|
| 1 | +# GitHub Copilot Instructions |
| 2 | + |
| 3 | +## 1. Repository Context & Mission |
| 4 | + |
| 5 | +**Repository:** `HttpUserAgentParser` (mycsharp) |
| 6 | + |
| 7 | +**Primary Goal:** |
| 8 | +Provide a **high-performance, stable, and broadly compatible .NET library** for parsing HTTP User-Agent strings, including integrations for ASP.NET Core and MemoryCache. |
| 9 | + |
| 10 | +**Core Design Principles:** |
| 11 | +- API stability over convenience |
| 12 | +- Predictable performance characteristics |
| 13 | +- Minimal allocations in hot paths |
| 14 | +- Full test coverage for all observable behavior |
| 15 | + |
| 16 | +--- |
| 17 | + |
| 18 | +## 2. Repository Structure (Authoritative) |
| 19 | + |
| 20 | +Copilot must understand and respect the architectural boundaries: |
| 21 | + |
| 22 | +- `src/HttpUserAgentParser` |
| 23 | + → Core parsing logic and public APIs |
| 24 | + |
| 25 | +- `src/HttpUserAgentParser.AspNetCore` |
| 26 | + → ASP.NET Core integration (middleware, extensions) |
| 27 | + |
| 28 | +- `src/HttpUserAgentParser.MemoryCache` |
| 29 | + → Caching extensions and cache-aware abstractions |
| 30 | + |
| 31 | +- `tests/*` |
| 32 | + → Unit tests for **all** shipped packages |
| 33 | + → Tests define expected behavior and are the source of truth |
| 34 | + |
| 35 | +- `perf/*` |
| 36 | + → Benchmarks for performance-sensitive code paths |
| 37 | + |
| 38 | +--- |
| 39 | + |
| 40 | +## 3. Standard .NET CLI Commands |
| 41 | + |
| 42 | +Use these commands consistently: |
| 43 | + |
| 44 | +- Clean: |
| 45 | + `dotnet clean --nologo` |
| 46 | + |
| 47 | +- Restore: |
| 48 | + `dotnet restore` |
| 49 | + |
| 50 | +- Build: |
| 51 | + `dotnet build --nologo` |
| 52 | + |
| 53 | +- Test (all): |
| 54 | + `dotnet test --nologo` |
| 55 | + |
| 56 | +- Test (single project): |
| 57 | + `dotnet test <path-to-csproj> --nologo` |
| 58 | + |
| 59 | +--- |
| 60 | + |
| 61 | +## 4. Autonomous Execution Rules (Critical) |
| 62 | + |
| 63 | +Copilot is expected to work **independently and end-to-end** without human intervention. |
| 64 | + |
| 65 | +### Mandatory Quality Gates (Never Skip) |
| 66 | +- The solution **must compile** after every change |
| 67 | +- **All tests must pass** |
| 68 | +- New behavior **must include unit tests** |
| 69 | +- Public APIs **must not break** existing users unless explicitly intended |
| 70 | +- Changes must be **minimal, focused, and intentional** |
| 71 | + |
| 72 | +If any gate fails: |
| 73 | +1. Diagnose the root cause |
| 74 | +2. Fix the issue |
| 75 | +3. Re-run the full validation cycle |
| 76 | + |
| 77 | +--- |
| 78 | + |
| 79 | +## 5. Change Strategy & Scope Control |
| 80 | + |
| 81 | +When solving a task, Copilot should: |
| 82 | + |
| 83 | +1. **Analyze existing code first** |
| 84 | + - Prefer extension over modification |
| 85 | + - Reuse established patterns and helpers |
| 86 | +2. **Avoid architectural rewrites** |
| 87 | + - No refactors unless explicitly required |
| 88 | +3. **Preserve backward compatibility** |
| 89 | + - No breaking changes to public APIs |
| 90 | + - No silent behavioral changes |
| 91 | + |
| 92 | +If multiple solutions are possible: |
| 93 | +- Prefer the **simplest**, **most explicit**, and **least invasive** option |
| 94 | + |
| 95 | +--- |
| 96 | + |
| 97 | +## 6. Testing Requirements |
| 98 | + |
| 99 | +Every functional change must be fully tested: |
| 100 | + |
| 101 | +- Unit tests are mandatory for: |
| 102 | + - New features |
| 103 | + - Bug fixes |
| 104 | + - Edge cases and regressions |
| 105 | +- Prefer existing utilities from: |
| 106 | + `tests/HttpUserAgentParser.TestHelpers` |
| 107 | + |
| 108 | +Tests define correctness. If behavior is unclear, tests take precedence over assumptions. |
| 109 | + |
| 110 | +--- |
| 111 | + |
| 112 | +## 7. Performance Guidelines |
| 113 | + |
| 114 | +- Treat parsing logic as performance-critical |
| 115 | +- Avoid unnecessary allocations and LINQ in hot paths |
| 116 | +- Prefer spans, pooling, and cached results where appropriate |
| 117 | +- Update or add benchmarks in `perf/*` for performance-relevant changes |
| 118 | + |
| 119 | +--- |
| 120 | + |
| 121 | +## 8. Output Expectations |
| 122 | + |
| 123 | +Copilot should deliver: |
| 124 | +- Compilable, production-ready code |
| 125 | +- Complete test coverage for new behavior |
| 126 | +- Clear, intentional commits without unrelated changes |
| 127 | + |
| 128 | +**Do not stop early.** |
| 129 | +A task is only complete when **all quality gates pass** and the solution is fully validated. |
0 commit comments