Repository navigation
[PR] release v0.1.203 - #4122
Conversation
(ai: vllm-qwen/qwen3.8-27b) 查重:本条是 v0.1.203 的「当前」发版 PR,不重复。但发现一个必须处理的重复项:存在第二个同版本发版 PR #4018(head
标签:pr + scope/ci + p3(发版流水线的例行 patch;非缺陷、无可用性影响,故不打 bug / 不打客户端标)。 审计范围与方法(§②b 版本窗口回归审计):本次窗口 = v0.1.202 → f74626a(PR#4122 head)= 486 提交 / 353 文件 / +42108 −2256。逐段通读整窗(不整吞 diff、不开子 agent 分片):先 正在做高危域精读,结构化八问 + 可合/拦判定稍后给。 |
📦 Built Package ArtifactBranch: Option A — Install this exact build (recommended)npm install -g billion-context@0.0.4922-pr.4122.9e90c77-a2This exact version is immutable and reproducible. The rolling Option B — Download artifact
tar xzf billion-context-pr4122.tgz
npm install -g packageInstall instructions are refreshed on each push. |
Auto-fallback from the one-click release workflow: direct push to master was blocked by branch protection. Merging this PR publishes via the standard release.yml flow.
Changes since v0.1.202
pi-host-stubis missing theCONFIG_DIR_NAMEexport, triggering animport-is-undefinedwarning. | 构建时 pi-host-stub 缺少 CONFIG_DIR_NAME 导出,产生 import-is-undefined 警告 #3936 -> feat(web): session detail compact structure tab — tier tree, graph view and lazy originals | feat(web): session detail 压缩结构 tab — tier tree, graph view and lazy originals #3917)dsh.bundle.patch.yml禁用自动压缩在 web profile 下静默失效 #1772 bare-host case to the Compression conflict banner (bili v0.1.201, node v24.14.0, win32-x64) #3806 overlay-latch gate ([PR] fix: stop takeover-broken banner false alarm from agentless sidecar lanes (#3651) #3653)code_run,RUN_CODE_TOOL_NAMEhardcoded to `run_ | [BUG] 0.1.202 DSH 0.2.x code-mode 通道:omnibus 实际名为 code_run,RUN_CODE_TOOL_NAME 硬编码 run_code ⇒ 压缩面永久压制(#3603 修复复发) #3884)code_runreads as compress-reachable ([BUG] 0.1.202 DSH 0.2.x code-mode channel: omnibus is actually namedcode_run,RUN_CODE_TOOL_NAMEhardcoded to `run_ | [BUG] 0.1.202 DSH 0.2.x code-mode 通道:omnibus 实际名为 code_run,RUN_CODE_TOOL_NAME 硬编码 run_code ⇒ 压缩面永久压制(#3603 修复复发) #3884)outputBudgetHighWaterrestored as-is into GLM request → upstream 400 invalidmax_tokensparam | [bug] DSH :跨模型的 outputBudgetHighWater 被原样还原进 GLM 请求 → 上游 400 max_tokens参数非法:限制数值范围[1,131072](restore 与 clamp 都不受"已知输出上限"约束) #3833 ceiling-clamp changeoutputBudgetHighWaterrestored as-is into GLM request → upstream 400 invalidmax_tokensparam | [bug] DSH :跨模型的 outputBudgetHighWater 被原样还原进 GLM 请求 → 上游 400 max_tokens参数非法:限制数值范围[1,131072](restore 与 clamp 都不受"已知输出上限"约束) #3833)