Repository navigation
refactor: split cognitive>25 CUDA forward/prefill functions that block whole-file stages #5022
Closed as not planned
Description
What
Split the pre-existing cognitive-complexity > 25 functions that block any whole-file stage of their files:
forward_single_cuda_with_cache(cognitive 51)forward_single_full_cuda_with_cache(31)forward_qwen3_moe_cuda(60)forward_qwen3_moe_cuda_traced_with_plan(cyclomatic 34 / cognitive 76)prefill_all_layers_gpu(29)prefill_multi_prompt(41)
Then fold #4971's first_token_logits.rs / token_logits.rs back into their home modules.
Done when
Each function is at or under the configured complexity limit with behaviour unchanged (existing tests green), and the two #4971 files are folded back.
Proposed by la-0.71.1a (inbox 2026-10-10 02:33:37Z). Refs #4971.
Activity
noahgift commentedon Oct 14, 2026
Folded into #5309 as a checklist row (TRIAGE-100, 2026-10-14): the open-ticket cap is 100, so this ticket is now tracked on that master and closed as not planned. Nothing is lost: reopen it when its work resumes, and an open PR for it reopens it.
Agent: aprender-clerk