Skip to content

Workqueue - #33

Merged
thanos merged 4 commits into
mainfrom
workqueue
Jul 30, 2026
Merged

Workqueue#33
thanos merged 4 commits into
mainfrom
workqueue

Conversation

@thanos

@thanos thanos commented Jul 30, 2026

Copy link
Copy Markdown
Owner

No description provided.

thanos added 4 commits July 30, 2026 12:43
… paths.

Usage

gnosis scan s3://my-bucket
gnosis scan s3://my-bucket/docs/api --region us-east-1 --no-tui --export
Behavior

Lists keys under the bucket/prefix; skips …/ directory markers
Relative paths = keys under that root (s3://b/docs/a.rs with prefix docs → a.rs)
Reads object bodies (bounded by --max-size)
Same basename excludes as FS (target, node_modules, .git, knowledge.okf)
Default AWS credential chain; optional --region
No Git enrichment for S3
Code

src/connectors/s3.rs — URI parse, list/get, MemoryS3Backend for tests
ScanSource::{Filesystem, S3} on ScanConfig
Pipeline branches FS vs S3 for discover/read

closed #24
Persist analyze_object jobs in redb with scan ids, CLI/TUI controls
(list, pause/stop, purge, rerun), auto-retry backoff, and a TUI
fix so :jobs output is visible.

 - closed #25 add a work queue to process artifacts
 - closed #26 add to the cli and TUI functionality to see jobs and ability to filter by their state. so list and detailed views will be needed
 - closed #27 jobs list should show the timestamp and if completed how log it took
 - closed #28 add gnosis jobs purge <age> (e.g. 5d) that deletes jobs older than that duration.
 - closed #29 add the ability to jobs rerun a list of jobs given by a comma delimited list of job ids
 - closed #30 when we run a scan - we should create a scan id.
 - closed #31 add Pause / unpause / stop for jobs and scans
 - closed #32 add auto retry on failure with progressive backoff
…USTSEC-2026-0098, 0099, 0104).

That old crate came in via aws-sdk-s3’s default rustls feature, which maps to legacy aws-smithy-runtime/tls-rustls → rustls 0.21. Modern TLS already comes from default-https-client.

Fix: depend on aws-sdk-s3 with default-features = false and enable sigv4a, http-1x, default-https-client, rt-tokio (no legacy rustls). Also dropped the stale RUSTSEC-2024-0436 ignore from deny.toml.

Coverage failed the ≥80% lines gate, not the tests. All tests passed; CI exited 1 after writing lcov.info because line cover was ~79.7%.

Biggest holes were Redb control paths (pause / unpause / stop / schedule_retry / fail / reclaim) that only the memory store was exercising.

I added focused tests covering those, plus JobStore::count and small S3 path helpers in pipeline.rs. Locally we’re now at 83.37% lines with --fail-under-lines 80 passing.
@thanos
thanos merged commit b39bb1d into main Jul 30, 2026
11 of 12 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant