7 commits

Author	SHA1	Message	Date
Snider	4669cc503d	refactor: replace fmt.Errorf/errors.New with coreerr.E() ... Co-Authored-By: Virgil <virgil@lethean.io>	2026-03-16 21:08:52 +00:00
Claude	72120bb200	feat: pass --parallel N to llama-server for concurrent inference slots ... Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 23:13:19 +00:00
Claude	b7342ec819	fix: only retry startServer on process exit, not timeout ... Distinguishes retryable failures (process exited, e.g. port conflict) from non-retryable ones (60s timeout, e.g. stuck server). Avoids 3x timeout penalty. Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 21:43:06 +00:00
Claude	c50a8e9e9b	feat: retry port selection in startServer on process failure ... Co-Authored-By: Virgil <virgil@lethean.io> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-02-19 21:40:05 +00:00
Claude	c07f37afe9	fix: guard nil exitErr wrapping, document concurrency invariant ... Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 21:38:01 +00:00
Claude	2c4966e652	feat: detect server crash before Generate/Chat calls ... Co-Authored-By: Virgil <virgil@lethean.io>	2026-02-19 21:34:46 +00:00
Claude	9aa7f624ba	feat: server lifecycle and helpers for llama-server subprocess ... Adds server.go with the process lifecycle layer that manages spawning llama-server, waiting for readiness, and graceful shutdown. Includes three helper functions (findLlamaServer, freePort, serverEnv) and the full startServer/waitReady/stop lifecycle. The serverEnv function critically filters HIP_VISIBLE_DEVICES to mask the Ryzen 9 iGPU which crashes llama-server if not excluded. Co-Authored-By: Virgil <virgil@lethean.io> Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-02-19 21:08:07 +00:00