Repository navigation
fix: restore the missing personaNamespace import dropped by the #2731 merge (fix #3064; surfaced by PR #3055) - #3068
Closed
ranxianglei wants to merge 1 commit into
Closed
ranxianglei wants to merge 1 commit into
ranxianglei wants to merge 1 commit into
Conversation
Owner
Author
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #3064
What happened. Since the #2731 merge (
fa7dc4e1b),src/server/handle.ts:811callspersonaNamespace(...)on the new google-lane dsh-persona branch, but the symbol is never imported →tscfails with TS2304. tsup/esbuild bundles without type-checking, sodist/still ships — with a free-variable reference that throwsReferenceError(→ 502) whenever the google+dshPersona path executes. First visible casualties: PR #3055's ci-image Typecheck step and its dsh plugin lane.Root cause (verified against the merge tree). This is not a simple forgotten import in #2731's own branch — that branch was internally consistent (all four wire sites switched to
personaNamespace, import included). The damage landed in the MERGE resolution offa7dc4e1b: for the three pre-existing wire sites (anthropic/openai/responses) it kept the master-side version (personaConversationOrderfromsession-resolve.js), kept #2731's NEW google branch verbatim (callingpersonaNamespace), but the import line followed the master side — so the google call became unbound. Commit-by-commit evidence:623d3078c(pre-merge master) andfc3d91382(#3056 head) both show 4×personaConversationOrder+ the session-resolve import; PR headd66bd9e4dshows 5×personaNamespace+ the persona-anchor import; mergedfa7dc4e1b/dafff5418(current master) show the mixed state with NO persona-anchor import.Fix. One added line in
src/server/handle.ts:The minimal fix, deliberately NOT re-routing the google site through the
personaConversationOrderadapter: the merge resolution already chose which resolver each pre-existing wire keeps, and this PR only restores compilation — it changes zero wire behavior. Whether the google lane should eventually move to the order-based resolver too is an open parity question left to the owner (OP checklist item 1 offered exactly this either/or).Verification.
tests/side-request-isolation.test.tssubteste2e: google-lane dsh side request resolves verbatim — no kernel state, no persona fork (#2730)fails (the same 502 seen in fix: self-restart replacement binds-or-dies on the freed port instead of hopping (#3044) #3055's CI); with the line it passes.npm run typecheck: red (single TS2304) → green.tests/side-request-isolation.test.ts+tests/dsh-persona-fingerprint.test.ts— 34/34 pass.npm run buildgreen. Full-suite verdict: remote CI on this head.Behavior-change report (mandatory). Restores intended behavior only; no drift. Old state: master does not typecheck + latent runtime ReferenceError on the google+dshPersona path. New state: that path executes exactly as #2731 wrote it (fingerprint-ladder persona fork,
|sub:<fp>keys). No threshold/default/timing/output-format change anywhere. CI lanes: zero workflow/test-gate changes in this PR — all e2e/etc. lanes stay exactly as they were (open), per owner instruction; the push-event typecheck evaluation (issue checklist item 2) is deferred per owner ("后面再搞规矩").