Commit Graph
4896 Commits
Author SHA1 Message Date
song fa565eafb1 fix(pricing): add image cache pricing field 2026-09-11 15:57:38 +08:00
song eacff0a629 fix: retain upstream config and image pricing updates 2026-09-11 15:57:38 +08:00
song c0d5119375 feat(openai): route OAuth image requests through native Codex Images
Use shared model routing for image forwarding and account tests, retain Luna orchestration for legacy models, and add Image 2.5 pricing and usage accounting.
2026-09-11 15:57:09 +08:00
Wesley Liddick 40cf3f50a6 Merge pull request #6968 from ouchihao/codex/issue-6903-deepseek-vision-modality
fix(codex): advertise DeepSeek vision image input
2026-09-11 14:01:36 +08:00
Wesley Liddick 75b7dd1e0b Merge pull request #6969 from ouchihao/codex/issue-5882-peak-rate-validation
fix(groups): return bad request for invalid peak-rate settings
2026-09-11 14:01:14 +08:00
Wesley Liddick 642d20b8ac Merge pull request #6988 from ZeroClover/fix/codex-manifest-display-name
fix(openai): keep Codex manifest display names in account model lists
2026-09-11 14:00:42 +08:00
Wesley Liddick 310f8b7fa2 Merge pull request #6964 from xlplbo/fix/apicompat-agent-message-chat-bridge
fix(apicompat): chat 桥保留 Codex agent_message 任务正文
2026-09-11 13:59:41 +08:00
Wesley Liddick 6254aba5e9 Merge pull request #6575 from doublepi123/feat/antigravity-gemini-3.8-flash
feat(antigravity): support Gemini 3.7 and 3.8 Flash models
2026-09-11 13:58:54 +08:00
Wesley Liddick 7145484fde Merge pull request #6971 from ouchihao/codex/issue-1890-ttft-request-details
fix(ops): show first-token latency in TTFT request details
2026-09-11 13:58:37 +08:00
Wesley Liddick fd300ab6f3 Merge pull request #6974 from alfadb/fix/deepseek-v41-flash-pricing
fix(billing): 适配 DeepSeek V4.1-Flash 新价与 pro→Flash 计费切换
2026-09-11 13:58:23 +08:00
Wesley Liddick 88011a6a2e Merge pull request #6783 from asmoyou/fix/large-image-request-memory
perf(gateway): reduce memory amplification for large Responses image requests
2026-09-11 13:58:11 +08:00
Wesley Liddick 14029e50aa Merge pull request #6661 from whhphd/fix/grok-media-slot-release
fix(grok): release abandoned media slots and preserve video ownership
2026-09-11 13:57:07 +08:00
Wesley Liddick 3fb03cdac1 Merge pull request #6960 from xlplbo/fix/openai-ws-exec-scope-preemption
fix(openai-ws): ctx_pool 会话抢占与会话状态按 codex 线程隔离
2026-09-11 13:56:52 +08:00
Wesley Liddick 8f9a9a255b Merge pull request #6965 from xlplbo/fix/openai-ws-pool-reader-loop
fix(openai-ws): 连接池常驻读循环应答上游 ping,根治空闲连接 1011 keepalive ping timeout
2026-09-11 13:56:23 +08:00
xlplbo 4543ddc5c1 fix(openai): WS 执行作用域按 request_kind 分道 2026-09-11 13:37:59 +08:00
matt 5af7e7438a fix(ci): check deferred sql.DB.Close errors in setup tests
PR #6499 merged without CI ever running on its head commit (fork run stayed
action_required), landing three bare `defer db.Close()` calls in
internal/setup/setup_test.go. The repo's errcheck config sets
disable-default-exclusions: true, so those are reported and golangci-lint has
been red on main ever since, failing every subsequent PR run.

