Skip to content

No nightly on 2018-12-07 #56586

Description

@RalfJung
$ rustup update
info: syncing channel updates for 'stable-x86_64-unknown-linux-gnu'
info: syncing channel updates for 'nightly-x86_64-unknown-linux-gnu'
info: checking for self-updates

   stable-x86_64-unknown-linux-gnu unchanged - rustc 1.31.0 (abe02cefd 2018-12-04)
  nightly-x86_64-unknown-linux-gnu unchanged - rustc 1.32.0-nightly (14997d56a 2018-12-05)

As you can see, the latest nightly still has a commit dating from 2018-12-05, meaning it is the nightly created on 2018-12-06 (shortly after midnight). There was no nightly created today.

Cc @rust-lang/infra

Activity

  1. added
    T-infraRelevant to the infrastructure team, which will review and decide on the PR/issue.
    C-bugCategory: This is a bug.
    on Dec 7, 2018
  2. alexcrichton commented on Dec 7, 2018

    @alexcrichton
    Member
    Dec  7 00:03:38 ip-10-202-181-185 release-nightly: Updating submodule src/tools/miri
    Dec  7 00:03:38 ip-10-202-181-185 release-nightly: fatal: reference is not a tree: 61f20761d3124f5a1b1caee8aa15637cc7f92d8e
    Dec  7 00:03:38 ip-10-202-181-185 release-nightly: Unable to checkout '61f20761d3124f5a1b1caee8aa15637cc7f92d8e' in submodule path 'src/tools/miri'
    Dec  7 00:03:38 ip-10-202-181-185 release-nightly: failed to run: git submodule update --init --recursive src/tools/miri
    

    @RalfJung or @oli-obk was miri force-pushed recently perhaps?

  3. RalfJung commented on Dec 7, 2018

    @RalfJung
    MemberAuthor

    No, but I made a PR updating miri to a commit that's in a miri PR, not in miri master. Because GH adds PR commits to the target repo I thought that should work, and it did work locally and on CI.

    I now made a branch in the miri repo containing that commit, I don't know if having it in a branch vs. a PR makes a difference though (I think it should not).

    Is the release machine using an ancient version of git? Seems like the same problem that @jethrogb had (#56307 (comment)) with git 2.7 (which is almost 3 years old).

  4. o01eg commented on Dec 7, 2018

    @o01eg
    Contributor

    Is the release machine using an ancient version of git? Seems like the same problem that @jethrogb had (#56307 (comment)) with git 2.7 (which is almost 3 years old).

    I got same issue with git 2.19.

  5. jethrogb commented on Dec 7, 2018

    @jethrogb
    Contributor

    This should sort itself out then, now that the commit is available upstream, no? But yeah, it would be good if such things were caught in CI. I think it would be reasonable to configure git in the most conservative way w.r.t. finding commits.

  6. RalfJung commented on Dec 7, 2018

    @RalfJung
    MemberAuthor

    I got same issue with git 2.19.

    Now that is strange. I don't have this issue.

    $ git --version
    git version 2.19.2
    

    How do you reproduce this?

  7. Mark-Simulacrum commented on Dec 7, 2018

    @Mark-Simulacrum
    Member

    git on RCS is currently 2.7.4 so that might explain at least part of the problem. I've filed rust-lang/rust-central-station#119 to try and update the Ubuntu used for RCS...

  8. alexcrichton commented on Dec 8, 2018

    @alexcrichton
    Member

    We got a nightly last night, so closing! I think this is otherwise diagnosed and being fixed

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

    C-bugCategory: This is a bug.T-infraRelevant to the infrastructure team, which will review and decide on the PR/issue.

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions