Run the reset contract over every indicator, not only the tested ones

[mkzung]

Description

Tests/Indicators/IndicatorResetContractTests.cs, one new file. It constructs every indicator in the assembly by reflection, at three periods, and asserts two things: the same series replayed around a Reset() gives the same values and the same ready flag, and the indicator is back in its default state afterwards, using TestHelper.AssertIndicatorIsInDefaultState.

Related Issue

Closes #9726

Motivation and Context

CommonIndicatorTests<T> states eleven contracts and runs them only for an indicator with a test class deriving from it: 115 of the 202 concrete indicators. Six reset defects were merged this month out of that gap.

The case for it is not a bug count, since it is green on master. It is what happens when the six merged fixes are reverted, keeping the fixture:

Failed!  Failed: 21, Passed: 1121, Skipped: 70, Total: 1212

merged fix	cases	how it shows
#9686 advance decline watermark	12	IsReady differs at index 1, across the four indicators sharing the base
#9687 ConnorsRSI previous input	3	the value differs, first at index 4
#9688 SuperTrend previous close	1	the value differs at index 0, at period 1 only
#9694 MidPrice rolling extremes	2	the value differs at index 0, at periods 2 and 14
#9705 IntradayVwap running sums	3	IsReady true after the reset
#9706 NewHighsNewLows readiness	0	the fixture registers assets, so the untracked state never arises

Three notes on the shape, since each was a choice:

Three periods, because of #9688. SuperTrend assigned its previous close on the early return it takes while the average true range warms up, so at any period above one that return fires on the first update and the stale value never shows.

The price series never revisits a level. A repeating one caught two of the six, because a value carried across a reset is one the indicator meets again a few bars later anyway.

Anything unreachable is an Assert.Ignore carrying its reason, so the runner counts it and names it rather than passing over it. On master that is 70 cases over 18 indicators: 36 the fixture cannot construct, 18 that throw on the series, 16 where a constructor refuses that period and the ignore quotes the indicator's own message.

Two of the five that throw are worth a look on their own: AutoRegressiveIntegratedMovingAverage is #9714, and ChoppinessIndex at a period of one is #9724.

Requires Documentation Change

No.

How Has This Been Tested?

Passed!  Failed: 0, Passed: 1142, Skipped: 70, Total: 1212, Duration: 391 ms

on b0006b29a, plus the reverted-fix run above. The neighbouring fixtures are unaffected: IndicatorTests, ChoppinessIndexTests, MidPriceTests, SuperTrendTests, IntradayVwapTests, ConnorsRelativeStrengthIndexTests, AdvanceDeclineDifferenceTests and NewHighsNewLows together give 0 failed.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Refactor (non-breaking change which improves implementation)
• Performance (non-breaking change which improves performance. Please add associated performance test and results)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Non-functional change (xml comments/documentation/etc)

Checklist:

• My code follows the code style of this project.
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• All new and existing tests passed.
• My branch follows the naming convention bug-<issue#>-<description> or feature-<issue#>-<description>