Studio: support llama.cpp semantic version output - #9127
Conversation
|
@codex review |
|
Codex Review: Didn't find any major issues. 🚀 Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
for more information, see https://pre-commit.ci
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
|
@codex review |
|
Codex Review: Didn't find any major issues. More of your lovely PRs please. Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Problem
Studio uses the numeric llama.cpp build to display the installed version and to decide whether a markerless source build is older than the latest Unsloth prebuilt.
llama.cpp recently changed what
llama-server --versionprints:Studio's parser expected the build number immediately after
version:. With the new output it read the semantic version's leading0, rejected that as an unknown build, and lost the real build number10470.This affects markerless installations such as source builds and custom llama.cpp paths. Prebuilt installations are unaffected because they read their version from
UNSLOTH_PREBUILT_INFO.jsoninstead of runningllama-server --version.For an affected installation, Studio could:
Fix
build NNNNfield when it is present on theversion:line.version: NNNNformat so existing binaries continue to work.2.0.0for legacy build number 2 when no build field is present.The update-status regression now drives the real parser with semantic output and verifies that source build 10470 is not offered prebuilt b10360 as an update.
Testing
python -m pytest -q studio/backend/tests/test_llama_cpp_update.py studio/backend/tests/test_llama_cpp_freshness.py(115 passed)python -m ruff check studio/backend/utils/llama_cpp_update.py studio/backend/tests/test_llama_cpp_update.pygit diff --check