Switch them to the codebase's existing convention
`defer func() { _ = db.Close() }()`.
2026-09-11 11:45:59 +08:00
xlplbo 613722eee4 fix(openai): HTTP 转发的执行作用域改从原始请求计算 2026-09-11 11:36:29 +08:00
xlplbo d0ca057ca4 fix(openai): WS 会话被抢占时先发关闭帧再取消 2026-09-11 10:25:50 +08:00
xlplbo e4c369bd56 fix(openai): HTTP 经 WS 池转发按执行作用域取会话状态键 2026-09-11 10:25:50 +08:00
Wesley Liddick f4dd88b001 Merge pull request #6847 from ouchihao/codex/issue-4064-custom-page-button-visibility
feat(settings): allow hiding the custom page open button
2026-09-11 09:34:11 +08:00
Wesley Liddick 2493b0dc31 Merge pull request #6924 from smyyyyy2025/fix/codex-user-agent-validation
fix: validate Codex User-Agent before identity parsing
2026-09-11 09:33:19 +08:00
Wesley Liddick 264cbbec26 Merge pull request #6499 from soulxyz/fix/setup-database-bootstrap
fix(setup): avoid hardcoded postgres connection
2026-09-11 09:32:01 +08:00
Wesley Liddick 87e01596b7 Merge pull request #6850 from reallyfeishi/fix/channel-monitor-utc-buckets
fix: anchor channel monitor date_bin buckets to UTC
2026-09-11 09:31:22 +08:00
Wesley Liddick 43f9383d40 Merge pull request #6869 from Garfielk/codex/fix-heartbeat-envelope-fields
fix(openai): accept full heartbeat bootstrap envelopes
2026-09-11 09:30:45 +08:00
Wesley Liddick 685e97a3a6 Merge pull request #6904 from m11y/fix-scheduling-threshold-snapshot
fix(scheduler): preserve threshold metadata
2026-09-11 09:30:00 +08:00
Wesley Liddick e71d291b39 Merge pull request #6917 from ouchihao/codex/issue-2149-monitor-base-path
fix(monitor): support upstream base URL paths
2026-09-11 09:29:25 +08:00
Wesley Liddick 4adcef4ff4 Merge pull request #6943 from xiuai-labs/fix/keep-client-system-cache-control
fix(gateway): 不再剥离客户端 system 上的 cache_control / stop stripping client system cache_control
2026-09-11 09:28:55 +08:00
Zero Clover 6feefb7aec fix(openai): keep Codex manifest display names in account model lists 2026-09-11 03:18:00 +08:00
alfadb ab9bd9e877 fix(billing): 适配 DeepSeek V4.1-Flash 新价与 pro→Flash 切换
- flash 低谷价更新为 $0.15/$0.60/$0.003 per MTok(2026-09-10 官方降价)
- deepseek-v4-pro 自 2026-09-14 04:00 UTC 起按 Flash 价计费(上游路由变更)
- 计费时点显式驱动 pro→Flash 切换,定价展示与计费两条路径同源
- deepseek-flash 保留 reasoning_effort=max;补 Codex 模型清单与前端白名单
- 价表 JSON 新增 deepseek-flash 条目并同步 flash 系价格
2026-09-10 22:10:29 +08:00
chihao.ou b112805ffa test(ops): align query matcher formatting with gofmt 2026-09-10 21:21:08 +08:00
chihao.ou d3bd614ba7 test(groups): initialize a valid rate multiplier in peak validation case 2026-09-10 21:17:42 +08:00
chihao.ou e9f5add97c fix(ops): show first-token latency in TTFT request details 2026-09-10 21:13:17 +08:00
chihao.ou ab3b398b5f fix(groups): return bad request for invalid peak-rate settings 2026-09-10 21:05:47 +08:00
chihao.ou 5032d06697 fix(codex): advertise DeepSeek vision image input 2026-09-10 21:04:56 +08:00
xlplbo bed1e3c362 fix(openai): 重试强制新建连接并放宽轮次预检超时 2026-09-10 20:35:53 +08:00
xlplbo a5b07b2960 feat(openai): WS 连接池常驻读循环应答上游 ping 2026-09-10 20:35:53 +08:00
xlplbo df8eceba0c fix(apicompat): chat 桥保留 Codex agent_message 任务正文 2026-09-10 20:25:38 +08:00
xlplbo 95a04c8dff test(apicompat): additional_tools 走 chat 桥的回归用例 2026-09-10 19:50:26 +08:00
xlplbo cd1ee1d1ad fix(openai): WS 接入会话状态改用执行作用域键 2026-09-10 17:50:24 +08:00
xlplbo 9e7c2d713c fix(openai): WS 会话抢占改用执行作用域键 2026-09-10 17:50:24 +08:00
xlplbo 06b61ba96f feat(openai): 解析 codex 线程标识派生 WS 执行作用域 2026-09-10 17:50:23 +08:00
r266-tech 1c0932e173 feat(gateway): retrieve models from the visible catalogue 2026-09-10 15:57:17 +08:00
nekocode a4edda36d5 fix(gateway): 不再剥离客户端 system 上的 cache_control
OAuth mimic 路径无条件删除 system[*].cache_control,不打日志也不报错。
把稳定前缀锚在 system 上的客户端因此拿到 cache_creation 与 cache_read
双 0 —— 不是没命中,是连写都没写。

