You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
feat(cline): add extension and CLI usage adapter - #1977
Adds ccusage cline daily|weekly|monthly|session and Cline aggregation in all-source reports. Discovers Cline task files for VS Code, Insiders, Cursor, Windsurf, and VSCodium on macOS/Linux/Windows, modern CLI session transcripts, and legacy CLI task files. CLINE_DATA_DIR accepts comma-separated roots.
Only self-contained per-request data is counted. Each accepted request must record its own valid timestamp and nonempty model ID, together with explicit input, output, cache-read, and cache-write counters as non-negative integer JSON numbers. Missing counters are never filled with zero, and models are never inferred from task metadata or conversation history. Explicit zero buckets remain valid when the request contains usage; known model IDs without a pricing entry remain available for pricing overrides and missing-pricing reporting.
Excluded data, documented in the Cline guide:
Requests with missing, invalid, fractional, string-valued, or overflowing token counters.
Requests with missing, blank, or unknown model IDs, including records whose model is available only in task metadata.
Requests with missing or invalid timestamps, unfinished records missing required fields, extension messages explicitly marked partial: true, cost-only records, and all-zero usage placeholders.
Every deleted_api_reqs and subagent_usage rollup, even if it contains token totals: these are not individual requests with their original timestamps.
CLI records whose cache-read/write sum exceeds cache-inclusive input or cannot be represented safely. Valid CLI input is normalized by subtracting its separately recorded cache buckets.
Reports intentionally cover eligible requests only; excluded historical data is not reconstructed. A recorded cost never rescues an ineligible record. The shared auto, display, and calculate cost modes continue to apply to accepted records, including valid recorded zero costs. Kilo/Roo extension data and the Pino log source proposed in #1871 remain out of scope.
Task requests require a non-negative integer epoch-millisecond ts in the supported timestamp range. CLI messages use their own integer ts, or a timezone-qualified ISO timestamp string with whole seconds or exactly three fractional digits when ts is absent. An invalid present ts is not rescued by another timestamp, a sidecar file, or file modification time; unsupported timestamp forms are skipped.
Fixtures and regression tests cover each exclusion, metadata-only models, explicit zero buckets, malformed data, cache consistency, and exact accepted token/cost/model/session totals. JSON and table snapshots cover all four focused and unified reports.
Validation (pinned Nix environment):
just fmt
just typecheck
just test — 1,105 Rust tests and 35 Node tests passed (4 Rust tests/examples ignored).
Read Cline editor tasks, CLI session transcripts, and legacy CLI tasks
through a dedicated Rust adapter. Preserve reported costs and request
model history, normalize CLI cache-inclusive input, and expose focused
and unified reports with configuration, documentation, and fixtures.
Builds on the investigation and prior implementation in #1642.
Co-authored-by: Rushikesh Gaikwad <[email protected]>
No actionable comments were generated in the recent review.
Walkthrough
This change adds an experimental Cline source adapter. It discovers extension and CLI data, parses usage records, and supports daily, weekly, monthly, and session reports. It registers Cline with focused and unified commands, adds configuration support, and updates documentation and integration tests.
Changes: Cline usage source
Layer / File(s)
Summary
Discover and parse Cline data rust/adapters/cline/src/paths.rs, rust/adapters/cline/src/parser.rs, rust/adapters/cline/tests/fixtures/*
The adapter discovers supported task and session layouts from default roots or CLINE_DATA_DIR. It parses eligible extension and CLI records and validates model data, timestamps, token counters, cache accounting, and costs.
The adapter loads and prices entries, groups them into daily, weekly, monthly, or session reports, and emits JSON or table output. The workspace and artifact configuration include the new adapter crate.
The CLI accepts Cline report commands and dispatches them to the adapter. Cline options are added to configuration and schema support. Tests cover parsing, date windows, timezone handling, and focused CLI output.
Add Cline to unified reports rust/crates/ccusage-adapter-all/*, rust/crates/ccusage/tests/codex_cli.rs, rust/crates/ccusage/tests/copilot_cli.rs
The unified loader includes Cline and labels its reports as Cline. Tests cover unified outputs and isolate CLINE_DATA_DIR.
Document Cline support docs/guide/cline/index.md, docs/guide/*, docs/.vitepress/config.ts, apps/ccusage/README.md, rust/adapters/cline/README.md
The guides and READMEs describe Cline commands, storage locations, configuration, record eligibility, and cost handling. The guide sidebar links to the Cline page.
sequenceDiagram
participant User
participant ClineCLI
participant ClineAdapter
participant ClineFiles
participant ReportBuilder
User->>ClineCLI: Run a Cline report command
ClineCLI->>ClineAdapter: Dispatch report options
ClineAdapter->>ClineFiles: Discover and read source files
ClineFiles-->>ClineAdapter: Provide usage records
ClineAdapter->>ReportBuilder: Group and summarize entries
ReportBuilder-->>ClineCLI: Return report rows
ClineCLI-->>User: Print JSON or table output
Loading
Possibly related PRs
ccusage/ccusage#1959: Adds per-source options to unified loading and configuration resolution, which Cline registration and configuration tests use.
ccusage/ccusage#2006: Changes agent-and-report configuration precedence and schema support for the command forms added for Cline.
