mirror of
https://github.com/hanzoai/s3.git
synced 2026-10-30 14:59:09 +00:00
zooqueenandHanzo Dev
ad6662859d
Seven compiled executables were committed to the repo: stress 33 MB s3-rdma-sidecar/demo-server 23 MB s3-rdma-sidecar/sidecar 22 MB test/kafka/kafka-client-loadtest/loadtest 18 MB test/kafka/simple-consumer/simple-consumer 9 MB test/kafka/simple-publisher/simple-publisher 9 MB test/kafka/kafka-client-loadtest/tools/simple-test 9 MB Every one is a Mach-O arm64 executable — a local macOS build, on a repo whose fleet is linux/amd64, so nothing could run them anyway. They are build outputs with their source sitting beside them; each of those packages still builds, which is the proof they were outputs rather than inputs. Nothing references them pre-built (the `stress:` Makefile targets are test targets, not this file). .gitignore did not cover them, which is why they came back. It also still named the pre-rebrand path — /seaweedfs-rdma-sidecar/bin, for a directory now called s3-rdma-sidecar — so it was ignoring nothing at all. Both fixed. This trims the checkout every clone and every CI job pays for. The blobs stay in history; removing them there is a rewrite, and this repo has already been through one — not worth a second without a reason better than disk. Verified: the affected packages build; vet 0 findings, 26 packages ok, 0 failures. Co-authored-by: Hanzo Dev <[email protected]>