Niko1221 da170f8ae7 serve: /config and /load bodies are marked read, so the close does not wait for the drain (#594)

2026-10-05 13:16:53 +02:00

..

web

Merge branch 'f-139-ux' into f-139-rel

2026-10-03 20:03:53 +02:00

__init__.py

Strata: Qwen3.8-Flash-Next on a 12-24 GB NVIDIA GPU + 64 GB RAM, one-click install

2026-09-24 18:42:20 +02:00

chat_golden.json

Strata: Qwen3.8-Flash-Next on a 12-24 GB NVIDIA GPU + 64 GB RAM, one-click install

2026-09-24 18:42:20 +02:00

chat_template.jinja

Strata: Qwen3.8-Flash-Next on a 12-24 GB NVIDIA GPU + 64 GB RAM, one-click install

2026-09-24 18:42:20 +02:00

frontend.py

fix(serve): safely handle malformed or partial tool call arguments in history

2026-10-05 13:16:53 +02:00

mcp_fake_server.py

Tools from MCP servers in the web app's chat (serve/mcp.py)

2026-09-28 17:28:01 +02:00

mcp.py

mcp: a server that stops quotes its last log line reliably (the stdout reader waited no time for the stderr reader, so the reason was sometimes missing - the flaky test_a_server_that_fails_is_left_out)

2026-09-30 22:13:46 +02:00

responses.py

serve: the OpenAI Responses API, POST /v1/responses (#451)

2026-10-03 18:35:35 +02:00

runconfig.py

web: a Model settings view for documented run config keys (#564)

2026-10-03 18:35:02 +02:00

server.py

serve: /config and /load bodies are marked read, so the close does not wait for the drain (#594)

2026-10-05 13:16:53 +02:00

structured.py

#334 follow-up: jsonschema is an optional, lazily imported package (no new hard dependency)

2026-10-01 08:27:57 +02:00

telemetry.py

serve: AMD readings in the Monitor tab from Linux sysfs; free VRAM on HIP (#301)

2026-10-01 08:41:14 +02:00

test_control_cancel.py

serve: a Stop during the prompt read goes to the engine at once (#879 abort latency)

2026-10-05 11:03:50 +02:00

test_detok.py

tokenizer: a word longer than 64 symbols merges with a heap and a linked list, not a rescan per merge (#268)

2026-09-30 20:37:45 +02:00

test_fatal_recovery.py

serve: invalidate the engine after fatal verification timeouts

2026-10-05 13:16:53 +02:00

test_frontend.py

serve: STRATA_KEEP_EMPTY_TURNS=1 restores rendering of empty assistant turns (#843)

2026-10-05 13:16:53 +02:00

test_lifecycle.py

#333 follow-up: the unload keeps its terminate/kill fallback; EngineStuck instead of a catch-all RuntimeError

2026-10-01 08:26:28 +02:00

test_mcp.py

fix(serve): report a tool call cut off by the end of the output as unfinished

2026-09-30 20:37:57 +02:00

test_monitor.py

web: a Conversation cache card in the Monitor tab (#596)

2026-10-03 18:31:20 +02:00

test_parallel.py

serve: validate stop in Responses and count_tokens; tests for stop in --batch and after a thinking budget (#454)

2026-10-05 13:16:53 +02:00

test_reasoning_tools.py

serve: a tool call inside the reasoning counts only for a declared tool (#804, strict form of #787)

2026-10-05 13:16:53 +02:00

test_responses.py

serve: /v1/responses refuses a file image from another origin (#553); a bad data: URL or unreadable file is a 400 (from #582)

2026-10-05 13:16:53 +02:00

test_runconfig.py

web: a Model settings view for documented run config keys (#564)

2026-10-03 18:35:02 +02:00

test_security.py

serve: read the /load and /unload body before replying (#630)

2026-10-03 18:12:54 +02:00

test_server.py

serve: /config and /load bodies are marked read, so the close does not wait for the drain (#594)

2026-10-05 13:16:53 +02:00

test_structured.py

#334 follow-up: jsonschema is an optional, lazily imported package (no new hard dependency)

2026-10-01 08:27:57 +02:00

test_vram.py

Hot VRAM resize, opt-in: a segmented expert cache, VRAM <reserve>, POST /v1/vram (#533)

2026-10-03 18:52:00 +02:00

test_winjob.py

serve: on Windows the engine, vision encoder and MCP servers end with the server

2026-09-30 14:01:52 +02:00

winjob.py

serve: on Windows the engine, vision encoder and MCP servers end with the server

2026-09-30 14:01:52 +02:00