The option now fails loud without replayFixture instead of silently
mounting nothing, and its JSDoc states the interplay with the
consumption check. The fixture is a non-empty header row (no longer a
0-byte placeholder), and the model-label assertion polls for
DeepSeek-V4-Flash (the directory loads asynchronously) instead of
reading the attribute once.
Master's credential-lifecycle work taught the editor card that a pi-ai
profile names `apiKeyEnv` only when a key is actually stored, so a route
left blank keeps its provider-native auth path. The create card kept
writing the derived reference unconditionally, so a route declared for a
credential chain or ADC was born pointing at a reference nothing sets —
and now rendered a red missing-key dot for it.
Both cards apply one rule. The obsolete assertion moves with the behavior
(the with-key case is covered by the neighbouring test), and the merged
Models e2e golden shows the declared route unmarked rather than flagged.
A bare /plan command never calls a model, so the scaffold's replay row
did not mount and the model directory was empty: the trigger rendered
the short fallback label, which fits beside the chip even on the
pre-fix layout, silently defanging the regression. The scaffold gains a
replayProvidersOnly option (provider catalog without a recorded script,
consumption check skipped), the test mounts it, and asserts the trigger
aria-label contains DeepSeek-V4-Flash before measuring — verified that
removing the wrap fix makes the test fail (click areas disjoint: false).
A row's stored profile could not tell a hand-declared gateway from a
shipped provider whose models someone narrowed — both look identical from
outside the adapter — so the Models page had no way to mark the routes a
deployment added itself.
The directory entry now carries `declared`, answered by the owning adapter
against its own installed catalog, and the page renders a Custom tag from
it. Absence stays "this adapter draws no such distinction" rather than
"shipped", so a route no adapter claims is labelled neither way.
Also records the default-route work's Agent Note and the e2e evidence for
all three changes: the composer switch writing the section, and the Models
page declaring a route with its own reasoning effort.
ChatView's bottom-follow recognized only wheel gestures as reader input,
so touch panning, scrollbar dragging, and keyboard paging could not leave
the bottom of a streaming transcript. Replace the wheel listener with
device-agnostic attribution: a scroll position deviating from the
observed-top ledger of the last delivered or written scrollTop is reader
input. Adds keyboard-paging and touch-style fling e2e scenarios (red under
the old implementation) and the bilingual Agent Note triplet.
The /plan command handler commits plan/mode active immediately on the
live agent (the lifecycle-chrome precedent), so the test drops the
recorded fixture, the record/replay mode split, and the turn-settled
wait. The golden comparison stays in replay/refresh modes; the fixture
file is removed and the note describes the no-model path.
- Drop the review-history narrative from checkDoneValue's JSDoc (the "in the
previous implementation … now avoids" clause); state the current contract only.
- Assert encodeJsonPlain on the same 100k-deep value the metering test uses:
its headline contract is stack-safety (JSON.stringify would throw), but no
test exercised the encoder on a deep value.
Address the latest review round:
- WireFrameShapesCoverUnions checked only union ⊆ roster, so removing a frame
from a message union (e.g. dropping ReplyErr from ReplyMessage) left the check
true while the public TS union diverged from the wire. Replace it with a
bidirectional equivalence between MessageFrames and the roster's message-frame
value types (nested Namespace/ErrorClass/DoneErrorField excluded): both a frame
added to a union without a roster entry and a frame removed from a union now
fail typecheck (both verified).
- The mirror e2e's python3 probes imported protocol.py without -B, writing
py/__pycache__/*.pyc into the (un-ignored) source tree. Add -B to both.
- Refresh the metering prose (checkDoneValue JSDoc + README both sides + Agent
Note both sides): the incremental-work list no longer says "per-key
JSON.stringify" now that jsonStringBytesUpTo scans without stringifying;
re-record the README and Agent Note i18n pairings.
