Repository navigation
Fat-jar policy and status: jllama ships no fat jars since its modular natives - #50
Merged
Merged
Conversation
… natives java-llama.cpp 5.2.0 ships the GPU backends as additive module jars next to the platform's library jar and attaches only the Maven artifacts to its GitHub Release; the all-backends fat jars, the CPU fat jar and the agent's core-less jar are gone, and with them jllama's copy of sign-fatjars.sh. - policies/fat-jar-release-assets.md: jllama's row, the smoke table (smoke-natives / smoke-natives-macos / smoke-agent.sh on the smoke sets), the agent paragraph, the sign-fatjars.sh notes (srcmorph's alone now) and the classifier list note for srcmorph's bump (cpu-windows-x86, msvc-windows-*, sycl-fp16/fp32 are gone; a GPU classifier is one module). - crossrepostatus.md: the shared-file list, the bytecode-floor placement and the release-asset smoke paragraph. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AytmJF9faEiQEVt6eetQS2
7 of 8 tasks
Merged
5 of 6 tasks
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
policies/fat-jar-release-assets.md: java-llama.cpp 5.2.0 ships the GPU backends as additive module jars next to the platform's library jar and attaches only the Maven artifacts to its GitHub Release. Its row in the per-repo table says so (no fat jar; replacementsjbang --deps …andexamples/server/pom.xml), the smoke table namessmoke-natives/smoke-natives-macos/smoke-agent.shon the smoke sets, the agent paragraph describes the thin jar on Central, and thesign-fatjars.shnotes record that the script is srcmorph's alone now.net.ladenthin:llamabump has to drop (cpu-windows-x86,msvc-windows-*,sycl-fp16/sycl-fp32) and that a GPU classifier is one module jar needing the platform's library jar.crossrepostatus.md: the shared-file list, the bytecode-floor placement and the release-asset smoke paragraph follow.Companion to the java-llama.cpp PR of the same branch (modular natives).
Test plan
check-natives.py,publish.yml).Related issues / PRs
Refs the java-llama.cpp PR "Modular natives" on
claude/hopeful-pascal-9jlbqb.Checklist
CONTRIBUTING.mdandCODE_OF_CONDUCT.md🤖 Generated with Claude Code
https://claude.ai/code/session_01AytmJF9faEiQEVt6eetQS2
Generated by Claude Code