Repository navigation
Pin toolchains and split CLI orchestration - #45
Merged
Merged
Conversation
CI and release jobs installed whatever `stable` was current, so Rust 1.99 silently introduced new Clippy lints and changed the compiler that builds shipped binaries. Pin 1.99.1 in rust-toolchain.toml and install that toolchain in every workflow so local, CI, and release builds match.
The release live suite depended on an unrecorded local PyTorch env, and the compiled-Triton case failed late on Python 3.12, missing NumPy, or a runtime that registers only libcuda.so.1. Add environment-pytorch.yml (Python 3.11, torch 2.3.0, Triton, NumPy 1.x), check Dynamo, Triton, and NumPy before starting the workload, and point Triton at the CUDA link stub only when no unversioned libcuda.so is registered.
Move clap arguments, command failures, CPU and syscall inventory, live collection, hidden development commands, and human rendering out of main.rs into dedicated modules with explicit imports. main.rs keeps the entry point and the public measure, validate, doctor, and discover flow. No behavior changes.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
rust-toolchain.tomland install it in CI and release jobs, so local, CI, and shipped builds use the same compiler and Clippy. Floatingstablepreviously broke CI when Rust 1.99 added lints.environment-pytorch.ymlfor thePYTORCH_ENVlive fixture (Python 3.11, torch 2.3.0, Triton, NumPy 1.x). The compiled-Triton test now checks Dynamo, Triton, and NumPy up front, and falls back to the CUDA link stub only when no unversionedlibcuda.sois registered (WSL container runtimes).xprobe/cli/src/main.rs(3962 → 866 lines) intoargs,failure,inventory,live,dev, andrendermodules with explicit imports. Line-for-line move; no behavior change.Verification
just fmt-check,just lint,just test(153 tests) on the pinned toolchainjust test-agent-contract,test-bpf-live,test-cupti-live,test-multisource-live,test-injection-live,test-nvtx-liveon RTX 3060test-pytorch-cuda-livepasses with the new env and fails fast with an explicit message on a Python 3.12 / torch 2.3 envenvironment-pytorch.ymlimports torch 2.3.0+cu121, Triton 2.3.0, NumPy 1.26.4 with Dynamo supported