8000 Follow-ups from #3482: resume refusals that lost their next step, docs scope, and two weak tests · Issue #3507 · gominimal/minimal · GitHub
[go: up one dir, main page]

Skip to content

Follow-ups from #3482: resume refusals that lost their next step, docs scope, and two weak tests #3507

Open

Description

@norrietaylor

Low-severity points left open on #3482 at e49b92b4. None blocks the merge.

  1. Some resume refusals lost their next step. crates/minimal-client/src/attach.rs (about 914) makes every refusal in minvmd's "could not give the box an address: " family fatal, not only the DNS service case. A refusal because the address range is full, or because the address is quarantined, now ends the command with the bare reason. Before, it reached the guest's BoxHostRowEnded, whose "run the same command again" is the right step for a quarantine that clears. Make only the reasons that name a command fatal, or give each one a step the user can run.
  2. The docs overstate. docs/reference/cli-min.md (about 311-313) says the reason names the command that checks the setup. That holds only for the DNS service case. The paragraph covers attach only, but min box exec and min session exec (exec.rs, about 348) fail on the same refusal with exit 1 and are not mentioned.
  3. One test proves nothing. attach_error_stays_on_screen_after_the_unwind (crates/minimal/src/tests.rs, about 7855) passes on the tree before fix(minimal,minimal-client,minimald): a resume that minvmd refuses shows minvmd's reason, and the #3482. The daemon-side test attach_refusal_sends_an_exit_status_before_the_close covers the fix. Make the client test fail without it, or remove it.
  4. A copied reason. NO_ANSWERER's text is copied by hand in box_registration.rs (about 1038) and vm_host.rs (about 1500). Nothing fails when minvmd's text in crates/minvmd/src/net/answerer.rs changes. Read it from the source, as the prefix test does.

Activity

  1. added
    bugSomething isn't working
    priority:nice-to-haveNice to have. Applied by the inbox-route classifier or a human.
    on Nov 17, 2026
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

    bugSomething isn't workingpriority:nice-to-haveNice to have. Applied by the inbox-route classifier or a human.

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions

      0