From 98e489eb6e1b92a51c260c2a00cff32d88e0e044 Mon Sep 17 00:00:00 2001 From: Xuan Son Nguyen Date: Fri, 14 Aug 2026 00:47:07 +0200 Subject: [PATCH] nits --- tools/server/server-context.cpp | 2 +- tools/server/server-queue.cpp | 2 +- tools/server/server-queue.h | 1 + 3 files changed, 3 insertions(+), 2 deletions(-) diff --git a/tools/server/server-context.cpp b/tools/server/server-context.cpp index d88d23337a..a7d382886a 100644 --- a/tools/server/server-context.cpp +++ b/tools/server/server-context.cpp @@ -694,7 +694,7 @@ struct server_slot { // caller need to update prompt.tokens after a successful call to keep track of the processing progress // note: this is not a member of server_slot because we want to run it inside yield_to_queue // slot is passed as const to avoid accidental modification of the slot state -// only llama_context / mtmd_context / mbatch are allowed to be used inside +// some pointers are allowed to be used, they are not used by to_json() static int process_mtmd_chunk(const server_slot & slot, mtmd::batch_ptr & mbatch, size_t idx, size_t & n_tokens_out) { GGML_ASSERT(slot.mctx); const auto & mctx = slot.mctx; diff --git a/tools/server/server-queue.cpp b/tools/server/server-queue.cpp index 588ff8d9bd..d22527ffec 100644 --- a/tools/server/server-queue.cpp +++ b/tools/server/server-queue.cpp @@ -192,7 +192,7 @@ void server_queue::worker_stop() { } void server_queue::yield_to_queue(std::function && work) { - GGML_ASSERT(worker.thread.joinable() && "yield_to_queue() must be called from the start_loop() thread"); + GGML_ASSERT(worker.thread.joinable() && "yield_to_queue() requires start_loop() to be running"); QUE_DBG("%s", "yielding to queue\n"); diff --git a/tools/server/server-queue.h b/tools/server/server-queue.h index f154262067..69a3e554f5 100644 --- a/tools/server/server-queue.h +++ b/tools/server/server-queue.h @@ -4,6 +4,7 @@ #include #include +#include #include #include #include