fix(oracle): build the product library (not tests/benchmarks) so Infer# succeeds

[PhysShell]

Build the product library, not the whole solution

Third Dapper iteration. With .slnx matching (#23) and the shallow-clone deepen (#23) in, the build finally reached the projects — the core Dapper/Dapper.csproj and Dapper.SqlBuilder compiled fine. It then failed on the test/benchmark projects of the solution:

• MSB3021 — Dapper.Tests pulls Microsoft.Data.SqlClient.SNI, whose native-copy step is denied (/bin).
• CS1705 — the net472 benchmark (Dapper.Tests.Performance) hits a System.Runtime version skew.

Neither is the library. An Infer# leak scan wants the product code, not its test suite — and the tests/benchmarks are often unbuildable in a bare CI (no SQL Server, multi-TFM, native deps).

The fix

When no build input is given, prefer the product library over the solution:

1. a unique <repo>.csproj (repo basename) outsidetest/ tests/ benchmark* sample* example* →
2. else a lone root*.sln/*.slnx →
3. else a lone solution anywhere →
4. else the dir (continue-on-error → partial report).

Dapper's layout has no src/ (I verified against Dapper.slnx): the library is Dapper/Dapper.csproj at the repo root, which find -name "$name.csproj" selects uniquely (tests, benchmarks, and Dapper.SqlBuilder are excluded by name/path). The build input still overrides for anything unusual.

Verified

A Dapper-shaped fixture resolves to Dapper/Dapper.csproj; a name-mismatch repo falls back to the lone .slnx; an explicit build is respected. oracle.yml valid YAML · build-step bash parses · no comparator change (--selftest still 12/12).

After merge

Re-dispatch oracle (cross-tool), repo = DapperLib/Dapper, no build input — Dapper/Dapper.csproj is auto-selected, so Infer# should finally produce SARIF and we get the real three-way own / CodeQL(security-and-quality) / Infer# report.

🤖 Generated with Claude Code

---

Generated by Claude Code

Summary by CodeRabbit

• Chores

  • Improved CI/CD build target selection when no explicit build target is provided, preferring a uniquely named product-library <repo>.csproj (outside test/benchmark paths) before falling back to solutions or the target directory.

• Documentation

  • Updated the Infer# build-selection documentation to reflect the new prioritized fallback order and reaffirm that the build input overrides selection, including clarification of shallow-clone deepening behavior.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, you can upgrade your account or add credits to your account and enable them for code reviews in your settings.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 2fb11b92-f1e4-44db-91a5-508819c42a29

📥 Commits

Reviewing files that changed from the base of the PR and between b579253 and 994d39a.

📒 Files selected for processing (2)

• .github/workflows/oracle.yml
• docs/notes/oracle.md

🚧 Files skipped from review as they are similar to previous changes (2)

• docs/notes/oracle.md
• .github/workflows/oracle.yml

---

📝 Walkthrough

Walkthrough

The Infer# "Build the target" step in the Oracle workflow gains a new first-preference branch: when inputs.build is not provided, it searches for a uniquely named <repo>.csproj outside test/benchmark/sample/example directories before falling back to solution files or the target directory. The oracle documentation is updated to reflect this priority order.

Changes

Infer# build-target selection refinement

Layer / File(s)

Summary

Infer# build-target selection logic and docs .github/workflows/oracle.yml, docs/notes/oracle.md

Workflow adds a csproj-priority branch (lines 107–137) that finds a unique <repo>.csproj outside test/benchmark/sample/example paths when inputs.build is empty, falling back to unique root solutions, any unique solution, then the target directory with an updated disambiguation note. Documentation updates the fallback-behavior description to match this priority order and notes that shallow-clone deepening precedes the build.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• PhysShell/Own.NET#22: Modifies the same Infer# auto-target selection block in .github/workflows/oracle.yml and updates docs/notes/oracle.md with the same selection/failure behavior description.
• PhysShell/Own.NET#23: Also modifies the Infer# auto-build target selection logic in .github/workflows/oracle.yml, including the git fetch --unshallow --tags shallow-clone deepening step referenced in the docs.

Poem

🐇 Hop past the tests, skip the benchmarks too,
A lone .csproj is the first thing I'll pursue!
If none is unique, I'll sniff out a .sln,
And if all else fails — the target's the one.
🌿 Build smart, build clean, no ambiguity here,
The rabbit has spoken, the path is now clear!

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and specifically summarizes the main change: modifying the build target selection to prefer the product library over tests/benchmarks to ensure Infer# analysis succeeds.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/zen-pasteur-76hfs1

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.github/workflows/oracle.yml:
- Around line 123-125: The find command using -not -path filters is
case-sensitive, so paths with capitalized directory names like Tests/,
Benchmarks/, Samples/, or Examples/ will not be excluded. Replace the -path
options with -ipath in all four exclusion filters (for test, tests, benchmark,
sample, and example directories) to make the pattern matching case-insensitive,
ensuring that project directories with any capitalization variant are properly
excluded from the selection.
In `@docs/notes/oracle.md`:
- Around line 39-44: The documentation in the oracle.md file is incomplete and
does not reflect all implemented selection rules for build artifacts. Update the
paragraph describing the Infer# workflow (starting around line 39) to include
two missing behaviors: first, clarify that sample and example directory trees
are excluded from the build selection process, and second, add the
root-solution-first fallback tier to the fallback hierarchy, explaining that it
attempts to use a solution file at the repository root before falling back to
repo-wide solutions. This will ensure the documentation accurately reflects the
complete implementation.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 49a6eaf8-54f3-4c6f-85c8-6ebe1d7eca2c

📥 Commits

Reviewing files that changed from the base of the PR and between 32ad714 and b579253.

📒 Files selected for processing (2)

• .github/workflows/oracle.yml
• docs/notes/oracle.md