Updates from Acepe session#213
Merged
Merged
Conversation
…ooling Show the thinking header while streaming, wire token reveal into thought groups, and restore TOKEN_REVEAL_STEP_MS so main-body text cascades smoothly again. Extend the dead-code analyzer with export-barrel-only detection and an opt-in apply script for high-confidence deletions. Co-authored-by: Cursor <cursoragent@cursor.com>
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
2 issues found across 39 files
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="scripts/dead-code/apply-dead-code.ts">
<violation number="1" location="scripts/dead-code/apply-dead-code.ts:22">
P2: `--root` parsing consumes the next token even when it is another flag. Missing root value can silently disable `--apply`/other flags and point analysis at an invalid path.</violation>
</file>
<file name="package.json">
<violation number="1" location="package.json:17">
P2: Misleading name: `dead-code:apply` does a dry-run, not an apply. Rename to signal dry-run behavior (e.g. `dead-code:dry-run`) and reserve `dead-code:apply` for the actual deletion.</violation>
</file>
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
| for (let index = 0; index < argv.length; index += 1) { | ||
| const arg = argv[index]; | ||
| if (arg === "--root") { | ||
| repoRoot = argv[index + 1] ?? repoRoot; |
There was a problem hiding this comment.
P2: --root parsing consumes the next token even when it is another flag. Missing root value can silently disable --apply/other flags and point analysis at an invalid path.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At scripts/dead-code/apply-dead-code.ts, line 22:
<comment>`--root` parsing consumes the next token even when it is another flag. Missing root value can silently disable `--apply`/other flags and point analysis at an invalid path.</comment>
<file context>
@@ -0,0 +1,109 @@
+ for (let index = 0; index < argv.length; index += 1) {
+ const arg = argv[index];
+ if (arg === "--root") {
+ repoRoot = argv[index + 1] ?? repoRoot;
+ index += 1;
+ continue;
</file context>
| "scripts": { | ||
| "audit": "bun audit --audit-level high", | ||
| "check:dead-code": "bun scripts/dead-code/find-dead-code.ts --strict", | ||
| "dead-code:apply": "bun scripts/dead-code/apply-dead-code.ts", |
There was a problem hiding this comment.
P2: Misleading name: dead-code:apply does a dry-run, not an apply. Rename to signal dry-run behavior (e.g. dead-code:dry-run) and reserve dead-code:apply for the actual deletion.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At package.json, line 17:
<comment>Misleading name: `dead-code:apply` does a dry-run, not an apply. Rename to signal dry-run behavior (e.g. `dead-code:dry-run`) and reserve `dead-code:apply` for the actual deletion.</comment>
<file context>
@@ -14,6 +14,8 @@
"scripts": {
"audit": "bun audit --audit-level high",
"check:dead-code": "bun scripts/dead-code/find-dead-code.ts --strict",
+ "dead-code:apply": "bun scripts/dead-code/apply-dead-code.ts",
+ "dead-code:apply:force": "bun scripts/dead-code/apply-dead-code.ts --apply",
"test": "bun scripts/forbid-structural-tests.ts packages && bun run --cwd packages/ui test && bun run --cwd packages/desktop test && bun run --cwd packages/website test"
</file context>
|
| Filename | Overview |
|---|---|
| scripts/dead-code/find-dead-code.ts | Adds export-barrel-only candidate detection via barrel re-export tracing. One dead branch in resolveReexportTarget (both arms return the same value) may over-populate barrelConsumed and hide valid candidates; separately, dead regex code in parseImportNames after an earlier continue guard. |
| scripts/dead-code/apply-dead-code.ts | New dry-run/apply CLI for dead-code deletion. Safe default (dry-run) and opt-in --apply guard; unlinkSync is properly scoped to repo-relative resolved paths. |
| packages/desktop/src-tauri/src/acp/commands/session_commands/new_session.rs | Adds require_creation_attempt_for_deferred_bind to fail-closed when a deferred bind's creation row is missing, replacing a silent Option unwrap with a proper error. Tests cover both the happy path and the missing-attempt case. |
