Campus Task Hub

A WeChat mini program where students post and pick up small campus tasks, such as collecting parcels, sharing notes and selling used items, built by a six-member software-engineering team.

Period
Jul 2026
Context
Software engineering course project
Team
Team of 6
My role
Team lead and lead front-end developer
Status
Completed course project (July 2026)
  • WeChat Mini Program (WXML/WXSS
  • JavaScript)
  • Spring Boot REST APIs
  • JWT
  • Git/GitHub

Demo & artifacts

Live deployment on my own server of the team's web client and Spring Boot back end (July 2026 versions from the repository). The interface is in Chinese. Log in as 张三 / 1, or admin / 1 for the arbitration tools. Data is kept in memory and resets every day at 04:00 Beijing time, and no emails are sent. The mini program I built calls the same API but runs only in WeChat DevTools.

My contribution

  • Led the six-member team and split the work between front end, back end and testing.
  • Implemented more than ten core client pages: login, task listings and posting, orders, chat, wallet, reviews, profile management and real-name verification.
  • Built one shared request layer for every API call: base URL, JWT authentication, loading states, error handling, and redirects for expired sessions and unverified users. A configuration file switches between local, LAN and server environments.
  • Wrote the interface documentation, organised an API design review, and authored and ran about 70 manual test cases.

The Spring Boot back end was built by teammates. My work was the client, the interface contract between client and server, and testing. The repository's commit history under-records my share, because work was often committed by whoever finished it. I would write part of a page, a teammate would continue, and they committed the result.

How it works

Mini program request flowA page calls the shared request function. It picks the base URL for the local, LAN or server environment, attaches the JWT token and shows a loading state, then calls the Spring Boot REST API. Successful responses return data to the page; an expired session leads to the login page; an unverified user is sent to real-name verification; other failures produce one consistent error message.Page10+ core pagesShared request layer• base URL from env config(local · LAN · server)• attach JWT token• loading state• one place to read theresponse and route itSpring Boot APIbuilt by the back-end teamresponseSuccess → data to the pageExpired session → login pageUnverified user → verificationOther error → one clear message
The shared request layer. Pages call one request function instead of calling the API directly, so the base URL, the JWT token, loading state and the handling of expired sessions, unverified users and errors are written once.

Students post tasks or accept other people’s tasks; orders, chat, reviews, a wallet and real-name verification support the exchange. The client is a WeChat mini program; the back end is a Spring Boot REST API secured with JWT.

A problem I solved: integration failures between front end and back end

Integration was where things went wrong. Pages broke because the front end and the back end had assumed different data formats and different login states. Patching each page would have copied the same assumptions into more than ten places and left them free to drift apart again.

Instead of fixing each failure myself, I got both sides together to agree on conventions, wrote the interface documents, and built one request module that every page uses. It handles the base URL, JWT tokens, loading states, errors, expired sessions and unverified users in one place. I also added a configuration file so everyone could switch between local, LAN and server environments without editing pages.

The trade-off is that the request layer becomes a shared dependency: a change to it affects every page. That is why the conventions were written down in the interface documents and reviewed with the team.

Results & validation

  • The core flows of browsing, posting and accepting tasks run on the client against the team's REST API.
  • About 70 manual test cases were written and run against the integrated system.
  • Interface documentation and API design review materials were prepared for the team.
  • The web version and the back end are deployed for public testing (link above). On 19 November 2026 I checked login, the task hall, orders, the message centre and live WebSocket notifications in a browser.

All testing was manual, and the tests were not re-run for this website.

Limitations & next steps

  • Every page depends on the request layer, so it is the best candidate for automated tests (expired token, unverified user, network failure) to complement the manual cases.
  • The public demo is the web client, not the mini program. It runs on a bare IP address with a short-lived Let's Encrypt certificate that renews automatically. A domain name would give it a memorable address and would let the mini program use it too, since WeChat accepts only registered HTTPS domains for released mini programs.

Run / reproduce

Environment

  • WeChat DevTools
  • The team's Spring Boot back end running locally, on a LAN host or on a server
  • The miniprogram branch of the team repository

Steps

  1. Open the mini program in WeChat DevTools.
  2. Choose the local, LAN or server environment in the configuration file so requests reach a running back end.
  3. Log in, then browse, post and accept a task.