fix(cli): suppress passive cache hint under --quiet

[s0undt3ch]

Problem

The passive "your cache is big, consider pruning" hint (crates/toolr/src/main.rs, maybe_emit_cache_hint_from_argv) runs before clap parses argv and was suppressed only for TOOLR_NO_CACHE_HINT, tab-completion, and self cache …. It never honoured --quiet.

--quiet promises to "suppress non-error output", and the hint is non-error output — so under --quiet it leaked to stderr anyway, once the cache accumulated enough orphans to trip the threshold. That breaks the "--quiet must be silent" contract the mise enter-hook depends on: toolr project venv sync --quiet runs on every cd and must stay silent in non-toolr directories regardless of cache state.

It surfaced as a flaky-looking failure of project_venv_sync::sync_quiet_silently_exits_when_not_a_toolr_repo — but only on a developer machine whose real cache had crossed the orphan threshold (CI's cache is clean, so CI never saw it). The test was right; the product was wrong.

Fix

Scan argv for --quiet / -q and gate the hint on it. The scan stops at the -- separator so a wrapped command's own quiet flag (toolr project venv run -- pytest --quiet) is never mistaken for toolr's, and it matches short clusters (-dq) as well as the long form.

Tests

• Unit (argv_requests_quiet): long/short/cluster detection, absence, and the ---separator boundary in both directions.
• Integration (tests/cache_hint_quiet.rs): builds a dirty cache (12 orphans), asserts the hint fires without --quiet (so the suppression test can't pass vacuously) and is suppressed with it.

Verification

Full mise run test umbrella passes against a real dirty local cache with no XDG_CACHE_HOME isolation — 382 passed, 8 skipped — proving the fix makes the suite robust to cache state, not just the isolated case. cargo clippy clean.

Independent of #379 (deps upgrade); branches off main.

[s0undt3ch]

This change is part of the following stack:

• fix(cli): suppress passive cache hint under --quiet #380 ◀

_{Change managed by git-spice.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.8%. Comparing base (35b3d7e) to head (de97fb9).
⚠️ Report is 4 commits behind head on main.
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #380   +/-   ##
=====================================
  Coverage   93.8%   93.8%           
=====================================
  Files         96      96           
  Lines      16023   16063   +40     
  Branches     118     118           
=====================================
+ Hits       15041   15081   +40     
  Misses       972     972           
  Partials      10      10           

Flag	Coverage Δ
3.11	93.8% <100.0%> (+<0.1%)	⬆️
3.12	93.8% <100.0%> (+<0.1%)	⬆️
3.13	93.8% <100.0%> (+<0.1%)	⬆️
3.14	93.8% <100.0%> (+<0.1%)	⬆️
macos-latest	93.8% <100.0%> (+<0.1%)	⬆️
ubuntu-latest	93.8% <100.0%> (+<0.1%)	⬆️
windows-latest	90.6% <100.0%> (+<0.1%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
toolr-core	94.3% <ø> (ø)
toolr	92.9% <100.0%> (+<0.1%)	⬆️
toolr-py	96.0% <ø> (ø)
xtask	90.3% <ø> (ø)

Files with missing lines	Coverage Δ
crates/toolr/src/main.rs	99.0% <100.0%> (+0.5%)	⬆️