[BlivionIaG]

Summary

Deletes FetchCord 3.0 rewrite code (from 5bbd936 / #229) that nothing in production calls or displays, plus tests that existed only to exercise that dead API. User-visible presence is unchanged. No new functions, flags, or files.

Every listed finding checked out. Nothing was dropped.

• Fetch.fetch and Fetch._cache. Production reads snapshot() and the returned dict (__main__.py, scripts/validate.py). Removed test_fetch_missing_field_returns_not_found and the fetch.fetch asserts in test_fetch_snapshot_first_provider_wins (snapshot asserts kept).
• FieldProvider.name and the name = assignments on the providers and test fakes. Nothing reads the attribute.
• WAIT_INTERVAL_SECONDS, the ignored interval_duration parameter on Cycle.wait, and test_cycle_wait_with_interval. test_cycle_wait_returns_promptly_when_stopped still covers the early return.
• run_command(..., shell=...) and test_run_command_with_shell. Production passes argument lists; exec_bash is the shell path. subprocess.run keeps its default shell=False.
• Unreachable GPU branch in _format_module. parse_fastfetch_json already routes GPU through _format_gpu, which stays.
• Orphan id tables fetch_cord/resources/desktop.json and windowmanager.json. get_infos callers and RESOURCE_FILES never load them. Package-data globs left as they are.
• host: block in fetchcord_cmds.yml (the host cycle displays motherboard, resolution, and system type; the report has no host row) and the Darwin font entry whose command is true (prints nothing, so it never yields a value). Linux font still ends with || true, which is a shell fallback, not that entry.

The only docstring edits are the sentences on Fetch and Cycle.wait that named the deleted methods. No other comment or docstring cleanup.

Left alone

fetch_cord/native/Windows/host.py still fetches a host value into the snapshot on Windows. Nothing displays it. Left untouched because it belongs to another contributor; possible follow-up.

Explicitly out of this PR: the Windows system_type script, the provider list in scripts/validate.py, packaging (requirements.txt, black config, MANIFEST.in, Formula), and comment/docstring trims.

Proof

This environment has Python 3.12.3 (python is not on PATH; 3.12 is in the CI test matrix). Commands from .github/workflows/ci.yml:

python3 -m pip install -e ".[dev]" — succeeded.

$ ruff check .
All checks passed!

$ ruff format --check .
44 files already formatted

$ mypy fetch_cord tests
Success: no issues found in 36 source files

$ python3 -m pytest tests/ -q --cov=fetch_cord --cov-fail-under=80
197 passed, 3 skipped, 13 warnings in 1.47s
Required test coverage of 80% reached. Total coverage: 88.26%

rg after the deletion (no matches means the symbol is gone):

$ rg -n "WAIT_INTERVAL_SECONDS|interval_duration|desktop\.json|windowmanager\.json|test_run_command_with_shell|test_cycle_wait_with_interval|test_fetch_missing_field|_cache"
pyproject.toml:80:  | \.mypy_cache

That hit is the mypy cache exclude, not Fetch._cache.

$ rg -n "def fetch\(" fetch_cord/fetch.py
73:    def fetch(self, skip: set[str] | None = None) -> dict[str, str]: ...
84:    def fetch(self, skip: set[str] | None = None) -> dict[str, str]:
118:    def fetch(self, skip: set[str] | None = None) -> dict[str, str]:
154:    def fetch(self, skip: set[str] | None = None) -> dict[str, str]:

Those are the provider methods. Fetch.fetch(self, field) is gone. No fetch.fetch( calls remain.

$ rg -n "run_command\(" 
fetch_cord/tools.py:15:def run_command(command: list[str]) -> str:
fetch_cord/fetch.py:86:            raw = run_command(
scripts/validate.py:84:            run_command(
tests/test_fetch_cord.py:564:        result = run_command(["ls", "-la"])
tests/test_fetch_cord.py:729:            run_command(["fastfetch"])
tests/test_more.py:947:        self.assertEqual(run_command(["ls", "-la"]), "ls output")

No caller passes shell=. The remaining shell=True is exec_bash.

$ rg -n "module_type == \"GPU\"|_format_gpu" fetch_cord/info.py
51:def _format_gpu(result: dict[str, Any]) -> str:
102:        if module_type == "GPU":
103:            names = [name for name in (_format_gpu(item) for item in items) if name]

GPU formatting stays on the parse_fastfetch_json path. _format_module no longer has a GPU branch.

$ rg -n "get_infos\(" 
# cpus, gpus, os, terminal, shell, motherboards, system_types only

$ rg -n -A8 "^RESOURCE_FILES" fetch_cord/constants.py
# cpus.json, gpus.json, os.json, terminal.json, shell.json,
# motherboards.json, system_types.json

$ ls fetch_cord/resources/*.json
cpus.json  gpus.json  motherboards.json  os.json  shell.json  system_types.json  terminal.json

$ rg -n "^  host:" fetch_cord/resources/fetchcord_cmds.yml
# no matches

$ ls fetch_cord/native/Windows/host.py
fetch_cord/native/Windows/host.py