| packages/desktop/src-tauri/src/acp/client/cc_sdk_client/mod.rs | Clears pending_model_id when the capability catalog is empty, preventing a stale model from persisting across reconnects. Test coverage added. |
| packages/desktop/src/lib/acp/store/session-sequence-id-backfill.ts | New utility that resolves the sequence ID backfill priority: metadata wins, then graph, then pending creation. Logic is simple and well-tested. |
| packages/desktop/src/lib/acp/store/session-open-snapshot-applier.svelte.ts | Adds syncSessionSequenceFromPendingCreation called before completePendingCreation, ensuring the sequence ID is backfilled from either the graph or pending creation before the pending state is cleared. |
| packages/ui/src/components/agent-panel/agent-assistant-message.svelte | Refactors thinking-block collapse to use a derived isCollapsed driven by streaming state, and shows the Thinking header unconditionally while streaming. Token-reveal timing is now applied to the last thought group during streaming. |
| packages/desktop/src/lib/acp/components/agent-panel/logic/agent-panel-header-sequence-id.ts | New logic function resolving panel header sequence ID from metadata, falling back to pending creation sequence when metadata isn't yet populated. |
| packages/desktop/src/lib/acp/components/messages/assistant-message.svelte | Deleted (dead code cleanup) — replaced by the shared component in packages/ui. |
Sequence Diagram
%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
participant Panel as agent-panel.svelte
participant Store as SessionStore
participant Facade as SessionConnectionFacade
participant Coord as SessionCreationCoordinator
participant Applier as SessionOpenSnapshotApplier
participant Backfill as session-sequence-id-backfill
Note over Panel,Coord: Before session metadata arrives (pending state)
Panel->>Store: getPendingCreationSession(id)
Store->>Facade: getPendingCreationSession(id)
Facade->>Coord: getPendingCreation(id)
Coord-->>Panel: "{ sequenceId: N }"
Panel->>Panel: resolveAgentPanelHeaderSequenceId()
Note over Applier,Backfill: On snapshot open (session connects)
Applier->>Applier: syncSessionSequenceFromGraph(graph)
Applier->>Backfill: resolveSequenceIdBackfillForExistingSession()
Backfill-->>Applier: graphSequenceId ?? pendingSequenceId
Applier->>Store: "updateSession(sessionId, { sequenceId })"
Applier->>Coord: completePendingCreation(sessionId)
%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
sequenceDiagram
participant Panel as agent-panel.svelte
participant Store as SessionStore
participant Facade as SessionConnectionFacade
participant Coord as SessionCreationCoordinator
participant Applier as SessionOpenSnapshotApplier
participant Backfill as session-sequence-id-backfill
Note over Panel,Coord: Before session metadata arrives (pending state)
Panel->>Store: getPendingCreationSession(id)
Store->>Facade: getPendingCreationSession(id)
Facade->>Coord: getPendingCreation(id)
Coord-->>Panel: "{ sequenceId: N }"
Panel->>Panel: resolveAgentPanelHeaderSequenceId()
Note over Applier,Backfill: On snapshot open (session connects)
Applier->>Applier: syncSessionSequenceFromGraph(graph)
Applier->>Backfill: resolveSequenceIdBackfillForExistingSession()
Backfill-->>Applier: graphSequenceId ?? pendingSequenceId
Applier->>Store: "updateSession(sessionId, { sequenceId })"
Applier->>Coord: completePendingCreation(sessionId)
Reviews (2): Last reviewed commit: "chore(dead-code): track barrel named imp..." | Re-trigger Greptile
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
Comment on lines
+575
to
+580
| const pendingCreation = sessionStore.getPendingCreationSession(id); | ||
| return resolveAgentPanelHeaderSequenceId({ | ||
| sessionMetadataSequenceId: sessionController.sessionMetadata?.sequenceId, | ||
| pendingCreationSequenceId: pendingCreation?.sequenceId ?? null, | ||
| hasPendingCreationSession: sessionStore.hasPendingCreationSession(id), | ||
| }); |
There was a problem hiding this comment.
getPendingCreationSession(id) already returns null when there is no pending creation, so calling hasPendingCreationSession(id) separately is redundant — pendingCreation !== null is the exact same check. The two calls also go through separate delegation chains, making them subtly susceptible to diverging if the underlying map is mutated between them.
