PostProject 0.7 compatibility evidence¶
Baseline: 0.6.0-alpha.1. Candidate: 0.7.0-alpha.1.
Family |
Last change |
ABI |
blender |
kdenlive |
natron |
obs |
openassetio-manager |
Mechanically eligible |
Eligible with dependencies |
|---|---|---|---|---|---|---|---|---|---|
|
|
51 |
— |
ABI |
— |
ABI |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
ABI |
ABI |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
ABI |
— |
— |
no |
no |
|
|
51 |
ABI |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
ABI |
— |
ABI |
— |
— |
no |
no |
|
|
51 |
— |
— |
ABI |
— |
ABI |
no |
no |
|
|
32 |
— |
projection |
projection |
— |
— |
yes |
yes |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
ABI |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
ABI |
ABI |
ABI |
— |
— |
no |
no |
|
|
51 |
— |
ABI |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
|
|
51 |
— |
— |
— |
— |
— |
no |
no |
The table is generated by tools/compatibility_usage.py from the retained
traces below. ABI reachability is distinct from semantic correctness. Every
changed native family is marked as changed in 0.7 at ABI 51; prior release
eligibility is not inherited. C++ Result propagation retains its source
protocol, but this report grants no new compatibility promise.
The Kdenlive and Blender traces cover the shared-production workflow, including its deliberate conflict/redecision path. The Manager trace covers publishing; its historical label still identifies one test-host route. OBS and Natron inputs are their normal real-host traces. Focused negative and contract runs are separate evidence and do not add independent host counts. The absent second real renew/fail/cancel consumer remains a qualification gap.
Input provenance records
adapter commits, host scope and trace checksums. Reproduce the full report:
git show v0.7.0-alpha.1:docs/compatibility-families.toml > target/release-0.7-families.toml
python3 tools/compatibility_usage.py \
--manifest target/release-0.7-families.toml \
--trace kdenlive=docs/evidence/release-0.7/kdenlive.txt \
--trace blender=docs/evidence/release-0.7/blender.txt \
--trace openassetio-manager=docs/evidence/release-0.7/openassetio-manager.txt \
--trace obs=docs/evidence/release-0.7/obs-normal.txt \
--trace natron=docs/evidence/release-0.7/natron-normal.txt \
--projection kdenlive=docs/evidence/release-0.7/kdenlive-cpp.txt \
--projection natron=docs/evidence/release-0.7/natron-cpp.txt
No source recompilation or layout check is treated as cross-series binary replacement evidence. The released 0.6.x Result propagation promise and its published artifacts remain intact. All changed production APIs are experimental.