Give every platform the Codex plugin, not just Windows - #698
Merged
Conversation
CommitLore — record lintTrailers: clean — 1 commit in Trailer violations fail this check. Active constraints are informational — they are what the repository already decided, not a verdict on this PR. |
MongLong0214
force-pushed
the
fix-691-dead-hosts
branch
from
August 16, 2026 00:43
0ae6cfe to
1a6e8d4
Compare
install.ps1 ran `plugin install-codex`. install.sh carried the same step behind wire_codex_plugin, reachable only from wire_codex, which nothing calls -- so Windows installed the Codex plugin and macOS and Linux did not. The plugin registers the marketplace and installs commitlore@commitlore, the same layer as the Claude plugin; the MCP registration does not cover it. The step goes into the enumeration both installers already call, rather than back into either shell. That is what makes the platforms agree by construction instead of by two copies staying in step, and it is where the Claude plugin already lives. Deliberately narrow. This PR opened by also deleting install.sh's 291 lines of unreachable host wiring, and T-1121 went red: it asserts the same sentences in both installers, and removing the POSIX copy broke a parity that had been holding between a live script and a dead block. That failure is correct and the assertion is not weakened. The deletion belongs with install.ps1's consolidation, because install.ps1's wiring is not the dead twin of install.sh's -- it has ten live Add-Wired call sites and delegates at line 817, so Windows wires hosts twice today. The Codex split is itself evidence that the enumeration and those call sites are not equivalent, which is exactly why deleting them on the assumption that they are would be unfounded. Both land together in #697 with a host-by-host map. Blind review agreed on this scope over finishing the consolidation here. Limit: one owner per host step, in the place both installers call Blast: module Undo: easy Certainty: firm Provenance: authored Record-Id: r-697codx
MongLong0214
force-pushed
the
fix-691-dead-hosts
branch
from
August 16, 2026 01:07
1a6e8d4 to
04bd6f3
Compare
This was referenced Aug 16, 2026
Merged
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #697's user-facing half. #691 unchanged and still open — see below.
install.ps1ranplugin install-codex.install.shcarried the same stepbehind
wire_codex_plugin, reachable only fromwire_codex, which nothingcalls. Windows installed the Codex plugin; macOS and Linux did not. The
plugin registers the marketplace and installs
commitlore@commitlore— the samelayer as the Claude plugin, not something the MCP registration covers.
The step goes into the enumeration both installers already call, rather than back
into either shell. That is what makes the platforms agree by construction rather
than by two copies staying in step, and it is where the Claude plugin already
lives.
Why this PR shrank
It opened by also deleting
install.sh's 291 lines of unreachable host wiring(#691), and
T-1121went red — it asserts the same sentences appear in bothinstallers, and removing the POSIX copy broke a parity that had been holding
between a live script and a dead block.
That failure is correct, and the assertion is not weakened or exempted. The
deletion belongs with
install.ps1's consolidation, becauseinstall.ps1'swiring is not the dead twin of
install.sh's:Windows wires hosts twice today. And the Codex split is itself evidence that the
enumeration and those call sites are not equivalent — which is exactly why
deleting them on the assumption that they are would be unfounded, by an author
who cannot run Windows.
Both land together in #697, with a host-by-host map of those ten sites.
Blind review agreed on this scope over finishing the consolidation here.
68 tests passing across the installer files,
T-1121untouched. Canonicalartifact
2e853534.