How the check works
The app compares your program's console output with the exercise's expected output and shows a result bar above the output area:
- ✅ Output matches the reference — the exercise passes.
- ❌ Output differs — both the expected output and yours are listed so you can see exactly which line diverges.
The full workflow
- Open the project library and load a practice project into the workspace.
- Solve it with blocks or by editing code directly.
- Press Run and wait for compiling and execution to finish.
- Read the result bar: a match means it passes; otherwise compare the two outputs and locate the difference.
- Fix and run again until it passes.
Using reference-answer code
Practice projects can carry optional reference code, reachable through “View / Import C++ code”, and it stays editable after import. Teachers can use this two ways:
- Keep the reference to yourself and publish only the task plus expected output, so students solve it independently;
- Reveal the reference when reviewing, and compare alternative solutions side by side.
Topic cards: answering “why”
Sooner or later a learner asks: “it passes, but why does it work?” Topic cards exist for that moment:
- Content is organised by knowledge point, covering common exam topics and algorithm difficulties;
- With AI configured (Settings → AI knowledge), code explanations and error analysis become available;
- Topic content is archived, so reopening it is instant instead of repeating the request.
AI is optional — see the AI setup guide. Practice and automatic checking work without it.
FAQ
It says mismatch, but my answer looks right
Check the output format first: an extra space, a missing line break or an extra prompt line all count as a mismatch. Print exactly what the task asks for.
Why is free building never judged?
Because there is no expected output to compare against. Use a project-library exercise, or attach reference-answer code to a project.
Can I create my own exercises?
Yes. Package the statement and expected output as a practice project and, optionally, attach reference code — it can then be auto-checked.
The check fails although nothing looks wrong
A build failure produces no run output, so the check cannot pass. Fix the compile errors first (see the compile & run modes guide).
Also prepare the “why”
With AI knowledge configured, learners can get code and error explanations right where they get stuck.