Description
If you haven’t already, check out our contributing guidelines for onboarding and email contributors@expensify.com to request to join our Slack channel!
Version Number: 9.2.38-0
Reproducible in staging?: Y
Reproducible in production?: N
If this was caught during regression testing, add the test name, ID and link from TestRail: https://test-management.browserstack.com/projects/2219752/test-runs/TR-1472/folder/13176685/41236613/972473565
Email or phone of affected tester (no customers): ibellicotest+1953@gmail.com
Issue reported by: Applause - Internal Team
Action Performed:
Prerequisite: Have two accounts opened on two different devices or environments
- Open the Expensify app
- Log in as User A
- As User B, open any private room or create one if necessary
- Invite User A to room conversation
- As User A, navigate to room
- Tap on header and leave the room
- Return to LHN
- Note that now, an unresponsive chat that displays a "No activity yet" message on preview, is displayed instead of room
Expected Result:
When leaving a room conversation as member, chat should automatically disappear from LHN
Actual Result:
When leaving a room conversation as member, an unresponsive chat with "No activity yet" message as preview, remains displayed on "Inbox"
Workaround:
Unknown
Platforms:
Select the officially supported platforms where the issue was reproduced:
- Android: App
- Android: mWeb Chrome
- iOS: App
- iOS: mWeb Safari
- iOS: mWeb Chrome
- Windows: Chrome
- MacOS: Chrome / Safari
- MacOS: Desktop
Platforms Tested:
On which of our officially supported platforms was this issue tested:- Android: App
- Android: mWeb Chrome
- iOS: App
- iOS: mWeb Safari
- iOS: mWeb Chrome
- Windows: Chrome
- MacOS: Chrome / Safari
- MacOS: Desktop
Activity
melvin-bot commentedon Oct 24, 2025
Triggered auto assignment to @JmillsExpensify (
Bug), see https://stackoverflow.com/c/expensify/questions/14418 for more details. Please add this bug to a GH project, as outlined in the SO.melvin-bot commentedon Oct 24, 2025
Triggered auto assignment to @JS00001 (
DeployBlockerCash), see https://stackoverflowteams.com/c/expensify/questions/9980/ for more details.melvin-bot commentedon Oct 24, 2025
💬 A slack conversation has been started in #expensify-open-source
github-actions commentedon Oct 24, 2025
👋 Friendly reminder that deploy blockers are time-sensitive ⏱ issues! Check out the open `StagingDeployCash` deploy checklist to see the list of PRs included in this release, then work quickly to do one of the following:
ikevin127 commentedon Oct 25, 2025
🔄 I ran Vector1 on this deploy blocker issue and here are the top 3 possible offending PRs based on semantic similarity2 and reranking3:
Please drop a 👍 (correct) or 👎 (incorrect) as feedback regarding the results (will use to fine-tune the reranking model).
Note
PR # and name added as text only to avoid triggering regression automation on the issues which the above PRs are addressing in case the finding is incorrect.
Footnotes
Vector is currently a local-only prototype stage app that uses AI embeddings to find offending PRs for a given deploy blocker issue. Because of the early stage of the app, results must be taken with with a grain of🧂. How it works: grab deploy blocker issue's OP data (title, steps, expected / actual result) and embed -> grab all deploy checklist issue PRs data (title, explanation of change, tests, changed files, AI diff summary) and embed -> calculate cosine similarity of issue embedding to each PRs embedding -> get results -> perform reranking -> get final results. ↩
Semantic similarity, when applied to embeddings, measures the similarity in meaning between different pieces of text by comparing their vector representations in a high-dimensional "semantic space". Texts with similar meanings are mapped to vectors that are close together, and this "closeness" is quantified using metrics like cosine similarity. This technology enables search engines and chatbots to find relevant information even without exact keyword matches, providing more accurate and contextualized results. ↩
Reranking is using a machine learning model that refines the order of search results from an initial retrieval system to improve relevance. It is used as a second, more computationally intensive stage in a search pipeline to ensure that the most accurate and contextually relevant documents are shown first. ↩
JS00001 commentedon Oct 27, 2025
This is due to #71828
JS00001 commentedon Oct 27, 2025
Fixed by revert
[-]LHN - Unresponsive chat with "No activity yet" message seen on LHN when member leaves room[/-][+][Due for payment 2025-11-04] LHN - Unresponsive chat with "No activity yet" message seen on LHN when member leaves room[/+]melvin-bot commentedon Oct 28, 2025
The solution for this issue has been 🚀 deployed to production 🚀 in version 9.2.39-3 and is now subject to a 7-day regression period 📆. Here is the list of pull requests that resolve this issue:
If no regressions arise, payment will be issued on 2025-11-04. 🎊
melvin-bot commentedon Oct 28, 2025
@JS00001 @JmillsExpensify The PR fixing this issue has been merged! The following checklist (instructions) will need to be completed before the issue can be closed. Please copy/paste the BugZero Checklist from here into a new comment on this GH and complete it. If you have the K2 extension, you can simply click: [this button]