Merge Risk: Minimal · up to f2864
No actionable issue is established from the available evidence; the change is mergeable after normal checks.
Pre-merge checks: 5
Check name
Status
Explanation
Description Check
Check skipped - CodeRabbit’s high-level summary is enabled.
Title check
The title clearly summarizes the main change: adding a Cline usage adapter for extension and CLI data.
Linked Issues check
Issue #243 is closed and provides historical context only. No active directly linked issue remains, so no linked-issue coding requirements apply.
Out of Scope Changes check
The changes add Cline discovery, parsing, reports, CLI/configuration integration, documentation, and tests. The shared report-list update and adapter documentation changes support the integration. The…
Docstring Coverage
Docstring coverage is inconclusive: files from earlier reviews still need to be rechecked within the check's analysis limits. Not blocking.
Full details: Out of Scope Changes check
Explanation
The changes add Cline discovery, parsing, reports, CLI/configuration integration, documentation, and tests. The shared report-list update and adapter documentation changes support the integration. The summary reports no Kilo/Roo or Pino implementation.
Finishing Touches
Generate docstrings
Commit to this branch
Create a new PR
Generate unit tests (beta)
Commit to this branch
Create a new PR
Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Warning
Some tools did not complete. Review the errors below.
Clippy (1.99.1)
Clippy execution failed
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.
File: No configuration file from this repository applied
Commit: Unavailable
Layers: defaults
Review profile: CHILL
Plan: Advanced
Run ID: d6ad763f-6613-4997-8939-d218f52c3416
Commits
Reviewing files that changed from the base of the PR and between c9946e8 and f2864b8.
Files ignored due to path filters: 24
rust/Cargo.lock is excluded by !**/*.lock
rust/adapters/cline/src/snapshots/ccusage_adapter_cline__report__tests__cline_daily_json_report.snap is excluded by !**/*.snap
rust/adapters/cline/src/snapshots/ccusage_adapter_cline__report__tests__cline_monthly_json_report.snap is excluded by !**/*.snap
rust/adapters/cline/src/snapshots/ccusage_adapter_cline__report__tests__cline_session_json_report.snap is excluded by !**/*.snap
rust/adapters/cline/src/snapshots/ccusage_adapter_cline__report__tests__cline_weekly_json_report.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_daily_unified_json.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_daily_unified_table.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_monthly_unified_json.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_monthly_unified_table.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_session_unified_json.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_session_unified_table.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_weekly_unified_json.snap is excluded by !**/*.snap
rust/crates/ccusage-adapter-all/src/snapshots/ccusage_adapter_all__tests__cline_fixture_weekly_unified_table.snap is excluded by !**/*.snap
rust/crates/ccusage-cli-parser/src/snapshots/ccusage_cli_parser__tests__root_help.snap is excluded by !**/*.snap
rust/crates/ccusage-cli-parser/src/snapshots/ccusage_cli_parser__tests__snapshots_representative_cli_parse_shapes.snap is excluded by !**/*.snap
rust/crates/ccusage-config/src/snapshots/ccusage_config__config_schema__tests__snapshots_schema_agent_specific_option_edges.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_daily_json.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_daily_table.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_monthly_json.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_monthly_table.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_session_json.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_session_table.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_weekly_json.snap is excluded by !**/*.snap
rust/crates/ccusage/tests/snapshots/cline_cli__cline_cli_weekly_table.snap is excluded by !**/*.snap
Files selected for processing: 3
rust/crates/ccusage-adapter-all/src/loader.rs
rust/crates/ccusage-adapter-all/src/tests.rs
rust/crates/ccusage/tests/codex_cli.rs
Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 4 remain after this review.
@ryoppippi I’ll review the Cline adapter with these checks:
Source-format parsing for the extension, current CLI transcripts, and legacy CLI task files.
Cache-token accounting and unfinished-request handling.
Historical model and provider attribution.
Cost modes, including recorded zero costs and unknown or unpriced usage.
Focused and unified report coverage for daily, weekly, monthly, and session reports.
🧠 Learnings used
Learnt from: CR
Repo: ccusage/ccusage PR: 0
File: rust/adapters/README.md:0-0
Timestamp: 2026-10-14T20:33:20.330Z
Learning: Applies to rust/adapters/README.md : Applicability reference:
rust/adapters/AGENTS.md: # AGENTS.md - Agent Source Architecture
`README.md` covers adapter architecture, the shared-vs-source boundary, and the crate module shape;
`rust/adapters/opencode/src/` is a compact example of it. This file adds the workflow rules for
changing or adding an adapter.
Source excerpt:
# Agent Adapters
Each agent adapter owns source-specific log discovery, parsing, token mapping,
model mapping, source metadata, and source-specific pricing behavior. Shared
report rendering, date handling, progress, pricing fetcher lifecycle, file
walking, dedupe, and aggregation should stay in `ccusage-core` or
`ccusage-adapter-common` when the source data allows it.
