PR #2239 removed the old in-memory activation entry and renumbered the local-modification log so Cordis source publication is item 16 and the rescope is item 17. It updated vendor/README.md but left this exact edit expecting the rescope before an item 18, so the current master post-state matched neither side and pnpm run hygiene failed.
Treat item 16 as the pre-rescope anchor and append item 17 in the replacement. The forward edit now produces the checked-in ordering, while reversing it removes only the rescope entry and preserves the independent Cordis publication note.
Verified with the rescope-vendor unit suite, pnpm run rescope-vendor:check, pnpm run hygiene, and git diff --cached --check.
Master advanced from e03b51d7db to ee223e6545 after bot-thread resolution and the PR-body policy fix. Merge-forward preserves the reviewed issue commits and explicit checkpoint history instead of rewriting the branch.
The only textual conflict was the generated translation-pairing record for docs/config-catalog. Regenerate the English catalog from the combined source tree and re-record the bilingual hashes so current-master command-line changes and this PR's compression configuration coexist.
Verified with pnpm run verify-config-catalog, pnpm run verify-translation-pairing docs/config-catalog.md, and git diff --cached --check.
The abstract capability flag forces every backend to state whether it owns per-session raw artifacts, but TypeScript cannot express that a true flag requires replacing the concrete unsupported default. Without an implementer-facing obligation, a backend could advertise support and then fail with a contradictory unsupported diagnostic on first use.\n\nDocument the required pairing at the capability declaration. Keep readRaw concrete so backends that correctly report false inherit one fail-loud implementation instead of duplicating rejection code.
ApiProxyDefaults uses an exact optional property, so passing config.sessionExportCompressionLevel directly made the service object carry an explicit undefined that is not assignable to the resolved request shape. The full host build caught this distinction after the redundant fallback was removed.\n\nConditionally omit the property when Cordis has not supplied a value. Direct createApiProxy callers still receive the implementation-owned default, while configured plugin values pass through without introducing another defaulting site.
Master advanced again after the first merge-forward checkpoint, adding the web human-transcript command-input change. Preserve the earlier checkpoint and merge the new exact e03b51d7 base as a separate commit instead of rebasing or rewriting this merge-heavy stack.\n\nThe incremental merge is conflict-free and keeps the session-export review fixes based on the repository's current integration state.
Response-consumer cancellation already stopped lineage reads, persistence reads, and ZIP production, but the final attachment phase called readImage without the producer signal. A slow or stalled attachment backend could therefore keep working after the browser abandoned the download and prevent the producer from settling.\n\nExtend the attachment read seam with optional cancellation, forward it through the local backend into Node's filesystem read, and preserve the abort reason rather than wrapping it as a storage failure. The exporter now passes its combined request/consumer signal to every attachment read.\n\nCover both ownership boundaries: the local-store test proves filesystem forwarding and cancellation identity, while the assembled export test cancels a reader during a pending attachment provider call. Regenerate the Cordis API catalog and paired documentation so implementers can rely on the new contract.
The response stream uses a fixed 64 KiB byte high-water mark; no deployment setting controls it. Calling that queue configured incorrectly suggested another tuning surface and obscured the concrete memory bound.\n\nName the fixed capacity directly while preserving the separate bound of one synchronous fflate push beyond the queued bytes.
The pre-stream error boundary covers both the live-session flush barrier and the persistence read, but its response attributed every failure to reading storage. A flush failure therefore produced a misleading diagnostic even though the response correctly withheld private backend details.\n\nUse preparation as the shared operation name and cover the flush-failure path explicitly. Both preparation stages now retain one stable, path-safe HTTP 500 without pretending to identify the failing stage.
The Cordis schema supplies the normal plugin default, while createApiProxy also owns the fallback required by direct programmatic callers. Repeating the same nullish fallback in ApiProxyService created a third defaulting site without adding a distinct invariant.\n\nPass the validated config value through unchanged and leave createApiProxy as the single implementation boundary that turns an optional request value into the required compression specification.
The fixture comment still said the Trajectory action used window.fetch after the implementation moved to a temporary download anchor. That wording implied client-side response handling and buffering which the browser-download design deliberately avoids.\n\nDescribe the actual native download-manager handoff while retaining the important contract: the fixture download stub only satisfies the host type and is unreachable through fixture dispatch.
Refresh PR #2258 onto master at 5427cbcc19 so the session-log export fixes are evaluated and mergeable against the current repository.\n\nPreserve master's SDK-toolchain removal, client theme bootstrap, and Python finish-reason changes intact. The only textual overlap is the generated config-catalog pairing record: both English and Chinese catalogs merge cleanly with the PR's compression setting and master's package moves, so regenerate that sidecar from the merged owners instead of choosing either stale hash.
