Skip to content

Upgrade llama.cpp from b9842 to b9859#290

Merged
bernardladenthin merged 1 commit into
mainfrom
claude/amazing-franklin-gwgced
Jul 2, 2026
Merged

Upgrade llama.cpp from b9842 to b9859#290
bernardladenthin merged 1 commit into
mainfrom
claude/amazing-franklin-gwgced

Conversation

@bernardladenthin

Copy link
Copy Markdown
Owner

Summary

  • Upgrade pinned llama.cpp version from b9842 to b9859 across all configuration files and documentation
  • Update CMakeLists.txt, CLAUDE.md, README.md, and breaking-changes documentation
  • Add detailed changelog entries documenting upstream changes in this range (IPv6 URL handling, hf-split fixes, backend optimizations, WebUI updates)

Test plan

  • CI is green on this branch
  • Docs / CHANGELOG updated where applicable

Related issues / PRs

N/A

Checklist

  • I have read CONTRIBUTING.md and CODE_OF_CONDUCT.md
  • My commits follow Conventional Commits
  • No security-sensitive changes

https://claude.ai/code/session_019pnrHT4MP55QV8MWTo2pf1

Bumps the pinned llama.cpp tag in llama/CMakeLists.txt (GIT_TAG +
-DLLAMA_TAG), the README badge/link, and the CLAUDE.md version
references.

The b9842..b9859 range touches no header the project binds; it is
almost entirely upstream backend TUs (CPU nvfp4 vec_dot, CUDA fattn
head-dim-512 + get_rows_back grid clamp, OpenCL Q1_0 + Adreno binary
kernels, Vulkan Asahi, WebGPU NVFP4, Hexagon FA rework) plus internal
common/arg.cpp (hf-split primary) and common/http.h (IPv6 host
bracketing) changes — all compiled inside upstream TUs. No jllama.cpp
change required.

All four patches re-verified to apply cleanly against b9859 via
git apply --check over the fetched b9859 sources: the only changed
patch-target file is common/arg.cpp, whose b9859 edit
(common_models_handler_apply) is disjoint from patch 0001's
make_utf8_argv/common_params_parse hunks; 0002/0003/0004 target files
untouched in this range. OuteTTS generator anchors hold (tools/tts/tts.cpp
unchanged). Appends the upgrade rows to
docs/history/llama-cpp-breaking-changes.md.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019pnrHT4MP55QV8MWTo2pf1
@bernardladenthin bernardladenthin merged commit 68332c1 into main Jul 2, 2026
9 of 11 checks passed
@bernardladenthin bernardladenthin deleted the claude/amazing-franklin-gwgced branch July 2, 2026 09:54
@claude

claude Bot commented Jul 2, 2026

Copy link
Copy Markdown

✅ Code review complete.

Summary: This PR upgrades llama.cpp from b9842 to b9859 with thorough, consistent updates across all configuration and documentation files.

Findings:

  • ✅ All version references updated consistently (CMakeLists.txt, CLAUDE.md, README.md, breaking-changes doc)
  • ✅ No code changes needed (all changes in b9842→b9859 range are backend-internal or already-compiled TUs)
  • ✅ All four local patches (0001–0004) verified to apply cleanly against b9859
  • ✅ OuteTTS generator anchors unchanged; WebUI auto-follows the pinned tag
  • ✅ Comprehensive changelog entries document IPv6 URL handling, hf-split fixes, and backend optimizations
  • ✅ Commit message follows Conventional Commits with clear rationale
  • ✅ No formatting or typo issues

Verdict: Ready to merge. The upgrade is low-risk and well-documented.

@sonarqubecloud

sonarqubecloud Bot commented Jul 2, 2026

Copy link
Copy Markdown

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants