Commit Graph
15492 Commits
Author SHA1 Message Date
Tianyi Cui dbd28a8071 fix(benchmarks): keep composer focus during streamed input 2026-09-07 17:21:56 +08:00
Tianyi Cui 201987b374 perf(llm): scan file content without per-array callbacks 2026-09-07 17:18:18 +08:00
Tianyi Cui f277ec0e06 Merge published V3 root after master integration approval
Include the root merge ancestry from #3669. Preserve the already validated system-message child recording rather than restoring the root’s header-system representation. The resulting tree is unchanged from the preceding representation head.
2026-09-07 16:58:52 +08:00
Tianyi Cui 930e4d67ea Merge pull request #3669 from deepseek-harness/fix/spkv-root-latest-master
ci(session): bring protected V3 integration onto latest master
2026-09-07 16:45:19 +08:00
Tianyi Cui 476918da0a Merge latest master teardown and migration waiter fixes
Observe actual shared-preparation waiter admission and settle both callers before cleanup, preserving current V3 migration expectations. Adopt master’s cross-platform teardown regressions without changing budgets or Session data.
2026-09-07 16:15:57 +08:00
Tianyi Cui a062772d7d Merge latest master teardown and migration waiter fixes
Observe actual shared-preparation waiter admission and settle both callers before cleanup, preserving current V3 migration expectations. Adopt master’s cross-platform teardown regressions without changing budgets or Session data.
2026-09-07 16:15:57 +08:00
Tianyi Cui 4ac4e3d014 Merge pull request #3668 from deepseek-harness/fix/windows-subagent-teardown-ci
test(ci): synchronize Windows waiters and complete teardown
2026-09-07 15:52:01 +08:00
Tianyi Cui d9042f61e1 test(snapshot): refresh V3 human-steering child recording
Record current V3 output from master’s human inbox control scenario while retaining the V2 predecessor unchanged. No additional product behavior or validation change.
2026-09-07 15:48:25 +08:00
Tianyi Cui b7f66f4440 test(sdk): record native V3 human-steering inbox transitions
Refresh the current V3 child recording through the actual SDK scenario after master’s inbox-control change. Preserve the V2 predecessor inherited from master and retain system-message admission and current references.
2026-09-07 15:47:27 +08:00
Tianyi Cui 50cce9c28d Merge latest master human inbox controls into system-prompt layer
Preserve the V3 PTC SDK assertion and enable the new human-steering runtime fixture while carrying master’s inbox lifecycle changes. Retain prior deterministic CI repairs and immutable predecessor ownership.
2026-09-07 15:45:09 +08:00
Tianyi Cui fffd39b101 Merge latest master human inbox controls into V3 integration
Preserve the V3 PTC scenario while enabling the current subagent human-steering fixture environment. All product changes are inherited from master; no V3 migration semantics are altered.
2026-09-07 15:44:17 +08:00
Tianyi Cui e2b3800a7e fix(python): challenge fresh file state in live SDK smoke 2026-09-07 15:40:51 +08:00
Dudu-0223 96ead6091d feat(subagent): align human inbox controls (#3223)
* feat(subagent): align continuable human inbox controls

* fix: repair rebase documentation and close test persistence

* docs: refresh rebased module dependency graph
2026-09-07 07:39:39 +00:00
Tianyi Cui f7b1658ff2 fix(benchmarks): reserve system heads in generated V3 histories 2026-09-07 15:03:32 +08:00
Tianyi Cui 19d167cec2 Merge latest master request-freeze and frontend performance work
Preserve the V3 system-prompt projection while adopting master’s per-agent frozen-message provenance and frontend performance gates. Keep benchmark workload admission and diagnostics fixes intact, and retain both migration and first-open/reopen lifecycle guarantees.
2026-09-07 14:53:00 +08:00
Tianyi Cui 95400aa52c Merge V3 preset-reference migration into system-prompt representation
Compose the root’s exact legacy code-to-ptc preset conversion with the existing structural system-head migration. Keep dense source events, audited payload admission, local reference mapping, and native V3 extension behavior unchanged. Preserve the historical preset migration fixture byte-for-byte and test unsupported sparse/opaque sources explicitly.
2026-09-07 14:51:05 +08:00
Tianyi Cui 85dcebcda4 Merge remote-tracking branch 'origin/master' into fix/spkv-root-latest-master 2026-09-07 14:48:39 +08:00
Tianyi Cui 8d0f45febc Merge remote-tracking branch 'origin/release/session-log-v3' into fix/spkv-root-latest-master 2026-09-07 14:47:55 +08:00
Tianyi Cui 5608528e14 Merge pull request #3637 from deepseek-harness/perf/frontend-scenarios
test(perf): gate long-session browser and reconnect workflows
2026-09-07 14:42:15 +08:00
Tianyi Cui b53c95e458 Merge pull request #3664 from deepseek-harness/fix/scroll-follow-pending-sample
fix(chat): preserve pinned scroll-follow across layout changes
2026-09-07 14:38:43 +08:00
Tianyi Cui 4023879df7 test(codex): finish captured cleanup after sibling failures 2026-09-07 14:30:31 +08:00
Tianyi Cui a1d11f213f test(session): observe shared waiter admission before cancellation 2026-09-07 14:30:29 +08:00
Tianyi Cui 5b2c9ba2d2 Merge pull request #3666 from deepseek-harness/worktree/session-v3-preset-rename
fix(session): migrate legacy code preset references to ptc
2026-09-07 14:30:24 +08:00
Tianyi Cui d1e9254454 fix(web-tests): author preset and stats seeds in current format
Generated V0 logs put user surfaces before the first step and cannot migrate to a protected V3 system head without changing chronology. Author native current headers, open the first step before the empty system head and user, keep title sequence references aligned, and supply embedded assistant streams.

