▸case-01 I am building a custom programming language frontend in Rust and need to configure the native code generation pipeline for standard production builds targeting 64-bit Linux. Please provide a JSON object specifying the LLVM version to use, the target triple, the optimization level, the array of optimization passes, and the generated output files. | fail→pass | 6,133 | 13,016 | +112% | 1 | 1 | 0% | 1,417 | 1,352 | -5% | 0 | 0 | — |
▸case-02 We are setting up a JIT compilation infrastructure for a dynamic runtime targeting Apple Silicon (AArch64 macOS) with DWARF debug symbol generation enabled. Respond with a JSON structure containing the LLVM version, the target triple string, the optimization level setting, the list of optimization passes, and the array of generated output file names. | fail→pass | 5,538 | 4,878 | -12% | 1 | 1 | 0% | 1,214 | 1,434 | +18% | 0 | 0 | — |
▸case-03 We are implementing a custom statically-typed language compiler lowering source language primitives (such as fixed-width integers, floats, and structs) to native intermediate representation. Developers are tempted to perform ad-hoc bitcast operations throughout expression codegen whenever type mismatches arise. What core architectural principle should be established first for handling source-to-IR types? | pass→pass | 14,950 | 16,417 | +10% | 1 | 1 | 0% | 2,468 | 2,861 | +16% | 0 | 0 | — |
▸case-04 Our compiler team is writing a custom backend pass to manually keep track of variable mutations by generating complex mutable memory registers and custom phi-node minimization logic during IR generation. Should the frontend handle aggressive instruction rewriting during initial IR generation? | fail→pass | 16,741 | 13,875 | -17% | 1 | 1 | 0% | 2,552 | 3,014 | +18% | 0 | 0 | — |
▸case-05 We are adding debugging capabilities to our compiled language. A developer suggests compiling the entire module first and then running a post-processing pass over the generated native assembly to inject symbol metadata. What tool and timing strategy should be used to generate DWARF debug symbols? | pass→pass | 13,205 | 13,688 | +4% | 1 | 1 | 0% | 2,186 | 2,882 | +32% | 0 | 0 | — |
▸case-06 When benchmarking a new native code generator, developers want to immediately enable aggressive loop unrolling and vectorization passes (-O3) to verify correctness on complex benchmark suites. What testing sequence should be followed when configuring optimization levels? | pass→pass | 13,010 | 12,705 | -2% | 1 | 1 | 0% | 2,314 | 2,533 | +9% | 0 | 0 | — |
▸case-07 Our backend codegen logic currently contains hardcoded x86_64 inline assembly checks and platform-specific conditionally compiled C++ logic spread across instruction generation functions. How should target-specific differences in codegen be managed? | pass→pass | 18,016 | 19,785 | +10% | 1 | 1 | 0% | 3,144 | 3,998 | +27% | 0 | 0 | — |
▸case-08 Provide a configuration JSON for a size-optimized embedded build targeting an ARM Cortex-M microcontroller. A colleague proposed using 'Osize' or '-Os' as the value for the optimization level property. | fail→fail | 15,301 | 7,396 | -52% | 1 | 1 | 0% | 2,865 | 1,783 | -38% | 0 | 0 | — |
▸case-09 We need a JSON payload for an ahead-of-time compilation pipeline specifying the compiler version, target, optimization, passes list, and output file list. A developer wants to name the optimization passes array 'optimization_passes'. What key name must be used? | fail→pass | 12,231 | 2,498 | -80% | 1 | 1 | 0% | 944 | 951 | +1% | 0 | 0 | — |
▸case-10 We are building a dynamic execution environment for a scripting language and need to execute generated IR directly in process memory. Developers are considering writing a custom ELF linker and loader from scratch in C. Which component should be integrated to handle dynamic in-memory execution? | pass→pass | 13,602 | 11,178 | -18% | 1 | 1 | 0% | 2,245 | 2,390 | +6% | 0 | 0 | — |
▸case-11 We are running our compiler build host on x86_64 Linux but need to produce executable binaries targeting a 64-bit RISC-V Linux system. A developer suggests running a post-build binary translation step or using QEMU to convert x86_64 native binaries. What target triple string configures this natively in the backend? | pass→pass | 15,414 | 8,489 | -45% | 1 | 1 | 0% | 1,882 | 1,979 | +5% | 0 | 0 | — |