Two review findings on checkDoneValue's metering and the wire-mirror binding:
- The string/key byte check used a decoded-length lower bound and then called
JSON.stringify, which materializes the ~6x escaped copy before the over-budget
check — the hundreds-of-MB spike the metered walk exists to avoid. Add
jsonStringBytesUpTo, a non-allocating scan that computes the exact escaped
UTF-8 size (matching JSON.stringify byte for byte, including surrogate pairs
vs lone surrogates) and bails the instant it crosses the remaining budget; use
it for both string values and object keys.
- The frame roster in WIRE_FRAME_FIELD_ROLES was hand-written, so a frame added
to ChildToHost/ReplyMessage without a roster entry slipped past. Introduce
WireFrameShapes (name -> interface) as the canonical roster the roles map is
bound against, plus a WireFrameShapesCoverUnions compile-time assertion that
every message-union member appears in it (verified: adding a frame to a union
without a WireFrameShapes entry fails typecheck).
Address the remaining review findings on the wire-mirror layer:
- Export PROTOCOL_FD from protocol.ts as the TS-side source of truth the host
wires, and assert the Python constant against it in the mirror e2e instead of
a bare literal 3, so an fd drift on either side is caught.
- Correct the FrameFieldRoles JSDoc to point at the actual assertion site
(WIRE_FRAME_FIELD_ROLES's satisfies clause, not WIRE_FRAME_FIELDS).
- Drop the redundant explicit type annotation on WIRE_FRAME_FIELDS (the trailing
`as` cast already types it; Object.fromEntries returns an index signature).
- Refresh the mirror-test comment to describe the roles-map binding (a TS-side
add/remove/rename/optionality-flip fails typecheck; a Python-side change fails
the comparison).
The array-based FrameFields<T> only checked that listed names were members of
the frame's keys, so a field added to a TS interface (e.g. LogMessage.seq?)
left the existing arrays a valid subset — typecheck passed, and since the
constant and Python both lacked the field the mirror test passed too. The
JSDoc's claim that runtime covered this was false.
Replace it with WIRE_FRAME_FIELD_ROLES, a per-frame map keyed by field name
(`Record<RequiredKeys<T>, 'required'> & Record<OptionalKeys<T>, 'optional'>`),
so every interface key MUST appear with a matching required/optional tag: an
added field, a removed field, a rename, or an optionality flip all fail
typecheck at the roles map (verified). WIRE_FRAME_FIELDS is projected from it
as the sorted arrays the mirror test still compares to the Python TypedDicts.
Also drop the `export` added to the promoted frame interfaces (Namespace,
ErrorClass, RunMessage, DoneErrorField, ReplyOk, ReplyErr) — nothing outside
protocol.ts imports them, so the barrel surface is unchanged and knip stays
clean.
The previous mirror binding (FrameFields<keyof T>) only checked membership: it
could not see a TS-side optionality flip (truncated? -> truncated leaves keyof
unchanged) or a field added on one side, so the "depends on the TS
declaration" claim was overstated.
- Promote the inline frame shapes (Namespace, ErrorClass, DoneErrorField,
RunMessage, and the two Reply variants) to named interfaces so every frame
binds uniformly.
- Derive FrameFields from RequiredKeys<T>/OptionalKeys<T>, so `required` and
`optional` each accept only that side's keys. An optionality flip or a rename
now fails typecheck (verified: flipping LogMessage.truncated to required
errors at the constant).
- Enumerate EVERY public TypedDict in py/protocol.py in the mirror e2e (not a
name list taken from the TS side) and assert both the frame roster and each
frame's required/optional sets by exact equality, so a frame or field present
on only one side of the wire fails the test.
Two gaps from the previous round's fixes:
- checkDoneValue flagged a non-lossless number but skipped counting its encoded
bytes, so a value over budget ONLY through that number classified as
non-lossless instead of over-budget (e.g. [Infinity] at cap 3, whose encoding
is 10 bytes). Count the scalar's bytes even when flagging, so the budget check
wins as the JSDoc promises. Add cap-3 regression cases.