Give the preset child a current identified user message. Preserve all paging, geometry, preset assertions and committed historical fixtures. Validated both generators red/green through the source seed reader, one full build, all preset/stats/trajectory browser tests (11), and scoped type-aware lint.
2026-09-07 14:21:50 +08:00
Tianyi Cui dda0f05999 fix(web-tests): generate native system head before chat users
The long-chat builder emitted current V3 with user/message before system/message, so seed restoration rejected its protected head before browser assertions ran. Open the first step and append the system head before the user surface; keep Session-assigned title and tool-result references.

Add an independent seed-reader regression for current version, protected head order, title/tool references and closed turns. Observed protected-head failure before the fix; focused regression and full configured type-aware lint pass. Existing browser scroll assertions are unchanged; full built owner runs follow.
2026-09-07 14:17:08 +08:00
Tianyi Cui 12dd429331 Merge remote-tracking branch 'origin/master' into fix/spkv-root-latest-master 2026-09-07 14:02:51 +08:00
Tianyi Cui 2eccae5631 Merge branch 'release/session-log-v3' into worktree/session-v3-preset-rename 2026-09-07 14:00:22 +08:00
Tianyi Cui a4f6250be9 Merge remote-tracking branch 'origin/release/session-log-v3' into feat/system-prompt-surface-node 2026-09-07 13:56:08 +08:00
Tianyi Cui 086b94d4f6 Merge pull request #3655 from deepseek-harness/fix/v3-pwsh-current-fixtures
test(session): 同步最新 master 并修复 V3 PowerShell fixtures
2026-09-07 13:53:21 +08:00
Yichen Jiang ff33f79eb1 fix(session): migrate legacy code preset references to ptc 2026-09-07 13:52:22 +08:00
Tianyi Cui 9ef426d729 fix(chat): keep genuine near-floor scroll gestures pending 2026-09-07 13:46:30 +08:00
Tianyi Cui 421e3b075c test(subagent): preserve lane budgets and await teardown completion 2026-09-07 13:39:56 +08:00
Tianyi Cui d3ed6d6f88 Merge remote-tracking branch 'origin/master' into feat/system-prompt-surface-node 2026-09-07 13:35:06 +08:00
Tianyi Cui 57885bd73d Merge latest master prompt placement through V3 integration
Preserve system/message persistence while adopting named persona prefix/suffix and environment facts after reusable instructions. Resolve the Web assertion against the logged system message, not retired header.system. Keep the existing protected V3 integration PR as the root publication path; no rules are bypassed.
2026-09-07 13:34:43 +08:00
Tianyi Cui cbdf32447e fix(benchmarks): calibrate repeated hosted browser open median 2026-09-07 13:29:38 +08:00
Tianyi Cui 1a3af89a8d fix(benchmarks): calibrate hosted paging and trajectory endpoints 2026-09-07 13:29:38 +08:00
Tianyi Cui 620c2b0b27 fix(benchmarks): calibrate hosted frontend endpoints and preserve input overlap 2026-09-07 13:29:38 +08:00
Tianyi Cui 0ac1d4d865 test(perf): align browser provisioning with hosted benchmark runner 2026-09-07 13:29:38 +08:00
Tianyi Cui c51c16cdac test(perf): respect browser failover provisioning policy 2026-09-07 13:29:38 +08:00
Tianyi Cui 00e452aa03 docs(perf): confirm repeated frontend CI calibration 2026-09-07 13:29:38 +08:00
Tianyi Cui 82cd35467a docs(perf): record first frontend CI calibration 2026-09-07 13:29:38 +08:00
Tianyi Cui 9369585872 test(perf): bound browser observers to active response 2026-09-07 13:29:38 +08:00
Tianyi Cui 2927034ffe test(perf): retain successful CI measurement output 2026-09-07 13:29:38 +08:00
Tianyi Cui 7ac5e08246 test(perf): measure real history streams and live input overlap 2026-09-07 13:29:38 +08:00
Tianyi Cui 6ad74db2e3 test(perf): gate long-session browser and active reconnect workflows 2026-09-07 13:29:38 +08:00
Tianyi Cui 843c8723e5 docs(perf): link active hosted request calibration 2026-09-07 13:29:37 +08:00
Tianyi Cui e2b81cf8bf docs(perf): distinguish hosted request and catalog budgets 2026-09-07 13:29:37 +08:00
Tianyi Cui 8e270960ed test(perf): calibrate request history on standard hosted CI 2026-09-07 13:29:37 +08:00
Tianyi Cui 73edce1ae7 perf(agent-loop): reuse proven message freezes per agent 2026-09-07 13:29:37 +08:00
Tianyi Cui 84914c316d test(perf): calibrate baseline requests for hosted CI 2026-09-07 13:29:37 +08:00