Commit a53ff64
committed
pin the installer to 2026.8.27.5, not .4
The pin request named 2026.8.27.4 on the assumption it was about to be
released. It was not: .4 shipped 2026-08-27T10:18 and .5 shipped three hours
later, so .4 is the OLDER version, and mcpp already pins .5 as
src/xlings/xlings.cppm kXlingsVersion with check_version_pins.sh enforcing it
across the repo. xim-pkgindex agrees. .5 also has a behaviour .4 lacks --
declarations beat the index during resolution -- so pinning these forks to .4
would have installed a strictly weaker xlings than the rest of the ecosystem
standardises on.
The gap the request pointed at was real, just not where it said: these two CI
workflows were not pinned AT ALL. They ran `curl … | bash` and took whatever
was newest that morning, which is why this is worth doing -- an unpinned
installer means a red run cannot be told apart from an ecosystem change.
Analysis recorded in mcpp-index
.agents/docs/2026-08-30-gbm-cross-repo-closed-loop-plan.md section 15.1 parent dab11e1 commit a53ff64
1 file changed
Lines changed: 1 addition & 1 deletion
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
50 | 50 | | |
51 | 51 | | |
52 | 52 | | |
53 | | - | |
| 53 | + | |
54 | 54 | | |
55 | 55 | | |
56 | 56 | | |
| |||
0 commit comments