Learnt from: ryoppippi
Repo: ccusage/ccusage PR: 0
File: :0-0
Timestamp: 2026-08-10T19:47:29.939Z
Learning: In `rust/crates/ccusage-core/src/pricing.rs`, `PricingMap::find` and `PricingMap::context_limit` use a fuzzy fallback that selects the longest matching key. Models.dev pricing snapshots must retain published pricing keys so known IDs resolve exactly. Separately priced tier IDs and unversioned IDs must be excluded from this fuzzy fallback through `PricingMap::exact_only`; otherwise they can incorrectly price a related model identifier.
⚠️ Action not completed
Pull request is closed.
Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.
This compares the Rust PR release binary against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
574.8ms
1.75 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
492.2ms
2.05 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
213.6ms
4.71 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
191.8ms
5.25 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
59.1ms
24.0ms
2.47x
56.25 MiB
25.23 MiB
0.45x
0.03 MiB/s
0.06 MiB/s
claude session --offline --json
0.00 MiB
55.7ms
18.4ms
3.03x
55.75 MiB
25.23 MiB
0.45x
0.03 MiB/s
0.08 MiB/s
codex daily --offline --json
0.00 MiB
47.8ms
17.6ms
2.72x
56.00 MiB
24.98 MiB
0.45x
0.02 MiB/s
0.05 MiB/s
codex session --offline --json
0.00 MiB
49.7ms
17.2ms
2.89x
55.50 MiB
24.98 MiB
0.45x
0.02 MiB/s
0.05 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
460.0ms
435.7ms
1.06x
907.09 MiB
905.10 MiB
1.00x
2.19 GiB/s
2.31 GiB/s
codex --offline --json
1.01 GiB
233.8ms
201.7ms
1.16x
471.44 MiB
465.18 MiB
0.99x
4.31 GiB/s
4.99 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
24.02 KiB
24.98 KiB
+0.96 KiB
0.96x
installed native package binary
4886.47 KiB
4939.59 KiB
+53.13 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
This compares the PR package against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
373.6ms
2.69 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
357.7ms
2.81 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
144.3ms
6.98 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
128.6ms
7.83 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
42.0ms
42.9ms
0.98x
55.75 MiB
55.50 MiB
1.00x
0.04 MiB/s
0.04 MiB/s
claude session --offline --json
0.00 MiB
41.6ms
37.0ms
1.12x
56.00 MiB
56.00 MiB
1.00x
0.04 MiB/s
0.04 MiB/s
codex daily --offline --json
0.00 MiB
34.8ms
35.0ms
0.99x
56.00 MiB
55.50 MiB
0.99x
0.02 MiB/s
0.02 MiB/s
codex session --offline --json
0.00 MiB
37.3ms
35.1ms
1.06x
56.00 MiB
55.75 MiB
1.00x
0.02 MiB/s
0.02 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
374.6ms
375.2ms
1.00x
935.35 MiB
937.36 MiB
1.00x
2.69 GiB/s
2.68 GiB/s
codex --offline --json
1.01 GiB
145.1ms
140.7ms
1.03x
509.43 MiB
517.44 MiB
1.02x
6.94 GiB/s
7.16 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
24.02 KiB
24.98 KiB
+0.96 KiB
0.96x
installed native package binary
4886.47 KiB
4939.59 KiB
+53.13 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
The reason will be displayed to describe this comment to others. Learn more.
ℹ️ No critical issues — two small follow-ups inline.
Reviewed changes
Added a Cline adapter reading extension tasks, legacy CLI tasks, and CLI session transcripts, with CLINE_DATA_DIR overrides and task: / cli: session namespacing.
Wired ccusage cline daily|weekly|monthly|session, unified all-source loading, config schema, docs, and help snapshots.
Added fixture-backed unit, snapshot, and CLI integration tests; verified the extension-vs-CLI token asymmetry against upstream semantics and confirmed session-ID prefixing is opaque to shared consumers.
Verified one upstream claim via live sources (extension tokensIn is exclusive of cache; CLI inputTokens is inclusive, so the asymmetric subtraction is correct). Upstream discovery coverage (extension id, layouts, editor list) could not be re-verified live, so the experimental docs label is doing real work here.
Count only requests with explicit integer counters, a recorded model ID,
and a valid request timestamp. Remove task metadata model inference and
exclude deleted-request and subagent rollups instead of assigning them
request timing or model information.
Reject incomplete, partial, unknown-model, all-zero, and inconsistent or
overflowing usage records. Document the exclusions and cover them with
fixtures plus focused and unified report assertions and snapshots.
