You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: .agents/docs/2026-09-12-engine-gaps-after-the-sdk-batch.md
+21Lines changed: 21 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -625,3 +625,24 @@ code, and what was built instead.
625
625
docs/20, its Chinese copy, the CHANGELOG and the #617 record now state the
626
626
three steps; mcpp#621 tracks the diagnostic that would make the failure
627
627
self-explaining.
628
+
629
+
## 13. What landed, and where each claim was measured
630
+
631
+
| repository | change | measured on the published form |
632
+
|---|---|---|
633
+
| mcpp |#619, tag `v2026.9.12.2` at `37c6796a`| 12 CI workflows green on the PR head; release job green for the four platforms, the sealed manifest and `publish-ecosystem`; every archive matches its `.sha256` and is byte-identical on both `xlings-res/mcpp` mirrors; the extracted Linux binary reports `mcpp 2026.9.12.2`, statically linked and stripped |
634
+
| openxlings/xim-pkgindex |#824 (`487deaf`), the bot bump | three `["latest"]` lines on 2026.9.12.2; the artifact `xim-index-487deaf.tar.gz` matches the pointer's size and sha256 and lists the version; `xlings install mcpp@2026.9.12.2` in the sandbox installs it |
635
+
| mcpplibs/mcpp-index |#399 (`8963c6fe`), the `c++-abi` requirement per platform | the macOS leg of the first revision refused the member and produced the per-platform form; the published artifact carries `requires` under `linux` and `macosx`; a local-index package with the same entry is refused under llvm naming both implementations and satisfied by gcc |
636
+
| mcpplibs/openkal-emscripten |#1 (`d5bc117e`), tag `0.1.1`| the package's CI, pinned to 2026.9.12.2, held the task gate in three directions; the GitHub tag archive and the `mcpp-res` GitCode asset are byte-identical (40399 bytes, sha256 `a33359fc…`) |
637
+
| mcpplibs/mcpp-index |#400 (`794a456b`), openkal-emscripten 0.1.1 | the published artifact `mcpp-index-794a456.tar.gz` carries the three entries; `mcpp index update` with the released binary syncs them; in the sandbox the feature without the root table is refused naming the feature, and with it a task starts, runs and joins under node |
638
+
| mcpp |#620 (`1b26d62b`), the bootstrap pin | moved only after the release was published, mirrored and indexed; `check_version_pins.sh` reports building and bootstrapping from 2026.9.12.2 |
639
+
640
+
The sandbox scripts are `verify.sh` (sections 1 to 15, the engine) and
641
+
`verify-e1.sh` (openkal-emscripten through the index), both run with
642
+
`xlings subos use verify-09114 --sandbox` and the CN mirror set on mcpp itself.
643
+
Two CI failures during the sequence were the index pointer's propagation lag and
644
+
not the release: a bootstrap that read `xim@artifact:971571a` while the pointer
645
+
already named `487deaf`, cleared by rerunning the job.
646
+
647
+
Issues #609, #613, #614, #615 and #618 are closed with the measurement each
648
+
resolution rests on. mcpp#621 records the one diagnostic this batch did not add.
0 commit comments