- The mirror e2e compared the Python TypedDict keys against a hand-written
constant, so a field change on the TS side alone would not fail it, and the
reply frames were not probed at all. Introduce WIRE_FRAME_FIELDS in
protocol.ts, bound to each frame interface's key set via `satisfies` (a
renamed/removed field breaks typecheck — verified), and drive the mirror test
from it, now covering ReplyOk/ReplyErr too. The test therefore fails on
one-sided drift from either language.
Address the two standing review suggestions in this layer rather than deferring
them to PR #4:
- Extend tests/protocol-mirror.e2e.ts to read each py/protocol.py TypedDict's
required/optional key set and assert it against the wire field names
src/protocol.ts declares (global included, via functional TypedDict). The
round-12 class of drift — a renamed/dropped field, or one side making a field
optional the other requires — now fails a test instead of relying on review.
Field types remain review-guarded (no mechanical TS/Python equivalent).
- Drop the forward references to PR #4's internal mechanisms from this layer's
prose: the "256 MiB frame ceiling" figure and the "(index.ts)" fd-3 pinning
citation become an abstract "host-side inbound frame-size cap" so the JSDoc,
spec, README, and Agent Note describe only what this layer owns.
Update both README sides and the Agent Note (both languages) to state the
mirror is now executable, and re-record their i18n pairings.
checkDoneValue returned non-lossless the instant it hit a non-finite/negative-
zero number, before finishing the budget metering. A value that is BOTH over-
budget and non-lossless then classified by member order: `["<huge>", 1e400]`
gave non-lossless while `[1e400, "<huge>"]` gave over-budget — the same value,
two verdicts — which would drive the consumer to emit invalid-output vs
output-limit non-deterministically, contradicting the JSDoc promise that an
over-budget value is rejected as over-budget regardless. Record the number
violation in a flag and let metering finish; return non-lossless only once the
whole value is confirmed within budget. Add a regression test asserting both
member orders classify as over-budget.
Both README sides still described checkDoneValue as "one bounded traversal /
一次有界遍历" — the same overclaim already retracted in the code JSDoc and the
Agent Note. Reword both to match: the walk bounds only the incremental
allocation it adds (escaped-string copy, enqueued children, per-key stringify);
the frame's own width is parsed upstream and capped by the host's fd-3 receive
buffer, not re-bounded here. Re-record README.i18n.yaml.
- Replace four raw U+0000 bytes in protocol.spec.ts string literals with the
\0 escape so the source stays plain text (a bare NUL makes text tools treat
the file as binary); the runtime value is unchanged, so the bytes:8 NUL-escape
assertion still holds.
- Sync the Agent Note (both languages) with the corrected checkDoneValue
contract: the walk bounds only the incremental allocation it would add, not
the frame width, which is already parsed and capped upstream by the host's
fd-3 receive buffer. Drop the "prevents a hundreds-of-MB allocation" overclaim
that the code JSDoc already retracted. Re-record the note i18n pairing.
ownValues' JSDoc claimed the generator avoids a "second full-breadth
allocation before a single value is examined", but for...in still materializes
the key-name enumeration when the loop starts — the same JS limitation the
checkDoneValue rewrite now acknowledges. What the generator genuinely saves is
the extra VALUE array Object.values/Object.entries would copy; state that
precisely rather than implying sublinear startup.
checkDoneValue cannot bound object width sublinearly: JS has no lazy own-key
iterator (for...in materializes the key set), and done.value is already
JSON.parse'd before the check runs, so the frame's width is paid upstream. The
genuine width bound is the host's fixed 256 MiB fd-3 receive buffer (a later
stack layer). Reword the JSDoc and branch comments to claim only what holds —
the traversal caps the INCREMENTAL allocation the check would add (escaped
strings, enqueued children, per-key stringify) and refuses over-budget before
those secondary allocations — and drop the mid-count micro-check that JS cannot
honor. Replace the Proxy test (whose ownKeys allocated a 2M array, proving
nothing) with assertions that an over-budget string/array/object is refused
before its escaped copy or child enqueue.
The base seam branch resolved its DUNDER_MEMBER inconsistency by keeping
/^__.+__$/ and asserting `____` (empty middle between two `__` pairs) does not
match. Drop this branch's earlier /^__.*__$/ stopgap so the seam file is
byte-identical to its base: the earlier change only existed because the base
was self-inconsistent, and the base now owns a coherent decision.
