Repository navigation
Alignment gate: two-platform installer contract, shared Browser revision 2.1, stale passkey wording - #69
Closed
SashaMIT wants to merge 1 commit into
Closed
Alignment gate: two-platform installer contract, shared Browser revision 2.1, stale passkey wording#69SashaMIT wants to merge 1 commit into
SashaMIT wants to merge 1 commit into
Conversation
…er revision, no stale passkey wording The alignment gate lagged the contracts it guards: Installer. The public installer now serves Linux x86_64/aarch64 and Apple silicon macOS, mapped to the signed release platforms, and fails closed on everything else. The gate still required Linux-only wording and forbade the darwin arm. The rules now pin the contract as it is: the two-platform release lookup in help, the Linux and aarch64-darwin platform mappings, the refusal of non-Apple-silicon Macs, clean `die` on unsupported OS and architecture, and a forbid on any arm beyond the contract (Intel macOS, Windows shells). The banner rule follows the one-command finish: Home opens when a terminal is present, and the command to open it is printed when none is. INSTALL.md states the platform contract in one clause so the docs rule pins a real sentence. README and Getting Started keep their Linux-preview rules: the public endpoint still serves that preview until the candidate is published. Browser lifecycle revision. The breaking revision is 2.1 and lives in the shared elastos-common browser_protocol contract; the adapter imports it. The gate looked for a 2.0 literal inside the adapter. It now pins 2.1 in the shared contract, requires the adapter to expose it on the wire, forbids a private literal in the adapter, and requires Runtime to reject Engine pages that carry any other revision. Passkey. One comment in home-gui shell-windows.js said "passkey"; the ceremony rule is about code, but the wording was stale. Reworded; the rule is unchanged. Every new or changed rule was mutation-tested: breaking its target makes the gate fail under that rule's label, with rg and with the grep fallback. Co-authored-by: Cursor <cursoragent@cursor.com>
Contributor
Author
|
Closing: superseded by 0947d02 on |
5 tasks
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
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.
Small, independent. One commit on
fix/0.7.1-security; touches the gate, one doc clause, one comment. Nothing is weakened: every rule that changed still enforces a contract, and each was mutation-tested (break the target → the gate fails under that rule's label), withrgand with the grep/Python fallback.Installer (was: Linux-only)
The installer serves Linux
x86_64/aarch64and Apple silicon macOS, mapped to the signed release platforms, and fails closed on everything else. The gate still required Linux-only wording and forbade thedarwin)arm. Now it pins the contract as it is:linux) PLATFORM="${ARCH}-linux"andPLATFORM="aarch64-darwin"mappings are present;macOS release lookup requires Apple silicon);die "Unsupported OS"anddie "Unsupported architecture"are present;x86_64-darwin,msys|mingw|cygwin|windows).Banner rule follows the one-command finish:
Opening Home...when a terminal is present,Home is installed. Open it from a terminal:when none is.docs/INSTALL.mdgains one clause — "detects the platform (Linuxx86_64/aarch64or macOS Apple silicon)" — so the docs rule pins a real sentence. README and Getting Started keep their existing Linux-preview rules unchanged: the public endpoint still serves the 0.1.2 Linux preview until the candidate is published, as INSTALL.md says.Browser lifecycle revision (was:
2.0literal in the adapter)The breaking revision is
2.1(122dc4ade) and lives inelastos-common::browser_protocol; the adapter imports it. Rules now:BROWSER_ENGINE_PROTOCOL_VERSION: &str = "2.1"in the shared contract;"protocol_version": BROWSER_ENGINE_PROTOCOL_VERSIONon the wire;BROWSER_ENGINE_PROTOCOL_VERSION: &stror"protocol_version": "…"literal in the adapter;engine_protocol_version != BROWSER_ENGINE_PROTOCOL_VERSIONrejection ingateway_browser.rs.The v2 cleanup binding/result rules are untouched (they already pass).
Passkey
home-gui/shell-windows.js:174comment said "passkey"; reworded to "sign-in". The ceremony rule (nonavigator.credentials/ passkey routes outside Home, System, Inbox, Wallet, providers) is unchanged.Gates on this branch
check-wci-alignment.shOK (rg and fallback);home-entropy-check,browser-entropy-check,home-shell-regression-smoke,home-agent-shell-smoke,people-discovery-smoke,browser-window-close-handshaketest,diff --check,bash -nall green.Not touched here, as agreed: vendor-token drift, release artifact metadata (
home-agent darwin-arm64: artifact size), Carrierelastos-identity 0.6.0expectation, clippyneedless_returninsetup.rs. (Locallyhome-agent-cancellation.test.mjsfails with a Node 22.23 "Unable to deserialize cloned data" runner error, identical on the untouched tip; not an assertion.)Made with Cursor
Part of #88.
Part of #87.