git@vger.kernel.org mailing list mirror (one of many)
 help / color / mirror / code / Atom feed
From: Johannes Schindelin <Johannes.Schindelin@gmx.de>
To: Josh Smeaton <josh.smeaton@gmail.com>
Cc: git@vger.kernel.org
Subject: Re: Bug: diff-index reports unrelated files as deleted when fsmonitor is enabled
Date: Mon, 27 Mar 2023 09:49:53 +0200 (CEST)	[thread overview]
Message-ID: <8afb1988-df49-d759-ead8-1bd384f09c82@gmx.de> (raw)
In-Reply-To: <CAPbDM0d5Xe=KTLECDCJYtpzmBt2wqzZrMWVcbiisM+abopenig@mail.gmail.com>

Hi Josh,

On Sat, 4 Mar 2023, Josh Smeaton wrote:

> git write-tree and then git diff-index reports thousands of files as
> `deleted` when core.fsmonitor is true, even though the files are still
> present on the file system.

I suspect that the issue is a missing `git update-index --refresh`. While
`git diff` refreshes the index automatically, `git diff-index` does not
(when you use low-level commands like `diff-index` you need to do such
things explicitly).

If that does not fix your problem, I encourage you to distill the issue
into a simple MCVE (https://stackoverflow.com/help/mcve), preferably in
the form of a patch, say, to t7527 (see
https://github.com/git/git/blob/v2.40.0/t/t7527-builtin-fsmonitor.sh).

Ciao,
Johannes

>
> What did you do before the bug happened? (Steps to reproduce your issue)
>
> ```
> git config core.fsmonitor true
> touch new-file.txt
> git add new-file.txt
> git diff-index --ignore-submodules --exit-code --no-color
> --no-ext-diff $(git write-tree) --
> ```
>
> What did you expect to happen? (Expected behavior)
>
> I expected that no files would be reported as having changes.
>
> What happened instead? (Actual behavior)
>
> Thousands of unrelated files were reported as `deleted`.
>
> ```
> git diff-index --ignore-submodules --exit-code --no-color
> --no-ext-diff $(git write-tree) --
> :100644 000000 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
> 0000000000000000000000000000000000000000 D src/tests/__init__.py
> :100644 000000 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
> 0000000000000000000000000000000000000000 D src/tests/site/__init__.py
> :100644 000000 d8edaad92120781998e811277f966551b12fc823
> 0000000000000000000000000000000000000000 D
> src/tests/site/test_models.py
>
>  ... thousands of other files
> ```
>
> What's different between what you expected and what actually happened?
>
> Thousands of unrelated files were reported as having been deleted.
>
>
> Anything else you want to add:
>
> Myself and a few co-workers all ran into this issue in the last few days. It
> presented itself when pre-commit hooks were run (specifically with the
> pre-commit
> tool from github pre-commit).
>
> When committing changes, the pre-commit tool runs the equivalent of:
>
> `git diff-index --ignore-submodules --exit-code --no-color
> --no-ext-diff $(git write-tree) --`
>
> It then stashes and unstashes those changes, causing the files to be
> removed from
> the working tree.
>
>
> ```
> $ git config core.fsmonitor true
> $ git checkout -b test-branch
> Switched to a new branch 'test-branch'
>
> $ echo "test 3" >> newfile.txt
> $ git add .
> $ git status
> On branch test-branch
> Changes to be committed:
>   (use "git restore --staged <file>..." to unstage)
> new file:   newfile.txt
>
> $ git commit -a -m "will it fail?"
> [WARNING] Unstaged files detected.
> [INFO] Stashing unstaged files to
> /Users/josh/.cache/pre-commit/patch1677888420-10129.
> trim trailing whitespace.................................................Passed
> check python ast.....................................(no files to check)Skipped
> check for case conflicts.................................................Passed
> [INFO] Restored changes from
> /Users/josh/.cache/pre-commit/patch1677888420-10129.
> [test-branch 7c899582c19] will it fail?
>  1 file changed, 1 insertion(+)
>  create mode 100644 newfile.txt
>
> $ git status
> On branch test-branch
> Changes not staged for commit:
>   (use "git add/rm <file>..." to update what will be committed)
>   (use "git restore <file>..." to discard changes in working directory)
> deleted:    src/tests/__init__.py
> deleted:    src/tests/site/__init__.py
> deleted:    src/tests/site/test_models.py
>         ... lots more files ...
>
> no changes added to commit (use "git add" and/or "git commit -a")
>
> $ git status | wc -l
>    15298
>
> $ git restore .
>
> $ git status
>
> On branch test-branch
> nothing to commit, working tree clean
> ```
>
> It can be reproduced, in our repo, without using pre-commit at all:
>
> ```
>
> git diff-index --ignore-submodules --exit-code --no-color
> --no-ext-diff $(git write-tree) --
> :100644 000000 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
> 0000000000000000000000000000000000000000 D src/tests/__init__.py
> :100644 000000 e69de29bb2d1d6434b8b29ae775ad8c2e48c5391
> 0000000000000000000000000000000000000000 D src/tests/site/__init__.py
> :100644 000000 d8edaad92120781998e811277f966551b12fc823
> 0000000000000000000000000000000000000000 D
> src/tests/site/test_models.py
>
>  ... thousands of other files
> ```
>
> Using `update-index` with the `--no-fsmonitor` flag prior to diff-index does the
> correct thing:
>
> ```
> $ git update-index --really-refresh --no-fsmonitor
> warning: core.fsmonitor is set; remove it if you really want to
> disable fsmonitor
>
> $ git diff-index --ignore-submodules --exit-code --no-color
> --no-ext-diff $(git write-tree) --
> > no output
> ```
>
> The problem persists when running `update-index` without disabling fsmonitor.
>
> I have reported this issue to the pre-commit project, but it looks like an issue
> with fsmonitor + write-tree.
>
> See: pre-commit issues on github #2795
>
>
>
> [System Info]
> git version:
> git version 2.39.2 (others reproduced with 2.39.1)
> cpu: arm64
> no commit associated with this build
> sizeof-long: 8
> sizeof-size_t: 8
> shell-path: /bin/sh
> feature: fsmonitor--daemon
> uname: Darwin 22.3.0 Darwin Kernel Version 22.3.0: Mon Jan 30 20:38:37
> PST 2023; root:xnu-8792.81.3~2/RELEASE_ARM64_T6000 arm64
> compiler info: clang: 14.0.0 (clang-1400.0.29.202)
> libc info: no libc information available
> $SHELL (typically, interactive shell): /bin/zsh
>
>
> [Enabled Hooks]
> pre-commit
>

      reply	other threads:[~2023-03-27  7:50 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-03-04  1:55 Bug: diff-index reports unrelated files as deleted when fsmonitor is enabled Josh Smeaton
2023-03-27  7:49 ` Johannes Schindelin [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=8afb1988-df49-d759-ead8-1bd384f09c82@gmx.de \
    --to=johannes.schindelin@gmx.de \
    --cc=git@vger.kernel.org \
    --cc=josh.smeaton@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://80x24.org/mirrors/git.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).