add: symlinks to `.dvc/cache/ha/sh` but creates `.dvc/cache/ha/sh.csv`

Description

When adding files dvc add ny.csv, the files are correctly moved to the cache as.dvc/cache/ae/78daccbcbb27cd66956368c36bdd8f.csv. However, the created symlink points to .dvc/cache/ae/78daccbcbb27cd66956368c36bdd8f (missing the extension). This results in a dead link.

Reproduce

1. Enable symlinks dvc config cache.type symlink
2. Create a dvc repo within a Google Drive Folder on a Mac M1 (Idk if this is critical) dvc init
3. touch foo.csv && dvc add foo.csv
4. Confirm that cache contains a file with a hash but points to a file without it

Expected

I expect the created symlink to not be dead on creation.

Environment information

DVC version: 2.10.2 (brew)
---------------------------------
Platform: Python 3.9.12 on macOS-12.3.1-arm64-arm-64bit
Supports:
        azure (adlfs = 2022.4.0, knack = 0.9.0, azure-identity = 1.10.0),
        gdrive (pydrive2 = 1.10.1),
        gs (gcsfs = 2022.3.0),
        webhdfs (fsspec = 2022.3.0),
        http (aiohttp = 3.8.1, aiohttp-retry = 2.4.6),
        https (aiohttp = 3.8.1, aiohttp-retry = 2.4.6),
        s3 (s3fs = 2022.3.0, boto3 = 1.21.21),
        ssh (sshfs = 2022.3.1),
        oss (ossfs = 2021.8.0),
        webdav (webdav4 = 0.9.7),
        webdavs (webdav4 = 0.9.7)

Additional Information (if any):

I have been unable to reproduce this in a test example. I suspect the issue is with Google Drive more than dvc, but I figured I’d drop a line.

Thanks for a fantastic tool!

[tirkarthi]

I am unable to reproduce this locally on Linux using local filesystem. Docs in creating files through API indicate that google drive might add an extension based on file’s mime-type.

https://developers.google.com/drive/api/v3/reference/files/create

Subsequent GET requests include the read-only fileExtension property populated with the extension originally specified in the name property. When a Google Drive user requests to download a file, or when the file is downloaded through the sync client, Drive builds a full filename (with extension) based on the name. In cases where the extension is missing, Google Drive attempts to determine the extension based on the file’s MIME type.

See also : https://support.google.com/drive/thread/61927468/backup-snyc-adds-unwanted-file-extensions?hl=en

[awadell1]

Gotcha, thank you!