Repository navigation
feat(mlow): complete codec evidence and own oracle tooling - #1447
Conversation
|
Too many files changed for review (196 files, 150 file limit). Bypass the limit by tagging |
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
Important Review skippedToo many files! This PR contains 173 files, which is 23 over the limit of 150. To get a review, reduce the PR to 150 files or fewer by splitting it into smaller PRs or changing its base branch. Upgrade to a paid plan to raise the limit. This review couldn't start because sufficient usage credits or metered capacity aren't available. Add credits or update usage-based reviews in the billing tab, then retry. ⚙️ Run configurationConfiguration used: Repository UI (base), Organization UI (inherited) Review profile: ASSERTIVE Plan: Team Run ID: ⛔ Files ignored due to path filters (23)
📒 Files selected for processing (173)
You can disable this status message by setting the 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. Comment |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 58ce8101c7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
# Conflicts: # Cargo.lock
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 57c10e3b4a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Merging this PR will improve performance by 10.58%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ⚡ | Memory | crypt_payload_one |
86 B | 77 B | +11.69% |
| ⚡ | Memory | audio_unprotect_packet |
86 B | 77 B | +11.69% |
| ⚡ | Memory | sframe_decrypt |
102 B | 93 B | +9.68% |
| ⚡ | Memory | audio_protect_packet |
106 B | 97 B | +9.28% |
Tip
Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.
Comparing mlow-wasm-vectors (910fc7c) with main (dc6c9a3)
Footnotes
-
356 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩
📦 Binary size report
.text per crate
Top movers (cargo-bloat attribution)
Baseline: |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 4b2ed410a3
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 8b8688d2d6
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codspeedbot fix all regressions and noisy benchs |
Pushed
|
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a84c811004
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
The corrected J#10736 search weights keep a different set of survivor tracks, which widens the full-resolution C/E update and costs ~10% more work in `smpl_pitch` (CodSpeed read it as a 15% regression on `codec_stages::pitch_search`). The extra work is the point of the fix, so pay for it elsewhere in the same routine, all of it verified against the existing goldens and the wasm oracle vectors: - Drop the coarse C/E/H snapshots. They were written every call and never read again: three zeroed allocations and three full copies for nothing. - Write stage-1 C straight into the head of the C array the upsample stages expand from, instead of into a second buffer that was copied in. - Build both E1 blocks by filling as they are produced. Every element is written before it is read, so the zeroing a `vec![0.0; n]` pays for was pure overhead. - Memoize the fine-search argmax in a dense array indexed by the (start subframe, segment length, block) triple the blockseg table bounds, rather than a `HashMap`. The hashing cost more than the argmax it saved, and its `RandomState` made the row's cost vary per process. Locally under `codspeed run --mode simulation`, on one machine: 311.8 us before -> 277.9 us, i.e. below the 291.2 us the row measures with the pre-J#10736 weights; instruction time 227.8 us -> 200.0 us. All 131 MLOW tests still pass, bit-exact. `bench_jid_push_phash_form` measured no execution time at all, on this branch and on main: observing `buf.as_bytes()` after the call hands LLVM the pointer and the length but never the bytes, so the formatting stores were dead. Laundering the destination through `black_box` before the write makes the pointer escape; the row now measures 666 ns.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 041bfdc4d7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a47916d47e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9833229f65
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d5789a33fa
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9da4e594e2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 105d6c3d70
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a0093fbd2f
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d558404982
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0bf70ba861
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
2 participants
MLOW fixtures are reproducible against pinned WhatsApp Web J/S wasm captures. This PR completes the codec corpus, replaces repository Python/Bash automation with Rust
cargo xttasks, and keeps WhatsApp-specific oracle execution in unpublished host tooling.unwasm-coresupplies generic static analysis through a pinned Git dependency; WhatSpecwa-storerestores verified captures. Neither enters the application runtime graph. Both producer PRs are merged.cargo xtis a lightweight dispatcher for hashes, descriptors and CI tasks. Capture-backed commands run through the separate releasewhatsapp-oracle-taskpackage. Bases, typed recipes and locked hashes replace eight expanded specs in Git; CI regenerates and uploads the exact expansions. This removes about 122,000 added lines without changing derivation hashes or fixture payloads.cargo xt oracle conformancecombines generated WhatSpec checks, capture-backed tests, J/S rederivation and Rust VoIP/IQ tests. Audio/video trace capture and comparison support future differential adapters; full callback ABIs and end-to-end IQ/signaling equivalence remain explicit coverage gaps. The normal suite ignores 26 slow signaling research scenarios;--slowselects them separately.Local validation: oracle-core lib (68 tests), oracle-task/cli (14 tests) and fast capture-backed integration suites passed via nextest with no missing-capture skips; a real-engine boot validated the worker, mmap, rights and signature bounds. Clippy for all targets of the four tooling crates, formatting, and dependency checks for those crates passed. Heavy threading, rederivation, rustdoc and workspace gates run in CI. Merged main through #1448; the only merge conflict was the sqlite-storage Cargo.lock entries.