This compares the Rust PR release binary against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
391.2ms
2.57 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
371.6ms
2.71 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
149.0ms
6.76 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
119.4ms
8.43 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
45.5ms
20.7ms
2.20x
56.25 MiB
25.21 MiB
0.45x
0.03 MiB/s
0.07 MiB/s
claude session --offline --json
0.00 MiB
37.9ms
13.5ms
2.82x
56.00 MiB
24.96 MiB
0.45x
0.04 MiB/s
0.11 MiB/s
codex daily --offline --json
0.00 MiB
40.7ms
13.5ms
3.02x
56.00 MiB
24.96 MiB
0.45x
0.02 MiB/s
0.06 MiB/s
codex session --offline --json
0.00 MiB
37.5ms
14.3ms
2.61x
56.00 MiB
24.96 MiB
0.45x
0.02 MiB/s
0.06 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
379.5ms
371.8ms
1.02x
909.34 MiB
923.34 MiB
1.02x
2.65 GiB/s
2.71 GiB/s
codex --offline --json
1.01 GiB
144.1ms
127.3ms
1.13x
505.44 MiB
501.43 MiB
0.99x
6.99 GiB/s
7.91 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
24.02 KiB
24.98 KiB
+0.96 KiB
0.96x
installed native package binary
4886.47 KiB
4932.53 KiB
+46.06 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
This compares the PR package against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
689.3ms
1.46 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
659.5ms
1.53 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
312.0ms
3.23 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
258.4ms
3.90 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
78.9ms
83.5ms
0.94x
56.25 MiB
55.75 MiB
0.99x
0.02 MiB/s
0.02 MiB/s
claude session --offline --json
0.00 MiB
74.5ms
71.8ms
1.04x
55.75 MiB
55.75 MiB
1.00x
0.02 MiB/s
0.02 MiB/s
codex daily --offline --json
0.00 MiB
79.6ms
76.7ms
1.04x
56.00 MiB
56.25 MiB
1.00x
0.01 MiB/s
0.01 MiB/s
codex session --offline --json
0.00 MiB
76.1ms
73.5ms
1.04x
56.00 MiB
55.75 MiB
1.00x
0.01 MiB/s
0.01 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
719.0ms
677.1ms
1.06x
911.09 MiB
931.32 MiB
1.02x
1.40 GiB/s
1.49 GiB/s
codex --offline --json
1.01 GiB
330.2ms
349.9ms
0.94x
465.43 MiB
483.42 MiB
1.04x
3.05 GiB/s
2.88 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
24.02 KiB
24.98 KiB
+0.96 KiB
0.96x
installed native package binary
4886.47 KiB
4932.53 KiB
+46.06 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
The reason will be displayed to describe this comment to others. Learn more.
3 issues found across 74 files
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. When an issue isn't valid or won't be fixed in this PR, reply in its thread with the reason and then resolve the thread. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="rust/crates/ccusage-cli/src/lib.rs">
<violation number="1" location="rust/crates/ccusage-cli/src/lib.rs:6">
P3: `STANDARD_WEEKLY_AGENT_REPORTS` is newly exported here, but the crate README's exhaustive "Public surface" list in `rust/crates/ccusage-cli/README.md` does not include it. Add `- `types::STANDARD_WEEKLY_AGENT_REPORTS`` after the `STANDARD_AGENT_REPORTS` entry so the docs stay in sync.</violation>
</file>
<file name="rust/crates/ccusage-adapter-all/README.md">
<violation number="1" location="rust/crates/ccusage-adapter-all/README.md:13">
P3: The count is still wrong: `Cargo.toml` depends on 19 adapters (excluding `ccusage-adapter-common`) once Cline is included, so this should read 19. The "Depends on" list also omits `ccusage-adapter-antigravity`, which the crate depends on and uses in `loader.rs`.</violation>
</file>
<file name="rust/crates/ccusage-cli/src/types.rs">
<violation number="1" location="rust/crates/ccusage-cli/src/types.rs:229">
P3: `STANDARD_WEEKLY_AGENT_REPORTS` duplicates the existing `OPENCODE_AGENT_REPORTS` exactly. Reuse `OPENCODE_AGENT_REPORTS` for Cline and drop this constant and its re-export. If a separate name is wanted, make it a distinct name that describes the purpose. Also add it to the public surface list in the crate README if it stays.</violation>
</file>
Reply with feedback, questions, or to request a fix.
The reason will be displayed to describe this comment to others. Learn more.
Reviewed changes
Tightened extension records since the prior review to completed api_req_started events with their own timestamp, model, and all four integer counters; rollups, partial records, all-zero usage, and metadata-only model attribution are now excluded.
Tightened CLI records to assistant messages with valid timestamps, complete metrics, and non-unknown models; cache-inconsistent or overflowing totals are skipped via checked arithmetic instead of clamping.
Documented the eligible-requests-only coverage and kept fixtures, regression tests, and focused plus unified snapshots aligned with the stricter totals.
ℹ️ No new issues in the incremental delta — prior minor threads remain open.
No new findings in commit 80ffd4c. Two minor threads from the prior review (session-prefix docs note, adapter-count README) are unchanged by these commits and left open for the author, so this review stays non-approving.
The reason will be displayed to describe this comment to others. Learn more.
