bug(fill): `fill` generates a `.meta` directory even if no tests executed

This is now noticeable as `fill` insists that the output directory is empty, cf https://github.com/ethereum/execution-spec-tests/pull/1608.

For example, no tests run as the collected tests do not run on the specified fork:

```
uv run fill --fork Frontier tests/istanbul/eip1344_chainid/
========================================================= test session starts ==========================================================
platform linux -- Python 3.11.12, pytest-7.4.4, pluggy-1.5.0
Generating fixtures for: Frontier
Only generating fixtures with stable/deployed forks: Specify an upcoming fork via --until=fork to add forks under development.
ethereum-spec-evm-resolver 0.0.5
solc: 0.8.24
EELS resolutions file: /home/dtopz/code/github/new/execution-spec-tests/eels_resolutions.json
rootdir: /home/dtopz/code/github/new/execution-spec-tests
configfile: pytest.ini
plugins: html-4.1.1, json-report-1.5.0, cov-4.1.0, xdist-3.6.1, regex-0.2.0, custom-report-1.0.1, metadata-3.1.1
collecting ... pytest-regex selected 3 tests to run for regex: .*
collected 3 items

------------ Generated html report: file:///home/dtopz/code/github/new/execution-spec-tests/fixtures/.meta/report_fill.html ------------
======================================================== no tests ran in 0.01s =========================================================
```

But a `fixtures/.meta` directory still gets created and an html report is written there. I'd consider this unexpected behavior.