The object branch counted every own key before applying the size bound, so a
forged done.value with millions of keys and a small cap forced an O(frame)
walk — contradicting the O(cap) guarantee the comment promised and able to
block the host event loop. Bail mid-count the instant the running minimum
encoding (braces + 4 bytes/entry + commas) crosses maxBytes, and drop the now
-redundant post-count check the loop subsumes. Add a Proxy-based test proving
a 2M-key object enumerates fewer than 1000 keys under a 64-byte cap.
Also correct the checkDoneValue JSDoc: per-scalar byte length is measured via
scalarJson (exact BigInt digits for beyond-safe integers), not JSON.stringify.
- Translate README.zh.md's Model Experience body and KV Cache line, which
were left verbatim in English.
- Convert half-width punctuation to full-width across the README.zh.md
Known Limitations bullets and the entire Agent Note Chinese side, per
docs/i18n translation-rules.md Typography (MUST use ,。:()in Chinese prose).
- Re-record both README and Agent Note i18n.yaml pairing hashes.
- Reword the workspace-constraints extra-files comment: this layer's py/
ships only the wire-protocol mirror; the spawned bootstrap arrives later.
The base seam branch still carries the buggy /^__.+__$/ (rejects `____`,
which its own reserved.spec asserts must match), so this stacked branch must
keep the /^__.*__$/ correction to stay green. Reword the JSDoc to the same
line count as the base so the CodeRuntime class does not shift, leaving the
cordis services catalog anchor identical to the base and confining this
branch's footprint on the seam file to the single regex character.
- Drop a CALL frame whose id is negative zero: it passes Number.isFinite but
the reply re-serializes it as `0`, colliding with a real call id `0`. The
honest child never issues `-0`.
- Document that validateChildFrame preserves a forged done frame's value and
error together on purpose, so consumers must check error before value.
- Cover the log-frame `truncated` rebuild branch: assert a literal-true flag
rides along and any other value (1, string, false) is dropped, closing the
protocol.ts branch the coverage gate flagged.
- Correct encodeJsonPlain's JSDoc: it matches compact JSON.stringify EXCEPT on
a beyond-safe-range integral double, where it emits the exact BigInt digits
(`...846976`) rather than the rounded `...847000` — the divergence the
"emits exact digits" test pins.
- Declare py/protocol.py's `global`-bearing frames (Namespace, CallMessage)
with functional TypedDict syntax so they carry the real wire key instead of
a `global_` attribute the wire never sends, and split optional-field messages
(Namespace/LogMessage/DoneMessage) into a required base plus a total=False
subclass so `type` and other required fields cannot be dropped. Widen
HostToChild to include the boot and run frames the host sends before replies.
- Reword the mirror e2e's py/ directory assertion to describe the source-tree
layout it actually checks.
- Drop the unused @deepseek-ai/dsh-code-runtime dependency: this layer
imports nothing from the seam (protocol.ts has no imports; the invariant
companion uses only cordis and dsh-invariants). The backend-core PR
re-adds it when PythonCodeRuntime consumes the seam. Fixes knip.
- Point the Agent Note's cross-reference to the seam note at the English
target on both language sides, per the bilingual-pairing contract (only
the language switcher flips to .zh.md). Re-record the sidecar.
- Add the Known Limitations section both READMEs require, covering the
cross-language guard's scope and the deferred runtime implementation.
- Regenerate the module graph for the dropped dependency edge.
The seam's dunder-member test (added in the base seam PR) asserts
`DUNDER_MEMBER.test('____')` is true and `test('__')` is false, but the
regex `/^__.+__$/` rejected `____`: the `.+` demanded a non-empty middle,
while `____` is two adjacent `__` pairs with an empty middle. Widen to
`/^__.*__$/` so a name with distinct leading and trailing `__` pairs
matches whether or not it has a middle, and align the JSDoc. Regenerate
the cordis service catalog for the merged seam source line.