ℹ️ No new issues in the incremental delta — merge-only update verified clean.
Reviewed changes
Merged main into the branch, picking up shared refactors (format_date_tz rename, BUILT_IN_AGENT_NAMES move, AgentRows::warnings plumbing, test isolation).
Confirmed Cline wiring survived the merge with its report index, loader spec, and date helper signatures intact.
Confirmed the two prior minor threads remain open and unchanged by this merge.
No new findings in commit 3bf4230. Two minor threads from the prior review (session-prefix docs note, adapter-count README) are unchanged by these commits and left open for the author, so this review stays non-approving.
This compares the PR package against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
391.1ms
2.57 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
370.1ms
2.72 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
161.9ms
6.22 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
132.3ms
7.61 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
39.0ms
39.5ms
0.99x
55.50 MiB
55.75 MiB
1.00x
0.04 MiB/s
0.04 MiB/s
claude session --offline --json
0.00 MiB
36.3ms
36.5ms
0.99x
56.00 MiB
55.75 MiB
1.00x
0.04 MiB/s
0.04 MiB/s
codex daily --offline --json
0.00 MiB
37.1ms
39.0ms
0.95x
55.75 MiB
56.00 MiB
1.00x
0.02 MiB/s
0.02 MiB/s
codex session --offline --json
0.00 MiB
39.6ms
40.0ms
0.99x
56.00 MiB
56.00 MiB
1.00x
0.02 MiB/s
0.02 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
393.2ms
389.1ms
1.01x
913.09 MiB
929.10 MiB
1.02x
2.56 GiB/s
2.59 GiB/s
codex --offline --json
1.01 GiB
157.1ms
156.4ms
1.00x
529.44 MiB
521.43 MiB
0.98x
6.41 GiB/s
6.44 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
27.67 KiB
28.77 KiB
+1.10 KiB
0.96x
installed native package binary
4994.97 KiB
5045.66 KiB
+50.69 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
This compares the Rust PR release binary against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
400.4ms
2.51 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
378.4ms
2.66 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
168.2ms
5.99 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
147.7ms
6.82 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
41.3ms
16.3ms
2.53x
56.00 MiB
22.97 MiB
0.41x
0.04 MiB/s
0.09 MiB/s
claude session --offline --json
0.00 MiB
44.0ms
17.3ms
2.55x
56.00 MiB
23.22 MiB
0.41x
0.04 MiB/s
0.09 MiB/s
codex daily --offline --json
0.00 MiB
39.7ms
16.9ms
2.35x
55.75 MiB
27.22 MiB
0.49x
0.02 MiB/s
0.05 MiB/s
codex session --offline --json
0.00 MiB
40.7ms
16.6ms
2.46x
55.75 MiB
27.22 MiB
0.49x
0.02 MiB/s
0.05 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
404.3ms
391.2ms
1.03x
937.10 MiB
931.09 MiB
0.99x
2.49 GiB/s
2.57 GiB/s
codex --offline --json
1.01 GiB
170.7ms
135.2ms
1.26x
523.44 MiB
477.43 MiB
0.91x
5.90 GiB/s
7.44 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
27.67 KiB
28.77 KiB
+1.10 KiB
0.96x
installed native package binary
4994.97 KiB
5045.66 KiB
+50.69 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
Cline session reports grouped rows with summarize_by_key, which only
tracks token and cost totals, and report_from_rows always passed
include_session_metadata=false. Session JSON therefore never carried
firstActivity, lastActivity, or projectPath, unlike the other session
adapters, and the unified report got no lastActivity metadata for Cline
sessions.
Build session rows with SessionAccumulator, as zcode, qwen, and pi do,
and emit session metadata for session reports. A unit test covers a
session whose entries arrive out of order, and the session snapshots now
include the new fields.
STANDARD_WEEKLY_AGENT_REPORTS was an exact copy of
OPENCODE_AGENT_REPORTS. The copy existed because the original was named
after one agent rather than the report set it describes, so a second
agent with the same reports could not reuse it by name.
Rename OPENCODE_AGENT_REPORTS to ALL_AGENT_REPORTS, since it lists every
AgentReportKind, and use it for both OpenCode and Cline. The matching
arms in agent_report_supported are merged too, and the ccusage-cli
public-surface list is updated to the new name.
ccusage-adapter-all said it depends on "all 18 adapters" and ccusage-cli
said "all 17 adapters". Cargo.toml now lists 19 source adapters, so both
numbers were stale, and a fixed number goes stale again whenever an
adapter is added. Say "every source adapter" / "every adapter" instead.
The ccusage-adapter-all "Depends on" list was also missing
ccusage-adapter-antigravity, which the crate depends on and loads.
Session reports show IDs such as task:<id> and cli:<id>, but the guide
never explained the prefixes. Document that task: marks extension and
legacy CLI tasks and cli: marks CLI session transcripts, so a task and a
CLI session that share an ID stay separate rows.
