test: replace fake and duplicate tests with real assertions

Fake tests rewritten to assert real behavior:
- audio_smoke: conversion tests now actually Convert() samples and verify
  output; waveform length/summary assertions tightened to exact values
- plugin_format_conversion: RowBytes/U8ToU16/LoadImageFile now call real
  production code (VideoParams::GetBytesPerPixel, sws scaler, OIIO decode
  of tests/img.png with known pixel values)
- core_color: HSV round trip now verifies fromHsv(toHsv(c)) == c instead
  of comparing toHsv against its own accessors
- core_bezier/node_inputimmediate: expected values replaced with
  independently derived constants instead of re-running the code under test
- common_commandlineparser/common_debug/common_jobtime: capture
  stdout/stderr/qDebug and assert actual output content
- proxy_manager: ProxyFinished test now drives a real proxy job instead of
  emitting the signal itself; proxy_dialog/panel/proxy/preferences/timeruler
  tests assert real widget state
- viewer_smoke/preview_autocacher/render_misc: zero-assertion tests given
  observable-state assertions or removed where nothing is observable

Duplicates removed:
- plugin_smoke_test.cpp: 18 tests duplicated from plugin_paraminstance /
  plugin_support_* / plugin_renderer_readback (751 -> 180 lines)
- module_smoke HumanStrings tests covered precisely by ui_humanstrings_test
- render_misc duplicate kDefaultInterpolation constant check

Removed by policy (skip allowed, never disabled):
- all DISABLED_ prefixes: re-enabled as real offscreen tests or deleted
- ffmpeg_decoder_hw: hardcoded personal path replaced with
  OAK_TEST_HW_DECODE_FILE env var, GTEST_SKIP when unset

Also:
- config_test: restore Config defaults after run (cross-test pollution)
- render_worker_footage: drop /tmp debug-output scaffolding
- previewaudiodevice construction test asserts the real bugfix

tests/gtest/core_bezier_test.cpp

@@ -81,8 +81,13 @@ TEST(CoreBezier, QuadraticXtoY)
 
 TEST(CoreBezier, CubicXtoT)
 {
+	// Independent expectation from the Bernstein basis: with x control
+	// values 0, 0.33, 0.66, 1.0 the curve expands to x(t) = 0.99t + 0.01t^3,
+	// and x(t) = 0.5 is solved by t = 0.5037592 (Newton-Raphson). The
+	// implementation bisects until |x(t) - x| < 1e-6 and dx/dt >= 0.99 on
+	// [0,1], so the returned t is well within 1e-5 of the true root.
 	double t = Bezier::CubicXtoT(0.5, 0.0, 0.33, 0.66, 1.0);
-	EXPECT_NEAR(t, 0.5, 0.01);
+	EXPECT_NEAR(t, 0.5037592, 1e-5);
 }
 
 TEST(CoreBezier, CubicTtoY)
@@ -98,9 +103,13 @@ TEST(CoreBezier, CubicXtoY)
 	Imath::V2d c(0.66, 1.0);
 	Imath::V2d d(1.0, 1.0);
 
+	// Independent expectation from the Bernstein basis: the x curve is
+	// x(t) = 0.99t + 0.01t^3, so x = 0.5 gives t = 0.5037592 (Newton-Raphson);
+	// the y curve is y(t) = 3(1-t)t^2 + t^3 = 3t^2 - 2t^3, which then yields
+	// y = 0.5056392. The implementation's 1e-6 bisection tolerance in x is
+	// amplified by dy/dt < 1.5, keeping the y error well under 1e-5.
 	double y = Bezier::CubicXtoY(0.5, a, b, c, d);
-	EXPECT_GE(y, 0.0);
-	EXPECT_LE(y, 1.0);
+	EXPECT_NEAR(y, 0.5056392, 1e-5);
 }
 
 TEST(CoreBezier, VectorConverters)