Suggested change
| const pendingCreation = sessionStore.getPendingCreationSession(id); | |
| return resolveAgentPanelHeaderSequenceId({ | |
| sessionMetadataSequenceId: sessionController.sessionMetadata?.sequenceId, | |
| pendingCreationSequenceId: pendingCreation?.sequenceId ?? null, | |
| hasPendingCreationSession: sessionStore.hasPendingCreationSession(id), | |
| }); | |
| const pendingCreation = sessionStore.getPendingCreationSession(id); | |
| return resolveAgentPanelHeaderSequenceId({ | |
| sessionMetadataSequenceId: sessionController.sessionMetadata?.sequenceId, | |
| pendingCreationSequenceId: pendingCreation?.sequenceId ?? null, | |
| hasPendingCreationSession: pendingCreation !== null, | |
| }); |
Resolve export-barrel-only false positives by following named imports through re-export chains, delete the unreachable slash-command section component, and remove the superseded desktop AssistantMessage renderer from the messages barrel. Co-authored-by: Cursor <cursoragent@cursor.com>
There was a problem hiding this comment.
1 issue found across 9 files (changes from recent commits).
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="scripts/dead-code/find-dead-code.ts">
<violation number="1" location="scripts/dead-code/find-dead-code.ts:947">
P2: Aliased named imports are resolved with the local alias instead of the exported symbol name. This undercounts barrel consumption and can produce false dead-code classifications.</violation>
</file>
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
| const token = typePrefix !== null ? typePrefix[1] ?? "" : trimmed; | ||
| const asIndex = token.lastIndexOf(" as "); | ||
| if (asIndex !== -1) { | ||
| names.push(token.slice(asIndex + 4).trim()); |
There was a problem hiding this comment.
P2: Aliased named imports are resolved with the local alias instead of the exported symbol name. This undercounts barrel consumption and can produce false dead-code classifications.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At scripts/dead-code/find-dead-code.ts, line 947:
<comment>Aliased named imports are resolved with the local alias instead of the exported symbol name. This undercounts barrel consumption and can produce false dead-code classifications.</comment>
<file context>
@@ -899,6 +899,267 @@ export function parseGitPorcelain(output: string): Candidate[] {
+ const token = typePrefix !== null ? typePrefix[1] ?? "" : trimmed;
+ const asIndex = token.lastIndexOf(" as ");
+ if (asIndex !== -1) {
+ names.push(token.slice(asIndex + 4).trim());
+ } else {
+ names.push(token.trim());
</file context>
Suggested change
| names.push(token.slice(asIndex + 4).trim()); | |
| names.push(token.slice(0, asIndex).trim()); |
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
1 participant
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.
Summary by cubic
Polishes the assistant Thinking experience with a streaming header and smoother token reveal, stabilizes first‑connect session identity/sequence, and upgrades dead‑code tooling for safer cleanup. Also removes orphan UI renderers and fixes pending model/deferred bind edge cases.
New Features
TOKEN_REVEAL_STEP_MSto 48ms (packages/ui,packages/desktop).resolveResolvableToolbarModelId, and use it in the composer toolbar (packages/desktop).session-sequence-id-backfill, store/applier updates).scripts/dead-code/apply-dead-code.tswith dry‑run and--apply(delete) modes plus optional--include-barrel-only; detectexport-barrel-onlyby following named imports through re‑export chains; remove unusedagent-input-slash-command-section.svelteand the legacy desktopAssistantMessageexport.Bug Fixes
pending_model_idwhen the capability catalog is empty in thecc_sdk_client; adds test coverage.require_creation_attempt_for_deferred_bind; adds tests.Written for commit 7ec96dc. Summary will update on new commits.