[TomCC7]

Linear project: Message Encoding Refactor

dimOS uses the published dimos-cdrgen==0.1.0 directly. The library owns parsing, generation, codecs, ownership checks and native build recipes. This PR adds SDK discovery, a C++ codec header adapter, one consumer test file, a short guide and Linux/macOS CI. Runtime transport changes and built-in message catalogs remain in later PRs.

Validation

At ee4feff4a967ea49c4d965a0ffba86fe9b811db2:

• Four Python/SDK consumer checks passed on Linux after merging main. The real C++ consumer separately compiled, ran and verified its CDR bytes offline (one case, 33.82 seconds), using pinned/hash-verified cached upstream sources. The parent test environment remains free of pip and catkin_pkg.
• Merges main fbe126e741c44e14ee1c23bb17a7e68e2b6a00d6 with an additive merge commit. Resolved only pyproject.toml and uv.lock: preserves main's lcm package rename and retains the pinned generator. Every pre-existing locked package version/source is unchanged; six generator-related packages are added. Offline lock consistency and applicable conflict-file hooks pass.
• Fixes the user-reported main macOS job, which failed because its C++ test environment lacked pip and catkin_pkg. The native test now creates its own build-tool environment with pinned dimos-cdrgen[native]==0.1.0; both main and dedicated CI use the same test setup. The parent test environment remains free of pip/catkin_pkg afterward.
• Native build dependencies stay isolated: the generator's PyYAML pin does not alter the main dependency lock. DIMOS_OFFLINE=1 also applies to provisioning the build tools.
• All applicable repository hooks passed. Earlier strict adapter typing and documentation branding checks remain applicable; this repair changes only the test and its workflow.

macOS was not run locally. The supplied failed job was inspected; replacement CI has not been polled or claimed green. This PR is open for review; merge approval is still required.

Ownership administration

The reviewer ownership requests remain open for maintainers. PyPI permission changes and package publication are outside this assistant's scope per the user's instruction; no administrative or publishing action was performed. Organization-source publication is also unchanged.

Downstream callers must migrate from dimos_message_build and generic dimos.message_codegen imports in their own PR scopes.

File change summary

Diff against merged main fbe126e741c44e14ee1c23bb17a7e68e2b6a00d6 at the head above:

Category	Files	Added	Deleted
SDK and codec adapters	2	46	0
Consumer tests	1	194	0
Documentation	1	49	0
Dependencies and CI	3	45	1
Automated Nix lock normalization	3	binary	binary

The dependency category includes binary-classified uv.lock, excluded from line counts. Total text changes: 334 additions and one deletion. The three Nix lock updates came from autofix.ci commit 9e52cd0a; they are not generator implementation. No upstream .msg snapshots or generated classes are checked in. Root conftest owns shared markers; the dedicated --noconftest commands declare their marker locally.

[codecov]

Codecov Report

❌ Patch coverage is 98.75000% with 1 line in your changes missing coverage. Please review.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
dimos/message_codegen/native_build.py	93.33%	0 Missing and 1 partial ⚠️

@@           Coverage Diff           @@
##             main    #4247   +/-   ##
=======================================
  Coverage   81.96%   81.97%           
=======================================
  Files        1657     1659    +2     
  Lines      154382   154462   +80     
  Branches    12992    12996    +4     
=======================================
+ Hits       126537   126614   +77     
- Misses      24244    24247    +3     
  Partials     3601     3601           

Components	Coverage Δ
Tests	96.84% <100.00%> (+<0.01%)	⬆️

Flag	Coverage Δ
OS-macos-latest	75.85% <72.50%> (-0.01%)	⬇️
OS-ubuntu-24.04-arm	75.62% <72.50%> (-0.01%)	⬇️
OS-ubuntu-latest	75.97% <72.50%> (-0.01%)	⬇️
Py-3.10	75.97% <72.50%> (+<0.01%)	⬆️
Py-3.11	75.96% <72.50%> (+<0.01%)	⬆️
Py-3.12	75.96% <72.50%> (-0.01%)	⬇️
SelfHosted-Large	32.15% <27.50%> (-0.01%)	⬇️
SelfHosted-Linux	40.61% <27.50%> (-0.02%)	⬇️
SelfHosted-macOS	40.93% <61.25%> (+0.01%)	⬆️

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

Files with missing lines	Coverage Δ
dimos/message_codegen/test_integration.py	100.00% <100.00%> (ø)
dimos/message_codegen/native_build.py	93.33% <93.33%> (ø)

... and 5 files with indirect coverage changes

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[github-actions]

go2 replay realtime

Details

Benchmark suite	Current: 8ab9ab3	Previous: 13e4a21	Ratio
first frame wall	9.499 s	9.978 s	0.95
first frame cpu	20.307 s	20.862 s	0.97
run cpu	32.042 s	32.466 s	0.99
run cpu (user)	24.21 s	25.213 s	0.96
run cpu (system)	7.832 s	7.253 s	1.08
peak memory	2719.418 MB	2717.016 MB	1.00
peak threads	397 threads	406 threads	0.98
disk read	23.402 MB	28.266 MB	0.83
disk write	11.297 MB	11.234 MB	1.01
network (transport)	2714.792 MB	2715.069 MB	1.00
network (external rx)	0.122 MB	0.074 MB	1.65
network (external tx)	0.166 MB	0.338 MB	0.49

This comment was automatically generated by workflow using github-action-benchmark.