sphere

Run a transport test

This is a local research smoke test, not a node, not an external measurement and not E1 evidence. It exercises the classifier and the transport boundary on loopback so a reader can confirm that the published source builds and the protocol plumbing is present.

Prerequisites

Download and verify the signed source snapshot from code and source snapshots. Do not use an unofficial mirror.

base=https://sqhere.org/source/v0.1.0-alpha.6
curl -O "$base/sphere-source-v0.1.0-alpha.6.tar.gz"
curl -O "$base/MANIFEST.json"
curl -O "$base/MANIFEST.json.sig"
curl -O "$base/SHA256SUMS"
curl -O "$base/SHA256SUMS.sig"
curl -O "$base/source-release.pub"
openssl pkeyutl -verify -rawin -pubin -inkey source-release.pub \
  -in MANIFEST.json -sigfile MANIFEST.json.sig
openssl pkeyutl -verify -rawin -pubin -inkey source-release.pub \
  -in SHA256SUMS -sigfile SHA256SUMS.sig
sha256sum -c SHA256SUMS
tar xzf sphere-source-v0.1.0-alpha.6.tar.gz
cd sphere-public-v0.1.0-alpha.6

Build and run the loopback smoke

cargo build --release --locked -p sphere-net --bin sphere-probe
./target/release/sphere-probe serve \
  --ip-a 127.0.0.1 --ip-b 127.0.0.2 --port 40000 &
server=$!
sleep 1
./target/release/sphere-probe health \
  --primary 127.0.0.1:40000 \
  --alternate 127.0.0.2:40001 \
  --allow-local
kill "$server"

A successful run prints a nat-probe/2 line and a local class. Loopback is expected to classify as full-cone; that result describes the local test fixture, not your access network.

For an access-network classification, use the signed peer kit and the run a probe flow. Do not run this local fixture on a network you do not control, and do not use Docker bridge networking: the bridge would classify the container network instead of the access line.

What this does not establish