feat: launcher max-output env channel (BILI_LAUNCHER_MODEL_MAX_OUTPUTS) - #973
Merged
Merged
Conversation
Contributor
📦 Built Package ArtifactBranch: Option A — Install from npm PR tag (recommended)npm install -g billion-context@pr-973Each push to this PR publishes a new version under the Option B — Download artifact
tar xzf billion-context-pr973.tgz
npm install -g packageInstall instructions are refreshed on each push. |
Pure-proxy launcher clients now hand the proxy their configured per-model
max output alongside the window map:
- client-config: ModelWindow.maxOutput — codex model_max_output_tokens,
pi/omp maxTokens, opencode limit.{context,output} (object shape now
parsed), codebuddy maxOutputTokens; collectModelMaxOutputs()
- launcher: ensureProxyRunning({modelMaxOutputs}) → child env
BILI_LAUNCHER_MODEL_MAX_OUTPUTS; instance-compat check extended
- server: parse/lookup + #924 headroom rank runtime-info (#955) >
launcher env > configured > registry; instance file carries the map
Fixes #971
ranxianglei
force-pushed
the
2026-09-19_launcher-max-outputs
branch
from
September 19, 2026 01:25
6bcf29d to
0ddbe31
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #971
What
Closes the last runtime-info gap (#955/#956 covered native plugins; this covers pure-proxy launcher clients): the client's configured max output now travels the same launch-time env channel as the window map.
ModelWindow.maxOutput— codexmodel_max_output_tokens, pi/ompmaxTokens, opencodelimit.{context,output}(object shape now parsed alongside the legacy number shape), codebuddymaxOutputTokens; newcollectModelMaxOutputs()ensureProxyRunning({ modelMaxOutputs })→ child proxy envBILI_LAUNCHER_MODEL_MAX_OUTPUTS(separate env — windows JSON shape untouched); instance-compatibility check extended so a changed map restarts the proxy like windows doparseLauncherModelMaxOutputs/launcherMaxOutput+ consumed in the Codex Responses 不发送 max_output_tokens 时 output headroom 不生效，是否可 fallback 到模型 output 配置？ #924 output-headroom fallback at rank runtime-info ([native] runtime-info protocol: plugins report real model config (model / contextWindow / maxOutput) to the proxy #955) > launcher env > configured > registry; instance file carries the map for probe comparisonMatrix after this PR
Every path now reports model info: native plugins (headers + bootstrap POST), launcher-with-plugin (inherits the plugin), pure-proxy launcher (env channel: windows + max outputs).
Pre-flight
npm run typecheck✅npm test1911 tests / 1909 pass / 0 fail (2 E2E-gated skips)npm run build✅tests/headroom-output-fallback.test.ts(launcher stands in at 80k → nudge; plugin headers 180k outrank launcher 10k)maxTokensnow expect it parsed (tests/model-windows.test.ts)