Session-log ZIP entries always used DEFLATE level 6 even though compression level is a deployment tradeoff: CPU-constrained hosts may prefer low latency while bandwidth-constrained hosts may prefer smaller archives. A hardcoded level also violated the repository rule that deployment-varying plugin choices live in validated Config.
Add sessionExportCompressionLevel to ApiProxyService.Config as an integer 0-9 with default 6, resolve the same default once for direct createApiProxy callers, and pass the required level into the streaming module. Tests prove schema defaulting and rejection as well as a level-0 versus level-9 archive-size difference with identical extracted content. The generated config catalog, bilingual gateway README, and feature note document the knob and its tradeoff.
The ZIP loop checked desiredSize only after a push and responded to an overfull queue with setTimeout(0). A timer turn does not mean the consumer drained anything, so a slow or disconnected client still allowed the producer to enqueue the complete compressed archive while later artifact and attachment reads ran eagerly.
Give the ReadableStream a 64 KiB byte queuing strategy and block the single producer on a pull-released capacity gate whenever desiredSize is non-positive. Cancellation wakes that gate through the existing producer signal; synchronous fflate output is therefore bounded to the queue high-water mark plus one input push. A regression test exhausts timer turns without consuming and proves the next media entry remains unread until response pulling begins, and the bilingual contracts now describe the real bound.
Only the root raw-artifact read received the request signal. Lineage discovery and descendant reads could continue after disconnect, response-body cancellation did not stop the producer, and the root error boundary converted an abort rejection into an ordinary HTTP 500.
Combine request and response-consumer cancellation into the ZIP producer signal, forward it through every cancellable read, check it around the attachment seam, and terminate fflate exactly once when production stops. The pre-stream boundary now rethrows the original abort instead of translating it. Regression tests cover signal propagation, exact cancellation identity at the HTTP boundary, and a reader cancellation interrupting an in-flight descendant read; the bilingual host contract records these lifecycle semantics.
The exporter read persistence artifacts directly even when the requested root or a descendant was still live. Buffered session events could therefore be omitted from a successful download, so the advertised verbatim-artifact guarantee described storage accurately but captured an arbitrarily stale durability boundary.
Resolve each id against SessionStore and cross its authoritative flush barrier immediately before readRaw. Cold sessions remain a no-op, while live roots and descendants are made durable independently; this intentionally yields a per-session read-boundary snapshot rather than claiming an atomic lineage snapshot. A host-path regression test proves both artifacts change from stale to durable only through flush, and the bilingual host contract and Agent Note document the boundary.
A present zero-byte .jsonl.zstd file was treated as though no artifact existed because readRaw returned undefined when frame scanning found nothing. That contradicted both the plaintext path and the logical zstd reader, and it made the export endpoint answer 404 for on-disk corruption.
Treat a present artifact without a complete header frame as corruption and reuse the zstd reader's existing diagnostic. The regression test now distinguishes an existing empty file from an absent path, and the bilingual JSONL storage contract records that zero-frame artifacts reject alongside other header and frame failures.
SessionPersistence.readRaw previously used undefined for two unrelated states: a supported backend could not find the requested session, or the backend had no per-session artifact concept at all. The export endpoint consequently reported an existing SQLite-backed session as HTTP 404, which falsely diagnosed storage capability as session absence.
Make raw-artifact support an explicit backend capability. Unsupported backends now fail their inherited readRaw path loudly and the host answers 501 before reading, while undefined retains the single meaning of an absent artifact on a supporting backend. First-party backends, test providers, generated API catalogs, bilingual persistence docs, and export error contracts now state that distinction; focused tests cover both the 501 and the inherited rejection.
The export endpoint already streams a ZIP response, but the web client immediately converted that response into a Blob. That forced the complete archive through JavaScript memory before a download could start and coupled transport, buffering, object-URL lifetime, and filename handling to the trajectory view.
Navigate a temporary download anchor directly to the export endpoint instead. The browser now owns streaming and HTTP failure presentation, while a standalone delivery module owns URL construction and filename sanitization. Focused tests cover the handoff, rejection behavior, and the assembled session view; the package README and feature note record the new ownership boundary.
Master advanced from 1ac6ee70bb24a566cca74d99bdfe5b1bbc092940 to cb7b5af91e9f23bb300a1b7d4fb64b64abbdea01 while the prior retarget's fresh CI was finishing. Preserve both already validated merge checkpoints and incorporate the new live tip with another forward merge instead of rewriting history.