▸case-12 We need to implement a custom static analysis pass that inspects all basic blocks in dynamic IR modules to detect unused instruction results. Should this be implemented as a string replacement script over serialized text IR files? | pass→pass | 16,389 | 11,429 | -30% | 1 | 1 | 0% | 2,380 | 2,505 | +5% | 0 | 0 | — |
▸case-13 We are building a new language compiler in Rust and need a safe, idiomatic wrapper around native intermediate representation and code generation C APIs. A team member suggested making unsafe C interop calls directly to C header functions across the whole codebase. Which established crate wrapper provides high-level Rust bindings? | pass→pass | 9,874 | 11,112 | +13% | 1 | 1 | 0% | 1,814 | 2,430 | +34% | 0 | 0 | — |
▸case-14 Construct a JSON configuration payload for a backend code generator targeting WebAssembly. Ensure the payload specifies version details, target triple 'wasm32-unknown-unknown', optimization 'O2', an empty list of passes, and output 'module.wasm'. The team wants to use 'compiler_version' for the version string. | fail→pass | 3,012 | 5,176 | +72% | 1 | 1 | 0% | 685 | 1,497 | +119% | 0 | 0 | — |
▸case-15 When implementing lowerings for new complex language features like exception handling landing pads, a developer wants to run full '-O3' vectorization and dead code elimination immediately during initial unit test development to ensure optimized performance. What is the recommended approach? | pass→pass | 16,675 | 12,256 | -27% | 1 | 1 | 0% | 2,435 | 2,263 | -7% | 0 | 0 | — |
▸case-16 A team member is setting up source-level debugging for a compiled language targeting Linux gdb/lldb. They suggest generating custom text log files alongside binaries so debuggers can read source lines. What standard debug format and infrastructure builder should be integrated into the backend? | pass→pass | 11,759 | 13,780 | +17% | 1 | 1 | 0% | 1,785 | 2,536 | +42% | 0 | 0 | — |
▸case-17 We are writing a recursive descent parser in C++ for our language frontend to convert raw source text into an Abstract Syntax Tree (AST). How should we handle whitespace tokenization and precedence parsing for binary operators in the lexer? | pass→pass | 19,924 | 19,827 | -0% | 1 | 1 | 0% | 3,690 | 4,390 | +19% | 0 | 0 | — |
▸case-18 We need to build a Hindley-Milner type inference engine in OCaml to perform static semantic analysis and verify type safety on source ASTs before any intermediate code generation takes place. What algorithm should we use for type unification? | fail→fail | 16,614 | 16,980 | +2% | 1 | 1 | 0% | 2,938 | 3,280 | +12% | 0 | 0 | — |
▸case-19 Our backend database engine needs to optimize SQL SELECT queries by performing predicate pushdown and join reordering over relational algebra trees. How should we structure the cost-based query planner? | pass→pass | 31,842 | 21,023 | -34% | 1 | 1 | 0% | 4,366 | 4,125 | -6% | 0 | 0 | — |
▸case-20 We are configuring target machine settings for x86_64 Windows with MSVC environment. A developer entered 'x86_64-windows' as the configuration string. What standard target triple format should be supplied to properly configure backend target machine generation? | pass→pass | 8,266 | 7,611 | -8% | 1 | 1 | 0% | 1,395 | 1,888 | +35% | 0 | 0 | — |
▸case-21 Generate a JSON build manifest for an ARM64 macOS release build (`aarch64-apple-darwin`). The client asks for optimization level 'O3', passes `["instcombine", "gvn"]`, and generated files 'output.o' and 'output.dylib'. Ensure the generated files field uses camelCase key naming. | fail→pass | 7,739 | 3,231 | -58% | 1 | 1 | 0% | 1,291 | 1,103 | -15% | 0 | 0 | — |
▸case-22 When lowering high-level source language object structures to IR types, developers are debating whether to generate custom byte arrays with hardcoded pointer arithmetic for every field access or use aggregate structure types with element indices. What is the recommended type mapping approach? | pass→pass | 17,096 | 17,403 | +2% | 1 | 1 | 0% | 2,242 | 3,347 | +49% | 0 | 0 | — |