test: add run_full_fuzz.sh deep fuzz-campaign helper - #74
Merged
Conversation
Adds fuzz/run_full_fuzz.sh: runs every cargo-fuzz target in libFuzzer fork mode for a per-target budget, persisting the (resumable) corpus and collecting crash artifacts, with a clean/dirty exit status. Safe-by-default for any machine, not just a big workstation: - FORKS defaults to cores-2 (machine stays responsive; FORKS=$(nproc) for all), and the run prints estimated total time + peak RAM (~FORKS x RSS_LIMIT_MB) up front. - When run interactively it waits 5s so a heavy run can be aborted and re-launched with smaller knobs; skipped under nohup/redirection. - Tunables: SECS_PER_TARGET, FORKS, RSS_LIMIT_MB, TARGETS. Quick run: SECS_PER_TARGET=120 FORKS=2 ./fuzz/run_full_fuzz.sh (~14 min, 2 cores). Also gitignores fuzz/*.log (campaign logs) and adds a CONTRIBUTING "Fuzzing" section pointing devs at the CI smoke (fuzz.yml), ad-hoc cargo-fuzz, and this helper. Validated with `bash -n` and a config-arithmetic simulation across machine profiles; not run live (leaving the CPU free). Signed-off-by: Nelson Spence <nelson@projectnavi.ai>
Nelson Spence (Fieldnote-Echo)
requested a review
from Navi Bot (project-navi-bot)
as a code owner
May 26, 2026 14:04
Review Summary by QodoAdd deep fuzz campaign helper script and fuzzing documentation
WalkthroughsDescription• Add run_full_fuzz.sh script for deep fuzzing campaigns across all targets • Script runs libFuzzer in fork mode with configurable per-target budgets • Safe-by-default tuning: FORKS defaults to cores−2, estimates time/RAM upfront • Interactive abort window and environment knobs for laptop-friendly runs • Document fuzzing workflow in CONTRIBUTING.md with script usage examples Diagramflowchart LR
A["Developer"] -->|runs| B["run_full_fuzz.sh"]
B -->|builds| C["cargo fuzz targets"]
C -->|runs fork mode| D["libFuzzer per-target"]
D -->|collects| E["Corpus & Artifacts"]
E -->|reports| F["Campaign Summary"]
G["CONTRIBUTING.md"] -->|documents| B
File Changes1. fuzz/run_full_fuzz.sh
|
Code Review by Qodo
1.
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a comprehensive fuzzing script, fuzz/run_full_fuzz.sh, to run deep fuzzing campaigns across all cargo-fuzz targets, along with updated documentation in CONTRIBUTING.md and gitignore rules. The feedback focuses on improving the script's portability and usability, specifically by adding fallbacks for macOS/BSD systems that lack nproc and the GNU-specific date -Is flag, and by ensuring that interrupting the script (e.g., via Ctrl-C) correctly aborts the entire sequential loop rather than skipping to the next target.
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
![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: f8baafd2bf
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 (@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 (@codex) address that feedback".
Copilot started reviewing on behalf of
Nelson Spence (Fieldnote-Echo)
May 26, 2026 14:28
View session
There was a problem hiding this comment.
Pull request overview
Adds a developer-facing helper for running an exhaustive, resumable local cargo-fuzz campaign across all ordvec fuzz targets, and documents the recommended fuzzing workflow (CI smoke vs. local deep runs).
Changes:
- Add
fuzz/run_full_fuzz.shto run all seven fuzz targets sequentially with libFuzzer fork mode, persistent corpus/artifacts, and an end-of-run artifact summary. - Update
fuzz/.gitignoreto ignore campaign log files. - Extend
CONTRIBUTING.mdwith a Fuzzing section pointing to CI smoke runs and the new local helper.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.
| File | Description |
|---|---|
fuzz/run_full_fuzz.sh |
New deep fuzz-campaign runner script with configurable budgets/forks/RSS caps and crash artifact reporting. |
fuzz/.gitignore |
Ignore *.log produced by the full campaign helper. |
CONTRIBUTING.md |
Document fuzzing workflow and how to run the full local campaign. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Addresses PR #74 review (gemini/qodo/copilot/codex): - Portable CPU count: nproc -> getconf -> BSD/macOS sysctl -> 1, so a missing nproc on macOS no longer leaves NCPU empty under set -u. - Portable timestamps: GNU 'date -Is' -> 'date -u +%Y-%m-%dT%H:%M:%SZ' (a now() helper), which BSD/macOS date supports. - Ctrl-C trap (SIGINT): stops the whole campaign instead of killing one target's fuzzer and marching to the next. - Propagate exit codes: each target's rc is captured; a non-zero run sets any_fail and the final status is non-zero (was only logged before, Codex P1). - Iterate the parsed array (for t in "${_targets[@]}") instead of unquoted word-splitting over $TARGETS. Validated: bash -n, now(), nproc fallback, and the propagate-on-failure logic. Signed-off-by: Nelson Spence <nelson@projectnavi.ai>
project-navi-bot
approved these changes
May 26, 2026
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.
run_full_fuzz.sh— deep fuzz-campaign helper for devsAdds
fuzz/run_full_fuzz.sh: a helper to run a deep local cargo-fuzz campaign across all seven targets. Complements the bounded CI smoke (fuzz.yml) with an exhaustive, resumable run for pre-release / ad-hoc deep fuzzing.What it does: runs each target in libFuzzer fork mode for a per-target wall-clock budget, persists the corpus (cumulative / resumable across runs), collects any crash/oom/timeout/leak artifacts, and exits non-zero if any are found.
Safe-by-default (not just for a big workstation)
The original was tuned for a high-core box; this makes it friendly to laptops/smaller machines:
FORKSdefaults tocores − 2so the machine stays responsive (setFORKS=$(nproc)to use every core).~FORKS × RSS_LIMIT_MB) up front, so you see the commitment (the full default run is ~21h and tens of GB).nohup/redirection so detached campaigns start immediately.SECS_PER_TARGET,FORKS,RSS_LIMIT_MB,TARGETS. Laptop-friendly quick run:SECS_PER_TARGET=120 FORKS=2 ./fuzz/run_full_fuzz.sh # ~14 min, 2 coresAlso
fuzz/.gitignore: ignore*.log(the campaign writesfuzz/full_fuzz_run.log).CONTRIBUTING.md: a Fuzzing section pointing devs at the CI smoke, ad-hoccargo +nightly fuzz run, and this helper.Validation
bash -n(syntax) + a config-arithmetic simulation across machine profiles (beast / laptop-quick / 2-core / 1-core) — confirms safeFORKS, sane time/RAM estimates, no arithmetic errors. Not run live in CI/locally on purpose (the corecargo +nightly fuzz build/fuzz run … -fork=… -max_total_time=…invocation is the same one already proven by a multi-hour campaign).