61 lines
4.5 KiB
Markdown
# Source integrity and validation notes
| |||
| |||
This document distinguishes the checks performed while preparing the repository from earlier MetaTrader runtime observations. It is not a trading-performance report.
| |||
| |||
## Source provenance
| |||
| |||
The baseline contains the existing **Smart Liquidity Hunter v1.11** source distributed through:
| |||
| |||
- [MT4 — MQL5 Code Base 79545](https://www.mql5.com/en/code/79545)
| |||
- [MT5 — MQL5 Code Base 79546](https://www.mql5.com/en/code/79546)
| |||
| |||
During package preparation on **15 November 2026**, both public source downloads were compared with the corresponding repository files. Each file contains 21,713 bytes and has SHA-256:
| |||
| |||
```text
| |||
e3a686f1df2c2e6ff39b8b6401417d2bf688214611073923934eb5e5ed18d4cc
| |||
```
| |||
| |||
The source files are intentionally byte-identical. Conditional compiler directives provide the platform-specific sections. They use ASCII-compatible UTF-8 without a byte-order mark and retain their original CRLF line endings. No source modification or license change was made for this repository package.
| |||
| |||
The original MIT license is preserved in [LICENSE](../LICENSE). The two historical chart PNGs are also unchanged copies of the existing project images. File hashes for the prepared package are listed in [SHA256SUMS](../SHA256SUMS).
| |||
| |||
## Archived runtime observations
| |||
| |||
The following observations come from existing reports generated before this repository packaging work. These reports identify the dashboard as **SLH v1.11**. Times below retain the reports' UTC time basis.
| |||
| |||
The original [MT4 report](../evidence/2026-11-14/MT4-runtime.txt) and [MT5 report](../evidence/2026-11-14/MT5-runtime.txt) are included unchanged.
| |||
| |||
| Observation | MetaTrader 4 | MetaTrader 5 |
| |||
| --- | --- | --- |
| |||
| Report time, UTC | 2026-11-14 20:58:38 | 2026-11-14 20:58:40 |
| |||
| Terminal build | 1492 | 6273 |
| |||
| Chart | EURUSD H4 | EURUSD H1 |
| |||
| Sample size | 1,000 bars | 1,000 bars |
| |||
| Bullish / bearish markers counted | 10 / 7 | 8 / 9 |
| |||
| Unfinished-candle signal buffers empty | Yes | Yes |
| |||
| Reported arrow-placement errors | 0 | 0 |
| |||
| Indicator instances on the inspected chart | 1 | 1 |
| |||
| Dashboard labels counted | 4 | 4 |
| |||
| |||
The MT4 report's latest bar is dated 13 November 2026. The MT5 report uses older historical data, with a latest bar dated 21 August 2024. The samples have different dates and timeframes and cannot be used as a cross-platform equivalence test. Marker counts are observations of indicator output, not executed trades, wins, or profitability statistics.
| |||
| |||
Archived compiler logs report zero errors and zero warnings for the MT4 and MT5 source filenames. Their recorded filesystem last-write times are 2026-11-14 20:56:16 UTC and 20:56:27 UTC respectively. The logs do not embed a source hash. No fresh compilation, new runtime test, or performance backtest was performed solely for this unchanged-source repository import.
| |||
| |||
Some adjacent older visual reports identify v1.1. They are not used here as proof of v1.11 reload behaviour or multiple-instance isolation.
| |||
| |||
## Reproducing a review
| |||
| |||
Record the source hash, terminal build, broker/server, symbol including any suffix, timeframe, inputs, available history, and candle timestamps before comparing results. Preserve the same price data and calculation window for comparisons.
| |||
| |||
The following checklist is provided for future verification; it is not a claim that new tests were executed during packaging:
| |||
| |||
1. Compile each source file in its matching platform's MetaEditor and retain the result.
| |||
2. Attach the indicator with the documented defaults and enough loaded history. Confirm that the dashboard reports the expected version.
| |||
3. Check that the unfinished candle has no bullish or bearish signal arrow. Arrow-buffer values are plotted marker positions beyond the candle, not trade prices.
| |||
4. For a marked closed candle, independently inspect the already-confirmed pivot, required penetration, close back across the level, and wick/closing-position filters.
| |||
5. Compare normal new-candle updates with a clean reload using identical history and inputs. Compare by candle timestamp and restrict the comparison to the common calculation window.
| |||
6. If testing multiple instances, use distinct `DashboardInstanceId` values and non-overlapping display positions.
| |||
7. If testing notifications, distinguish normal one-new-candle updates from attachment, history resets, and multi-candle catch-up updates, which suppress historical alerts.
| |||
8. Record observed failures and the exact reproduction conditions alongside successful observations.
| |||
| |||
For the executable rules and input defaults, see the source files and [USAGE.md](USAGE.md).
|