fix(gateway): silence idle stream duplicate warning

[KoNit-K]

What does this PR do?

This PR stops the gateway from emitting a possible duplicate send warning merely because a stream consumer object exists. It keeps the diagnostic when the turn has actual streaming-delivery evidence that was not reconciled with the normal final send.

Related Issue

Fixes #110590

Type of Change

• ✅ Bug fix (non-breaking change that fixes an issue)

Changes Made

• gateway/run_turn.py — require delivery evidence before logging the duplicate-send diagnostic.
• tests/gateway/test_stream_delivery_warning.py — cover the all-false Telegram-like consumer and retain coverage for an unconfirmed preview mismatch.

How to Test

• scripts/run_tests.sh tests/gateway/test_stream_delivery_warning.py — 2 passed after the fix; the idle-consumer assertion failed on origin/main while the control warning assertion passed.
• scripts/run_tests.sh tests/gateway/test_stream_delivery_warning.py tests/gateway/test_duplicate_reply_suppression.py — 11 passed, including existing final-delivery suppression coverage.
• Verify an idle consumer with streamed=False, previewed=False, and content_delivered=False produces no possible duplicate send warning.
• Verify an unconfirmed preview remains observable through the existing warning, preserving the mismatch diagnostic.
• scripts/run_tests.sh tests/gateway/test_stream_delivery_warning.py
• scripts/run_tests.sh tests/gateway/test_stream_delivery_warning.py tests/gateway/test_duplicate_reply_suppression.py

Evidence

• BEFORE: the focused probe was RED on origin/main before this change
• AFTER: - [x] scripts/run_tests.sh tests/gateway/test_stream_delivery_warning.py — 2 passed after the fix; the idle-consumer assertion failed on origin/main while the control warning assertion passed. — 2 passed
• CONTROL: neighboring paths listed in How to Test still pass

Checklist

Code

• ✅ I've read the Contributing Guide
• ✅ My commit messages follow Conventional Commits
• ✅ I searched for existing PRs to make sure this isn't a duplicate
• ✅ My PR contains only changes related to this fix
• ✅ I've run relevant tests locally (see How to Test)
• ✅ I've added tests for my changes
• ✅ I've tested on my platform: macOS

Documentation & Housekeeping

• ✅ Documentation update: N/A unless noted in Changes Made
• ✅ cli-config.yaml.example: N/A
• ✅ CONTRIBUTING.md or AGENTS.md: N/A
• ✅ Cross-platform impact considered
• ✅ Tool descriptions/schemas: N/A

[alt-glitch]

This was generated by AI during triage.

Duplicate of #103160 — that open, earlier PR changes the same elif _sc is not None branch in gateway/run_turn.py to the same streamed or previewed or content_delivered gate (and downgrades the no-signal case to info). #105446 addresses the same false positive via an interim-only consumer flag. Consider consolidating on one of the existing PRs.

[kvnloo]

CHECK: subsumed by KEEP #111296 for #110590 / #111281.

Exact-head cross-link: this PR silences the idle-consumer duplicate-send WARNING when no delivery evidence exists. KEEP #111296 head dc714eac2136eac2d7eebdfe9535b348528bbaa4 is the narrower preview-only gate that matches the #111281 RCA (only _previewed without a confirming send is real duplicate risk). Prefer close-or-rebase after #111296 rather than merging both into gateway/run_turn.py.

effort: factory-easy
Abort if HEAD != 0111a3c