Skip to content

Pin the xllamacpp CPU swap to llm2's locked version - #274

Merged
WSHAPER merged 1 commit into
nextcloud:mainfrom
WSHAPER:fix/ci-pin-xllamacpp
Oct 19, 2026
Merged

WSHAPER merged 1 commit into
nextcloud:mainfrom
WSHAPER:fix/ci-pin-xllamacpp

Conversation

@WSHAPER

@WSHAPER WSHAPER commented Oct 19, 2026

Copy link
Copy Markdown
Contributor

The llm2 setup swaps the CUDA wheel for the CPU build so driverless runners can run it, but the unpinned reinstall resolves the newest PyPI wheel and bypasses llm2's poetry.lock. That is how the recent xllamacpp rename landed and broke every integration leg, see #269 and nextcloud/llm2#306.

This pins the swap to 2026.9.10809, the version llm2's lock pins, so the tested environment is the locked one again. Can be dropped once llm2 ships the binding fix (nextcloud/llm2#307).

🤖 AI (if applicable)

  • The content of this PR was partly or fully generated using AI

The unpinned reinstall resolved the newest PyPI wheel and bypassed
llm2's poetry.lock, so the xllamacpp hostname rename took down every
integration leg. Pins the swap to 2026.9.10809, the locked version,
until llm2 ships the binding fix.

Assisted-by: opencode:zai/glm-5.3
Signed-off-by: WSHAPER <42714629+WSHAPER@users.noreply.github.com>
@WSHAPER
WSHAPER merged commit 4da41f7 into nextcloud:main Oct 19, 2026
7 checks passed
@WSHAPER
WSHAPER deleted the fix/ci-pin-xllamacpp branch October 19, 2026 15:23
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