Tests failing for mercurial changed files

[lucasfcosta]

🐛 Bug Report

Currently, when cloning straight from the current version master the test for getting changed files in hg is failing.

I'm running the latest version of hg (5.0.1) in an old Mac I have but the same happens in any of the other machines I've cloned jest into.

The error I get is due to the pattern we use to specify a revset returning an empty revision range.

    Command failed: hg status -amnu --rev min((!public() & ::.)+.)^ /var/folders/p0/0npmk50s57v5sgzb_s9z2xmc0000gn/T/jest-changed-files-test-dir /var/folders/p0/0npmk50s57v5sgzb_s9z2xmc0000gn/T/jest-changed-files-test-dir/nested-dir /var/folders/p0/0npmk50s57v5sgzb_s9z2xmc0000gn/T/jest-changed-files-test-dir/nested-dir/second-nested-dir
    abort: empty revision range

      at makeError (node_modules/execa/index.js:181:11)

Related PRs:

• Improve hg revset used by jest-changed-files #7880
• Fix pattern for HG changed files #8066
• Make lastCommit and changedFilesWithAncestor work through config #5476

To Reproduce

Steps to reproduce the behavior:

1. Clone jest
2. Install hg
3. Install dependencies
4. Run tests (yarn test)

Expected behavior

The test linked above should pass since the withAncestor option should return both the current changes and the changes in the last commit. Currently, the pattern we use to specify a revset errors because the pattern specifies an empty range.

I managed to solve this locally myself, by using ancestors(.)::0 as the argument for --revset, but I'm not sure whether this solution covers all necessary edge-cases. Since I don't use mercurial myself I thought it was better to open an issue and see what other mercurial users think of this solution. If this is a satisfactory solution I can open a PR as I already have these changes commited to this branch.

Link to repl or repo (highly encouraged)

As per the steps to reproduce section, this repo is its own reproducible repo.

Run npx envinfo --preset jest

Paste the results here:

  System:
    OS: macOS High Sierra 10.13.6
    CPU: (4) x64 Intel(R) Core(TM) i5-5257U CPU @ 2.70GHz
  Binaries:
    Node: 10.16.0 - ~/.nvm/versions/node/v10.16.0/bin/node
    Yarn: 1.15.0 - /usr/local/bin/yarn
    npm: 6.9.0 - ~/.nvm/versions/node/v10.16.0/bin/npm

[KrzysztofKarol]

I have same problem.

$ git rev-parse HEAD
e3f4c65140f08a2ec81e5a8260704c1d201e33c1

  System:
    OS: macOS 10.15.1
    CPU: (8) x64 Intel(R) Core(TM) i7-4770HQ CPU @ 2.20GHz
  Binaries:
    Node: 12.13.0 - ~/.nvm/versions/node/v12.13.0/bin/node
    Yarn: 1.19.1 - /usr/local/bin/yarn
    npm: 6.12.0 - ~/.nvm/versions/node/v12.13.0/bin/npm

$ hg --version
Mercurial Distributed SCM (version 5.1.2)

 FAIL  e2e/__tests__/jestChangedFiles.test.ts (26.869s)
  ● gets changed files for hg

    abort: empty revision range

      334 |   });
      335 | 
    > 336 |   ({changedFiles: files} = await getChangedFilesForRoots(roots, {
          |                            ^
      337 |     withAncestor: true,
      338 |   }));
      339 |   // Returns files from current uncommitted state + the last commit

      at makeError (node_modules/execa/lib/error.js:58:11)
          at runMicrotasks (<anonymous>)
          at async Promise.all (index 0)
      at Object.<anonymous> (e2e/__tests__/jestChangedFiles.test.ts:336:28)

 FAIL  e2e/__tests__/onlyChanged.test.ts (93.791s)
  ● gets changed files for hg

    expect(received).toMatch(expected)

    Expected pattern: /PASS __tests__(\/|\\)file2.test.js/
    Received string:  "·
      ● Test suite failed to run·
        abort: empty revision range
    "

      299 | 
      300 |   ({stdout, stderr} = runJest(DIR, ['-o', '--changedFilesWithAncestor']));
    > 301 |   expect(stderr).toMatch(/PASS __tests__(\/|\\)file2.test.js/);
          |                  ^
      302 |   expect(stderr).toMatch(/PASS __tests__(\/|\\)file3.test.js/);
      303 | });
      304 | 

      at Object.toMatch (e2e/__tests__/onlyChanged.test.ts:301:18)

[SimenB]

Yeah, they fail for me as well. I've never used mercurial, so I have no real point of reference here. Maybe @scotthovestadt can help?

[lucasfcosta]

@SimenB I'd be happy to open a PR if the solution I've described when I opened the issue is valid, but given that I'm not an SVN expert, I'd like to confirm before doing it:

I managed to solve this locally myself, by using ancestors(.)::0 as the argument for --revset, but I'm not sure whether this solution covers all necessary edge-cases. Since I don't use mercurial myself I thought it was better to open an issue and see what other mercurial users think of this solution. If this is a satisfactory solution I can open a PR as I already have these changes commited to this branch.

[SimenB]

I have absolutely no clue, unfortunately... Maybe @quark-zju knows?

[quark-zju]

This is mostly likely related to phase configuration. In Mercurial, the common workflow is that the "master" branch has public commits, and feature branches have "draft" commits.

I suspect all commits are "drafts" and there are commits that should be public but are not public in these cases. In that case, hg phase --public master will probably fix it.

I think the revset can be changed to:

first(min(!public() & ::.)^+.^)

to work with incorrect phase setup (all commits are draft) while still being able to take advantage of phases to detect draft feature branch.

Explanation:

• min(!public() & ::.)^: This is the parent of the first commit of the feature branch, or an empty set if phases is not setup correctly.
• first(x+y): This picks x if x is not empty, or y if x is empty.

[github-actions]

This issue is stale because it has been open for 1 year with no activity. Remove stale label or comment or this will be closed in 14 days.

[github-actions]

This issue was closed because it has been stalled for 7 days with no activity. Please open a new issue if the issue is still relevant, linking to this one.

[github-actions]

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Please note this issue tracker is not a help forum. We recommend using StackOverflow or our discord channel for questions.