Ollama: Silent Path Fixes and Runner Cleanup
Today's activity centers on correctness, not features — fixes for storage collisions, misdirected configs, and failed model imports — plus a shared-code cleanup in the MLX runner. The highest impact is reversible manifest host encoding and single-shard split GGUF imports.
Duration: PT2M37S
Episode overview
This episode is a short developer briefing from Ollama.
It explains recent repository work in plain language.
- Show: Ollama
- Published: 2026-11-22T14:00:44Z
- Audio duration: PT2M37S
Transcript excerpt
This excerpt keeps the crawler page concise. Listen to the episode or use the RSS feed for the full update.
Hello, it's Sunday, November 22nd, 2026, and this is your Ollama developer briefing.
The takeaway today is correctness over features. Most of the activity fixes places where Ollama looked in the wrong folder or mishandled a name, alongside a cleanup that removes duplicated model setup code.
First, wrong-location bugs. One fix makes version two manifest host encoding reversible, in pull request 19835. Colons were escaped but percent signs were not, so two different registry hosts could map to the same storage folder and have their manifests served under the wrong name. The change escapes percent first,…
Second, smoother setup for models and automation. Pull request 19825 fixes split model imports dating back to issue 5245. Referencing a single shard of a split G G U F file now pulls in all shards, instead of failing with a shard-count error. And pull request 19830 adds a global no-animation flag that drops spinners…
Finally, maintenance. Pull request 19805, merged as commit 3-1-8-e-e-2-8, moves shared tensor helpers and common quantization and tokenizer setup for the M L X runner into one place, with forward passes untouched and no behavior change. Docs add two community…
What's next:…
Nearby episodes from Ollama
- Inkling on Mac Gains and Stream Reliability Fixes
- Adaptive Memory on Macs, Tighter Server Security
- Tool Call Reliability and Runner Efficiency
- Runner Reuse, Decision Models, and Thinking UI Fixes
- MLX Concurrency and Memory Tuning
- GPU Load Fixes and Clearer Errors
- Weekly Recap - Parallel Inference, Tool Fidelity, and Launcher Polish
- Qwen Tool-Call Reliability Fix