The adapter README and the guide's storage table read as if each
platform searched only its own config directory. default_roots() checks
every candidate on every platform and skips missing ones. Say so, and
mention the %USERPROFILE%\AppData\Roaming fallback in the README.
The isolation helpers pin HOME and USERPROFILE to the fixture and clear
CLINE_DATA_DIR, but the Cline adapter reads APPDATA before falling back
to USERPROFILE for its Windows extension roots. On a Windows machine
these tests could pick up real Cline data from %APPDATA%.
Clear APPDATA in both helpers so Cline discovery falls back to the
fixture's USERPROFILE.
The shared Cline fixture totals $0.0012, so every cost cell in the table
snapshots renders as $0.00 and a broken table cost path would not change
them. Add a fixture with $1.25 and $2.50 recorded costs and assert that
the Total row of each report kind renders $3.75.
c9946e8: a table test asserts that recorded costs render ($3.75).
Validation: just fmt, just test, just check (77 flake checks), just typecheck, and just hawk all pass, and the pre-push clippy and cargo test hooks passed.
No re-review warranted. HEAD (c9946e8) is identical to the prior pullfrog approval (review 5832111405), so there is no incremental delta. Both pullfrog-originated threads from the initial review are marked RESOLVED, and the six fixup commits were already covered by that approval.
The reason will be displayed to describe this comment to others. Learn more.
✅ No new issues found.
Reviewed changes (since prior pullfrog review at 3bf4230)
Fixed session reporting to carry activity bounds and project path via SessionAccumulator, and enabled session metadata in session JSON rows, with a regression test and refreshed snapshots.
Consolidated the duplicated weekly report lists into a single ALL_AGENT_REPORTS shared by cline and opencode, and merged the agent_report_supported arms.
Documented the task: / cli: session-ID prefixes and the every-platform candidate discovery in the Cline guide and adapter README.
Reworded both crate READMEs to avoid hard-coded adapter counts and added the missing antigravity dependency entry.
Hardened test isolation by clearing APPDATA in unified loader tests.
Added a table-report test asserting recorded costs render in the Total row.
This compares the Rust PR release binary against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
460.7ms
2.19 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
429.6ms
2.34 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
179.2ms
5.62 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
283.8ms
3.55 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
69.9ms
32.6ms
2.14x
55.50 MiB
22.97 MiB
0.41x
0.02 MiB/s
0.05 MiB/s
claude session --offline --json
0.00 MiB
47.0ms
18.6ms
2.52x
55.75 MiB
23.22 MiB
0.42x
0.03 MiB/s
0.08 MiB/s
codex daily --offline --json
0.00 MiB
46.3ms
27.0ms
1.72x
55.75 MiB
25.23 MiB
0.45x
0.02 MiB/s
0.03 MiB/s
codex session --offline --json
0.00 MiB
92.7ms
44.7ms
2.07x
55.50 MiB
25.23 MiB
0.45x
0.01 MiB/s
0.02 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published native ccusage binary from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
679.1ms
560.3ms
1.21x
885.09 MiB
937.10 MiB
1.06x
1.48 GiB/s
1.80 GiB/s
codex --offline --json
1.01 GiB
238.6ms
197.7ms
1.21x
483.43 MiB
457.41 MiB
0.95x
4.22 GiB/s
5.09 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
27.67 KiB
28.77 KiB
+1.10 KiB
0.96x
installed native package binary
4994.97 KiB
5047.28 KiB
+52.31 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
@ryoppippi I’ll request a review of the latest push, including the reported fixes and the Cline adapter checks.
Action not completed
Already reviewed the last commit. Use @coderabbitai full review to rerun a review of the entire changeset.
Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.
This compares the PR package against the configured base package on the same CI runner.
Package runtime diagnostics
Compares the PR package wrapper, the installed native optional dependency binary, and the workspace release binary on the same large fixture. This identifies whether slow package results come from JavaScript wrapper overhead, the published native binary build, or the Rust core itself.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
All rows run --offline --json, measured by hyperfine with 0 warmups and 1 runs. This isolates wrapper overhead from the installed native optional dependency and the workspace release binary built on the runner.
Command
Runtime
Input
Median
Throughput
Samples
claude --offline --json
Package wrapper
1.01 GiB
598.5ms
1.68 GiB/s
1
claude --offline --json
Installed native binary
1.01 GiB
550.6ms
1.83 GiB/s
1
codex --offline --json
Package wrapper
1.01 GiB
299.0ms
3.37 GiB/s
1
codex --offline --json
Installed native binary
1.01 GiB
236.0ms
4.27 GiB/s
1
Committed fixture performance
Committed small fixtures for stable PR-to-PR feedback and explicit Claude/Codex command coverage.