这个动作是 46e5ac9 引进时「system 必然被整个重写」的配套:客户端的
system 内容都搬进 messages 了,残留断点指着空气,删它是对的。后来
system 注入变成可配置,于是长出了原设计里不存在的组合 —— system 没被
重写,断点照删。剥离失去了赖以成立的前提,代码却还在。

三处调用点里 count_tokens 那处还是硬编码 true,与另两处的
!systemRewritten 语义不一致。

直接移除该选项而不是再加一个开关:注入开启时 system 已被整个替换、
留在上面的断点是我们自己拼进去的稳定锚点;注入关闭时断点就是客户端的
缓存意图。两种情形都没有删它的理由,而需要开关的是有取舍的行为。

顺带补上 count_tokens 的 4 块上限兜底:五条出口里只有它没在自己的转发
路径上调过 enforceCacheControlLimit,原先那句 strip 是这条路上唯一压制
断点数量的东西。拿掉之后「客户端 system + 客户端 messages + 刚注入的
tools[-1]」实测顶到 5 块,而上游对超限是 400。

messages 层的同类改写见 #2369,那次只落地了 messages 一半。

此前该行为零测试覆盖(stripSystemCacheControl 在 _test.go 里零命中),
本次补 4 个用例,其中上限兜底那条撤掉修复即红。

Claude-Session: https://claude.ai/code/session_01WnTW52pmELGy3g7KPEUXj2
2026-09-10 15:48:17 +08:00
smyyyyy 4e5632c3e3 fix: validate Codex User-Agent before identity parsing 2026-09-10 03:18:31 +08:00
chihao.ou 21323c31f8 fix(monitor): support upstream base URL paths 2026-09-09 20:02:25 +08:00
wangjunwei de76baaa0d fix(scheduler): preserve Anthropic threshold metadata 2026-09-09 17:13:37 +08:00
刘世俊 a0babc93dd fix(openai): accept full heartbeat bootstrap envelopes 2026-09-09 12:48:57 +08:00
hansnow 7ccc8a6f50 fix(openai): 支持 Image 2.5 并修复 OAuth 生图主控模型 2026-09-09 11:04:59 +08:00
lcy 29a36a10be Merge upstream/main into feat/antigravity-gemini-3.8-flash
Resolve comment-wording conflict in normalizeGeminiThinkingTierAlias by
keeping upstream's phrasing. Function body and behavior are unchanged.
2026-09-09 01:49:32 +08:00
reallyfeishi f180fbb884 fix: anchor channel monitor buckets to UTC 2026-09-08 23:02:06 +08:00