Skip to content

The docs follow where the storage authority, pointer and wheel landed - #1493

Merged
GLinnik21 merged 1 commit into
mainfrom
docs/platform-w8a-claims
Nov 20, 2026
Merged

GLinnik21 merged 1 commit into
mainfrom
docs/platform-w8a-claims

Conversation

@GLinnik21

Copy link
Copy Markdown
Collaborator

This is a docs follow-up to the platform plan (discussion #1460). It corrects the sentences that steps W8a (#1488, #1490, #1491) and W5c (#1492) made false. A doc-claim audit of those four PRs on main at fb35a0e found them. It is prose only, plus one code comment. No attribute, cfg or code line changes.

What changes

Left alone

  • rust-modules/plex/Cargo.toml still describes hostsim as selecting "the simulator's session paths". Nothing in plex reads hostsim since plex's session persistence reads the platform's storage authority instead of the build's CPU #1491. The file is certificate-pinned, and W8c removes the feature.
  • The plan's own baselines are records of the tree they cite:
    • F39's "21 sites";
    • the W8a per-directory counts;
    • the HostEvent list's Edit, LocaleChanged, LowMemory and Launch, which belong to W5b and W9.

Tests

On the committed tree at bda11f0:

  • make test-crate C=plx_plex T=persistence: test result: ok. 101 passed; 0 failed; 0 ignored; 0 measured; 727 filtered out
  • make test-crate C=plx_os T=storage: test result: ok. 150 passed; 0 failed; 0 ignored; 0 measured; 312 filtered out
  • make check-nodef: exit 0
  • ci/check-module-layers.py, ci/check-deps.sh and ci/check-statics.sh are green.
  • ci/test_readonly_generated.py (22 tests) and ci/test_readonly_dependencies.py (64 tests) pass.
  • No touched file is certificate-pinned. Both spellings of each path were checked against ci/proofs/*.json.

🤖 Generated with Claude Code

A doc-claim audit of #1488, #1490, #1491 and #1492 found these sentences
that the tree no longer supports:

- plex's guide and session/persistence.rs's module doc listed
  cleanup_temporaries among the public calls that hand the authority and the
  helper's transport to a *_via form. It reads the authority and dispatches
  through cleanup_temporaries_for(authority), which takes no transport,
  because the helper arm sweeps nothing. Both now name it as the exception.
- docs/module-layers.md marked L15e and L15g "planned" though W5a to W5c
  (#1484, #1487, #1492), W6b (#1489) and the first three W8a PRs (#1488,
  #1490, #1491) have landed. They now follow the L15d row's form.
- docs/platform-ports.md said telemetry's sites follow in their own PR (it
  landed first, as #1490, and installs the account-clear sweep on every
  platform), said the MemoryOnly gate sits above a target_arch selection
  that no longer exists in plex or telemetry, and described Pointer with
  buttons and shown members. An as-landed note now gives W5c's
  Pointer { x, y, edge } and Wheel { dy: i32 } and says why the other two
  are not carried.
- os/src/storage/mod.rs's comment on StoreError's allow(dead_code) gave a
  reason that stopped being true when both helper and files arms became
  compiled on every build. With the allow deleted no variant warns on the
  host or on armv7-unknown-linux-gnueabi, so the comment now says that and
  the attribute is left for the os directory's W8a PR to decide.

Co-Authored-By: Claude <noreply@anthropic.com>
@GLinnik21
GLinnik21 merged commit 7a1e6c5 into main Nov 20, 2026
14 checks passed
@GLinnik21
GLinnik21 deleted the docs/platform-w8a-claims branch November 20, 2026 15:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant