[burgholzer]

🤖 AI text below 🤖

Description

Extract nine independently useful fixes and simplifications from #2545 onto main after #2730, with one commit per change:

• Validate LLVM modules and native QIR calling conventions, ABI attributes, and direct runtime call types before JIT execution.
• Return compiler output-write failures without an LLVM stream-destructor abort.
• Find provider resources beside libraries loaded by relative path.
• Release scalar DD phase roots when interpretation fails.
• Reject truncated DD records, invalid node references, and skipped vector levels while preserving matrix level skipping.
• Remove failure propagation from infallible compiler helpers and simplify QCO yield binding.
• Avoid repeated benchmark outcome validation and probability evaluation.
• Parse benchmark specifications and compute canonical metadata once through the existing registry.
• Include pass-pipeline parser details in the owning MLIR diagnostic.

The throwing APIs, optional LLVM/MLIR build, serialization format, and QIR runtime failure policy remain unchanged. The additive benchmark parsing API is documented in its public header. The larger dependency and error-handling migrations stay in #2545; no new dependency is added.

Review retained the existing DD parser and caught an overstrict QIR attribute check: function-level alignstack remains accepted, with a positive regression. Bug regressions reproduced failures before their fixes. DD changes address record structure and non-finite weights; broader numeric-normalization work is outside this extraction.

Validation on Linux AArch64 with LLVM/MLIR 23:

• Release build and 4,004 runnable native tests passed.
• Fresh no-MLIR build and 770 runnable tests passed.
• Both native configurations skipped the existing unsupported job-ID test.
• Fresh editable package and all 60 Python DD tests passed.
• Repository lint and whole-file C++ lint passed (all 23 selected files, zero findings).

Windows and macOS require hosted CI. Codex assisted with extraction, review, tests, and this description.

Checklist

• The pull request only contains commits that are focused and relevant to this change.
• I have added appropriate tests that cover the new/changed functionality.
• I have updated the documentation to reflect these changes.
• The changes follow the project's style guidelines and introduce no new warnings.
• The changes are fully tested and pass the CI checks.
• I have reviewed my own code changes.

If PR contains AI-assisted content:

• Any agent that created, edited, or submitted GitHub content was explicitly authorized for that scope, as required by our AI Usage Guidelines.
• Every agent-authored or agent-edited public text body begins with the visible disclosure 🤖 *AI text below* 🤖 (titles are exempt).
• I have disclosed AI assistance in the PR description.
• I confirm that I have personally reviewed and understood all AI-generated content, and accept full responsibility for it.

[codecov]

Codecov Report

❌ Patch coverage is 95.29412% with 8 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/qdmi/common/DeviceConfiguration.cpp	42.8%	4 Missing ⚠️
src/bench/JSON.cpp	80.0%	2 Missing ⚠️
mlir/lib/Dialect/QIR/Execution/JIT/Session.cpp	97.2%	1 Missing ⚠️
src/dd/Serialization.cpp	97.7%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!