objects into JSON-compatible structures before serializing the webhook sample payload. **OPW-6295777** Description of the issue/feature this PR addresses: Current behavior before PR: Desired behavior after PR is merged: --- I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr Forward-Port-Of: odoo/odoo#271864
The messaging recipient field no longer offers to create a new contact when no usable name or email has been entered. This prevents accidental creation of unnamed contacts that can clutter contact lists and cause errors when opened later.
Original PR description
The chatter recipients input always added a "Create" option to its autocomplete dropdown, even when the field held no text. The feature that introduced this input,…
The chatter recipients input always added a "Create" option to its autocomplete dropdown, even when the field held no text. The feature that introduced this input, https://github.com/odoo/odoo/commit/3b985d2c4239fe702c1a56bef04fc112af39e954, wired that option to create a res.partner straight away. Selecting "Create" with an empty input therefore created a partner with no name, displayed everywhere as "Unnamed". The email popover that opens for a recipient without an address then offered a Discard button, but the empty partner was already saved. The same sequence could be repeated to accumulate nameless contacts, which later raise a missing name error when opened in the Contacts app. Only add the "Create" option when the typed term resolves to a name or an email, in getAutoCompleteSources of recipients_input.js. An input with no usable text has nothing to create from, which matches the standard many2one autocomplete that only offers creation once text is entered. Steps to reproduce: 1. Open any record with a chatter 2. Click "Send message" 3. In the recipients field type a character then delete it, leaving the field empty with the dropdown still open 4. Click the "Create" option 5. On the "What's the email address of Unnamed?" popover, click Discard => A blank "Unnamed" contact is created, and repeating the steps creates more Ticket [link](https://www.odoo.com/odoo/action-4043/5942428) opw-5942428 --- I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr Forward-Port-Of: odoo/odoo#250973
This update prevents an internal French leave validation test from failing when no user timezone is set during automated checks. It helps keep the testing pipeline reliable without changing employee-facing leave behavior.
Original PR description
Currently the test test_holiday_in_week is failing in runbot tests This is due to the fact that during the tests the user tz in env is not set and hence it fails in the pytz library. Added a fallback to avoid the issue runbot issue opw-237623 Description of the issue/feature this PR addresses: Current behavior before PR: Desired behavior after PR is merged: --- I confirm I have signed the CLA and read the PR guidelines at www.odoo.com/submit-pr Forward-Port-Of: odoo/odoo#248941
The disconnect button for French electronic invoicing has been renamed to accurately describe the service it affects. This avoids confusion caused by the previous wording, which incorrectly referenced Peppol instead of French electronic invoicing.
Original PR description
The name of the disconnect button for the France e-invoicing was incorrect as it referenced peppol and was fixed in this pr to be called Disconnect French electronic invoicing task-6266337 Forward-Port-Of: odoo/odoo#268536