The newer base adds session-log format guards and benchmark documentation. It merges automatically, including the generated config-catalog overlap, so no MCP feature resolution is required; this checkpoint makes GitHub evaluate PR #2187 against the master tip current at the final push.
Master advanced from c757901957abdfd87f1cd7b11ecab16f3050e65d to 1ac6ee70bb24a566cca74d99bdfe5b1bbc092940 while retarget verification was running. Preserve the already validated c757901957 checkpoint and merge the new tip forward instead of rewriting that checkpoint.
The newer master changes subagent output selection and removes the empty experimental package group. It merges cleanly with PR #2187, so no feature-resolution changes are required; this commit makes final mergeability and CI evaluate the branch against the live master tip.
Retarget PR #2187 from master 8a763b34645fc9371c0c21595ac65438d603d9bf to c757901957abdfd87f1cd7b11ecab16f3050e65d so the reconnect changes are evaluated against the current release and package metadata.
Resolve the mcp-client package.json conflict by preserving master's publishable-package metadata and workspace peer ranges while retaining the PR's direct @deepseek-ai/dsh-timeout peer and development dependency. The reconnect implementation therefore keeps its explicit runtime contract without discarding current-master release configuration.
The reconnect supervisor used connection lost for every transition into backoff, including an initial startup attempt that never established a connection and later retry attempts that also failed. That wording implied a previously healthy generation and obscured whether any tools had ever been registered.
Capture whether the generation had reached the established state before scheduling recovery. Established disconnects retain connection lost/reconnecting; startup and retry failures now report connection failed/retrying. The reconnect-disabled diagnostic uses the same distinction while preserving its concrete manual-recovery guidance.
Unit assertions cover established loss, initial failure, retry failure, and both reconnect-disabled branches. Focused package coverage remains 100%, and the bilingual Agent Note records the observable state vocabulary.
Strict startup intentionally wraps connection and synchronization failures with the server-qualified activation diagnostic while retaining the original error in Error.cause. The prior assertion checked only the wrapper text, so the causal chain could regress unnoticed and erase the actionable transport failure.
Assert the full wrapper message and object identity of the original connection error in cause. This keeps operator-facing context and the underlying SDK diagnostic independently stable without changing production behavior.
The failure-cap path already appends tool disposal to syncChain, but the existing tests only covered give-up after settled discovery. They could not detect a future change that disposed the old set immediately and then allowed a blocked re-sync to publish a new leaked generation.
Hold a list_changed fetch open, drive the reconnect budget to exhaustion, then release a different tool list. The test proves final cleanup runs after that in-flight swap and removes both the previous and late-published tool names while creating no attempt beyond the configured cap.
The supervisor selected strict startup registration with a shared isFirstSync flag. Because the MCP SDK may deliver tools/list_changed before connect() resolves, that notification could enter the sync queue first, consume the strict option inside its contained handler, and leave the actual activation sync non-fatal.
Pass startup intent explicitly to connectGeneration(). Only the plugin activation attempt receives the failOnStartupError registration policy; notification-driven syncs and later reconnect generations always use contained runtime semantics. Queue arrival order can no longer redefine startup behavior.
A regression test injects list_changed from inside connect(), keeps a foreign namespace squatter in place, and proves activation still rejects after the notification's contained sync. Focused package coverage remains 100%, and the bilingual reconnect note records the ownership rule.
The MCP SDK starts a fire-and-forget close when initialization fails. Its stdio transport clears its process field before that close finishes, so our second Client.close() could return immediately and the reconnect timer could launch a replacement while the original child was still alive.
Track the transport onclose signal for every client generation and gate failed-attempt backoff on both Client.close() settlement and that signal. Use the same barrier during plugin disposal. If the SDK's bounded stdio termination window expires without onclose, fail closed and report incomplete shutdown instead of risking overlapping server processes.
Regression coverage models the SDK's early-returning second close, delayed and missing close signals, pending-connect disposal, close rejection, and the terminal timeout path. The reconnect Agent Note and Chinese counterpart now record the quiescence contract.
Integrate master at 8b4ef532 before addressing review feedback so each bot issue remains isolated in its own descendant commit. Resolve the vendor package rescope by moving the PR's new MCP supervisor imports to @deepseek-ai/cordis and @deepseek-ai/schemastery, preserve the reconnect implementation and e2e coverage, and regenerate the merged config and module catalogs with fresh bilingual pairing records.
Tianyi Cui