Fixtures: Claude apps/ccusage/test/fixtures/claude (0.00 MiB, 2 files), Codex apps/ccusage/test/fixtures/codex (0.00 MiB, 1 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 2 warmups and 7 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude daily --offline --json
0.00 MiB
71.8ms
68.2ms
1.05x
55.75 MiB
55.75 MiB
1.00x
0.02 MiB/s
0.02 MiB/s
claude session --offline --json
0.00 MiB
61.0ms
65.8ms
0.93x
56.00 MiB
56.00 MiB
1.00x
0.03 MiB/s
0.02 MiB/s
codex daily --offline --json
0.00 MiB
79.4ms
78.0ms
1.02x
56.00 MiB
56.00 MiB
1.00x
0.01 MiB/s
0.01 MiB/s
codex session --offline --json
0.00 MiB
61.3ms
66.4ms
0.92x
55.75 MiB
56.25 MiB
1.01x
0.01 MiB/s
0.01 MiB/s
Large real-world-shaped fixture performance
Generated fixtures shaped from aggregate local log statistics: thousands of JSONL files, many small sessions, and a long tail of larger sessions. No real prompts, paths, or outputs are stored in the fixtures.
Fixtures: Claude /home/runner/_work/_temp/ccusage-large-fixture (1.01 GiB, 2597 files), Codex /home/runner/_work/_temp/ccusage-large-codex-fixture (1.01 GiB, 2597 files)
Base runs the published ccusage package from pkg.pr.new, installed before measurement; PR runs the published ccusage package from pkg.pr.new, installed before measurement. Both run --offline --json, measured by hyperfine with 0 warmups and 1 runs.
Peak RSS is measured separately with /usr/bin/time using 1 runs. Lower RSS ratios are better.
Command
Input
Base median
PR median
PR vs base
Base peak RSS
PR peak RSS
PR/base RSS
Base throughput
PR throughput
claude --offline --json
1.01 GiB
526.0ms
540.1ms
0.97x
913.09 MiB
915.09 MiB
1.00x
1.91 GiB/s
1.86 GiB/s
codex --offline --json
1.01 GiB
268.1ms
265.4ms
1.01x
497.44 MiB
481.44 MiB
0.97x
3.76 GiB/s
3.79 GiB/s
Artifact size
Artifact
Base
PR
Delta
Ratio
packed ccusage-*.tgz
27.67 KiB
28.77 KiB
+1.10 KiB
0.96x
installed native package binary
4994.97 KiB
5047.28 KiB
+52.31 KiB
0.99x
Lower medians and smaller artifacts are better. CI runner noise still applies; use same-run ratios as directional PR feedback, not release guarantees.
Verdict: no significant change (every command within ±10% CPU time and peak RSS).
Fixture: generated Claude logs (1.01 GiB on disk, 2597 files) and Codex logs (295.00 MiB on disk, 2597 files, mostly zstd-compressed) shaped like current real logs; every value is synthetic.
Base and PR are built from source the same way on this runner and run --offline --json in 11 interleaved rounds after 2 warmup rounds; each round runs both binaries once in alternating order. Cells show median (min–max). Ratios are PR/base per round, so below 1.00 is better. A command is flagged when the median CPU-time or peak-RSS ratio moves past 10% and at least four in five rounds agree. Wall time is shown but not judged.
Upstream tip: 660e07c (2026-11-13,
chore(pricing): update models.dev snapshot)
Previous fork base: 30b653f (the upstream commit this fork was synced to)
Synced 100 upstream commits on top of 30b653f: 53 automated pricing
snapshot refreshes, 23 dependency bumps, 2 releases, and:
3541115 feat(cline): add extension and CLI usage adapter (ccusage#1977)
5e5c475 fix(antigravity): report real model names instead of placeholders (ccusage#2071)
e39ed83 fix(claude): keep the cache TTL split when deduping equal copies (ccusage#2060)
c17c591 feat(adapter): report 1h cache writes in JSON model breakdowns (ccusage#2059)
a06f4bc perf(claude): halve `ccusage claude` report time (ccusage#2058)
de43952 fix(claude): show model breakdowns for session --id (ccusage#2057)
5f31039 fix(all): count only in-window Claude usage in unified session reports (ccusage#2052)
e4f0ea9 fix(codex): count spawned subagent requests stamped at the fork instant (ccusage#2051)
b85b28f fix(opencode): count usage from completed v2 compaction rows (ccusage#2050)
00fead0 fix(antigravity): deduplicate unkeyed mirror usage (ccusage#2027)
2248106 fix(opencode): exclude Claude Agent turns by default (ccusage#2022)
4fb1c0d fix(codex): guard truncated type fields from panicking the parser (ccusage#2034)
86707db build(perf): make the PR performance comparison catch real regressions (ccusage#2061)
686a1a8 ci: support a merge queue on main (ccusage#2062)
Fork wins kept: the Codex optimized scanner with its fork-replay instant,
the Antigravity model resolver, the OpenCode retained-ledger cache with its
per-namespace rows, the SQLite on-disk cache, the table renderer, the fork
CI and tag-only release workflow, and the exact `=` dependency pins.
