Skip to content

unsloth: repin ggml-org#25731 to the commit that merges onto b10705 - #161

Merged
danielhanchen merged 2 commits into
masterfrom
unsloth/repin-25731
Aug 31, 2026
Merged

unsloth: repin ggml-org#25731 to the commit that merges onto b10705#161
danielhanchen merged 2 commits into
masterfrom
unsloth/repin-25731

Conversation

@danielhanchen

Copy link
Copy Markdown
Member

Overview

Repins ggml-org#25731 from edee0e160a to 44eb88e9ab, one line.

Why

The pin preflight is failing on master:

- ggml-org/llama.cpp#25731 (edee0e160a) does not merge onto b10705 + the pins before it.
##[error]pins do not merge onto the current base tag

edee0e160a no longer applies to the current base tag. The upstream conflicts have since been resolved and the PR is MERGEABLE again as of 2026-08-31T08:06Z, with head 44eb88e9ab.

What this unblocks

The preflight stops at the first conflicting pin, so a single stale entry fails the whole run and no nightly is produced. Two consequences, and the second is the one that matters:

  1. No nightly means the AMD fixes pinned by unsloth: pin the two AMD fixes so the ROCm nightlies carry them #160, #157 and #158, do not reach users despite that PR having merged. Those fix output corruption currently being reported against our own GGUFs on gfx1151, including lemonade-sdk/lemonade#3160.
  2. Because the run aborted at #25731, those two pins were never dry-run merged. They are unverified rather than known good. ggml-cuda: avoid direct ROCm_Host compute on HIP integrated GPUs (port of ggml-org#25863) #158 in particular was branched from fork master and has to merge onto an upstream tag, which is the mismatch worth watching. This PR is what lets the preflight actually reach and check them.

Supersedes #100, which repinned this same entry to a commit for base tag b10369 and is now three tags stale.

edee0e1 stopped applying to the current base tag, which failed the pin
preflight. Because the preflight stops at the first conflict, it also meant the
two AMD fixes pinned by #160 were never dry-run merged at all.

44eb88e is the head after the upstream conflicts were resolved; the PR is
MERGEABLE upstream again as of 2026-08-31T08:06Z.
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Aug 31, 2026

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review Completed 2026-08-31T08:11:56.377345Z f47b6e4 PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

#157 and #158 were branched from fork master, which is right for a small PR
diff and wrong for a pin. Fork master has diverged from upstream (it deletes a
dozen upstream workflow files), so merging those commits onto the base tag
dragged the whole divergence in and conflicted modify/delete on twelve
.github/workflows files. That is what exited the preflight with no message
after 'ok #137'.

The pins that work are branched from an upstream commit that is an ancestor of
both master and the base tag, so they carry only their own delta. Both branches
are now rebased onto 11cd988 on that pattern, and verified locally to merge
onto b10705 for a net +35/-3 across 4 files, which is the two fixes and nothing
else.
@danielhanchen
danielhanchen merged commit 982782a into master Aug 31, 2026
4 of 5 checks passed
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.

1 participant