vector_vendor: checked-in binary, no network fetch

[Minipada]

Summary

• Replace vector_vendor's file(DOWNLOAD ...) with the official Vector 0.57.0 release tarballs (x86_64 + aarch64) checked into vector_vendor/prebuilt/ as plain committed blobs — same bytes, same SHA256 checksums already pinned in CMakeLists.txt.
• CMakeLists.txt now extracts and installs the Vector binary from the local checked-in tarball; no network access at build time.
• vector_path/VECTOR_PATH override is unchanged.
• Git LFS was evaluated and briefly used, then reverted. Make DC's vendor packages buildfarm-compatible and fix dc_simulation license metadata (prep for #134) #421's original plan explicitly said "no git-lfs"; mid-review I verified that assumption's stated reason (LFS content wouldn't survive bloom's release-tarball export step) doesn't actually hold — traced and empirically tested the real bloom/vcstools code, which does correctly smudge/archive real LFS content. LFS was reverted anyway for a simpler reason: GitHub's free LFS tier is 1GB/month, and each future Vector version bump would add ~106MB of new, non-deduplicated LFS objects — about 9 bumps before hitting that ceiling. A plain committed blob has no such quota; the tradeoff is the same ~106MB added to ordinary git history per bump instead. ADR-0002 documents both the trace (kept for the record) and the final call.
• .pre-commit-config.yaml: scoped exclude on check-added-large-files for the two tarballs (2MB default cap, by design).
• REUSE.toml: license annotation for the vendored tarballs (MPL-2.0, Vector Authors).
• Updated doc/CI comments that described the old download behavior.

Test plan

• git add -A && prek run --all-files --skip build-doc passes (REUSE, large-file, format checks all green)
• Ran ./tools/e2e/scripts/verify_network_isolation.sh (Network-isolated build verification harness for vendor packages #423's harness): colcon build --packages-select aws_sdk_vendor vector_vendor under --network=none shows Finished <<< vector_vendor — vector_vendor builds with zero network access. aws_sdk_vendor still fails on its git clone as expected (tracked separately by aws_sdk_vendor: flattened source, no network fetch #425); network-isolation-check in ci.yaml stays if: false until that lands too.
• Verified both checked-in tarballs' SHA256 match the pre-existing pinned checksums in CMakeLists.txt.
• Ran this in real CI on this PR (build-workspace job) before the LFS revert — full workspace build + colcon test passed building vector_vendor from the checked-in file. One unrelated flaky test (MeasurementMissionOpenRmfTest, a pre-existing dc_measurements lifecycle-node race, reproduced identically on jazzy's own last green commit) passed cleanly on rerun with zero code changes.

Closes #424

🤖 Generated with Claude Code

https://claude.ai/code/session_019hV3DGw8o1QKsYbo8fQaS9

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.93%. Comparing base (6403634) to head (ec17da8).

Additional details and impacted files

@@           Coverage Diff           @@
##            jazzy     #429   +/-   ##
=======================================
  Coverage   70.93%   70.93%           
=======================================
  Files         119      119           
  Lines        7401     7401           
=======================================
  Hits         5249     5249           
  Misses       2152     2152           

Flag	Coverage Δ
cpp-jazzy	70.93% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Minipada]

Superseded by #432 — consolidated with #431 into a single PR so the checked-in binary never has to be added and then removed again across separate rebase-merged commits on jazzy.