Skip to content

Follow-ups from Sasha's 0.8.0 pre-release review (#375) #434

Open

Description

@andersalm

These are the follow-ups from Sasha's pre-release review of #375 (comment 6230061841). Each finding was checked against the code or recorded as unverified. Release-relevant dispositions are on #375. This issue keeps the items that wait until after 0.8.0, so they live in one place.

  • Sign-in, sign-up and owner-recovery challenge pools: a few IPs can fill them, and behind a proxy every caller shares one quota bucket. Unverified. 0.8.0 does not set up remote or HTTPS Homes, so exposure is local.
  • Old maintainer key: it is retired, but no revocation is published. Decide how Runtimes and docs mark a retired key as revoked (Replace the maintainer release key and keep signing off the public server #186).
  • Installer trust-root override (Stop Intel Mac before release lookup and pin the verified head CID #398): remove the ELASTOS_MAINTAINER_DID env and flag override in 0.8.1 (option a). The out-of-band check stays as it is documented: compare the installer's printed Maintainer DID: with the README (README "Install from the publisher", docs/INSTALL.md "Compare and change the release signer").
  • Rollback scope: a normal rollback restores only the five journalled files, not the support binaries or migrated principal roots. Check against the Undo journeys that passed on installed alphas.
  • Crash recovery: a crash between claim_start and record_started leaves a permanent repair state.
  • Shutdown: shutdown doesn't stop Runtime process groups it didn't spawn directly.
  • Fresh source-home setup fails (developer path).
  • Passkey sign-in for credentials that are not discoverable: if 0.8.0 ships with this as a known issue rather than a fix, add an account-named sign-in fallback in 0.8.1. Irzhy decides the tradeoff that a name lookup could reveal which accounts exist.

Activity

  1. added this to the After 0.8.0 milestone on Oct 17, 2026
  2. added a commit that references this issue on Oct 17, 2026
  3. andersalm commented on Oct 17, 2026

    @andersalm
    ContributorAuthor

    Passkey fallback disposition, superseding the last checkbox's "if 0.8.0 ships with a known issue": the fallback ships in 0.8.0. Its disclosure is an accepted 0.8.0 limit, recorded on #375 (decision comment). Before any remote, LAN or HTTPS Home access ships, re-evaluate it: account-first sign-in with stronger limits, or remove the fallback once 0.7.1 credentials are migrated. This includes the case where other OS users on the same machine can call the localhost sign-in route.

  4. andersalm commented on Oct 17, 2026

    @andersalm
    ContributorAuthor

    Superseding comment 6241284019: the passkey fallback does not ship in 0.8.0. There are no public 0.7.1 users, only Anders's test accounts (#375). The branch fix/080-passkey-named-fallback (e6589f8cc) stays here as the option if a real migration need appears; its disclosure limit would need to be decided then.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions