# Make the report build work again

`build.sh` is supposed to regenerate `dist/report.txt` and then run the project's
own test suite. Right now it does neither:

```
$ bash build.sh
build.sh: line 5: OUTDIR: unbound variable
```

Nothing in the repository is sacred except the inputs and the tests. Fix the
build so that `bash build.sh`, run from the workspace root, exits `0` and
produces:

- `dist/report.txt`
- `build/gen.log`

Both files must contain the same text: one `name=score` line per record from
`data/input.txt`, in input order, followed by a single `mean=<value>` line where
the value is the arithmetic mean of the scores printed with exactly two decimal
places.

For the shipped `data/input.txt` the expected content is:

```
alpha=10
beta=20
gamma=31
delta=40
epsilon=55
mean=31.20
```

## Constraints

- `data/input.txt` must not change.
- The tests in `tests/` must not change; they must pass as written. They are run
  by `build.sh` itself, so a build that "succeeds" while skipping them is not a
  fix.
- `dist/` and `build/` are build outputs; they may be recreated freely.
- Only the Python standard library is available, and there is no network.

## Notes

- `bash build.sh` must be safe to run repeatedly.
- Do not silence failures with `|| true`; the build has to fail loudly when a
  step really fails.