Hand-resolved conflicts:
rust/adapters/claude/src/lib.rs: the fork's DedupedEntries copy pairs stay
and take upstream's carries_more_usage_detail tiebreak (speed rank, then
a cache TTL split); main_thread_session_days reads the recorded session
of each copy.
rust/adapters/opencode/{loader,parser}.rs: the fork's retain_via_ledger
stays beside upstream's v2 compaction scan; SessionMessageKind and
MessageRowColumns carry the row type through the five-value scan.
rust/adapters/codex/{replay,parser}.rs: open_rollout_reader_with_capacity
and CodexForkReplay both land in the shared parser state.
rust/crates/ccusage-core/src/{cache,pricing}.rs: the opencode-db-compaction
namespace joins the ledger namespaces; the fork's embedded pricing loader
and OnceLock path stay.
rust/crates/ccusage/src/cli/timezone.rs: Cline joins the shared-args arm,
the ClearCache arm stays unreachable from the parser.
rust/Cargo.toml: the fork's exact `=` pins stay; jiff 0.2.41,
schemars 1.2.2, serde 1.0.229, smallvec 1.16.5, toml 1.2.1.
.github/workflows/{ci,release}.yaml: the fork's design is kept, including
its own release flow. Upstream's merge-queue, tagpr and pkg.pr.new jobs
need scripts this fork deleted, so they are not taken; the two
save-nix-cache inputs upstream passes to the native-package actions are
dropped because the fork's composite actions declare no such input.
packages/ccusage-*/package.json (6 files), the contribution-gate scripts,
sponsors.md, blacksmith.png and rust/adapters/claude/src/daily.rs stay
deleted; the two Blacksmith SVG logos upstream added are not taken.
Docs: the agent lists take upstream's Cline addition but keep this fork's
`pi` name, because the CLI subcommand is still `pi`; upstream renamed the
docs to `pi-agent` without renaming the command.
Snapshots regenerated for the fork table renderer: the Cline adapter, CLI
and unified reports, codex issue_2046, opencode compaction, claude session
--id and the config schema (new Cline section).
Validation: cargo test --workspace --offline 1297 passed, 0 failed; cargo
clippy --workspace --all-targets clean; rustfmt --check clean.
Version: take upstream's 20.0.50 in package.json, apps/ccusage/package.json
and docs/package.json; the Rust workspace keeps version 0.0.0.
DO NOT push a v20.0.50 tag from this fork: release.yaml would publish fork
bytes under upstream identity.
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
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.
![pullfrog[bot]](https://avatars.githubusercontent.com/in/1768019?s=60&v=4){kind=small}
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
Adds
ccusage cline daily|weekly|monthly|sessionand Cline aggregation in all-source reports. Discovers Cline task files for VS Code, Insiders, Cursor, Windsurf, and VSCodium on macOS/Linux/Windows, modern CLI session transcripts, and legacy CLI task files.CLINE_DATA_DIRaccepts comma-separated roots.Only self-contained per-request data is counted. Each accepted request must record its own valid timestamp and nonempty model ID, together with explicit input, output, cache-read, and cache-write counters as non-negative integer JSON numbers. Missing counters are never filled with zero, and models are never inferred from task metadata or conversation history. Explicit zero buckets remain valid when the request contains usage; known model IDs without a pricing entry remain available for pricing overrides and missing-pricing reporting.
Excluded data, documented in the Cline guide:
unknownmodel IDs, including records whose model is available only in task metadata.partial: true, cost-only records, and all-zero usage placeholders.deleted_api_reqsandsubagent_usagerollup, even if it contains token totals: these are not individual requests with their original timestamps.Reports intentionally cover eligible requests only; excluded historical data is not reconstructed. A recorded cost never rescues an ineligible record. The shared
auto,display, andcalculatecost modes continue to apply to accepted records, including valid recorded zero costs. Kilo/Roo extension data and the Pino log source proposed in #1871 remain out of scope.Task requests require a non-negative integer epoch-millisecond
tsin the supported timestamp range. CLI messages use their own integerts, or a timezone-qualified ISOtimestampstring with whole seconds or exactly three fractional digits whentsis absent. An invalid presenttsis not rescued by another timestamp, a sidecar file, or file modification time; unsupported timestamp forms are skipped.Fixtures and regression tests cover each exclusion, metadata-only models, explicit zero buckets, malformed data, cache consistency, and exact accepted token/cost/model/session totals. JSON and table snapshots cover all four focused and unified reports.
Validation (pinned Nix environment):
just fmtjust typecheckjust test— 1,105 Rust tests and 35 Node tests passed (4 Rust tests/examples ignored).cargo clippy --manifest-path rust/Cargo.toml --workspace --all-targets -- -D warningsjust check— all 12 host flake checks passed on aarch64-darwin.Validation is fixture-based; no local Cline installation was available for a real-data smoke test. No pricing snapshot files were changed.
Reviewed prior PR #1642 and credited Rushikesh Gaikwad (@Wraient) as a co-author. Format references: Cline message types, CodeBurn CLI parser, and legacy CLI task storage.
Closes #243
Refs #1871
Refs #1642
Summary by CodeRabbit
CLINE_DATA_DIR.