如何将文件流式传输至远程Git仓库？无需本地仓库存储

Great question—this is a super common use case for automation workflows or environments where you don’t want to clutter your local filesystem with temporary files. While Git is fundamentally built around local repositories, there are two reliable ways to get content to a remote repo without maintaining a local working copy:

1. Use Git's low-level commands to create objects and push directly #

Git’s core relies on content-addressable objects, so you can skip the working directory and staging area entirely by manipulating these objects directly. Here’s a step-by-step breakdown:

• First, generate a blob object (Git’s way of storing file content) from your streamed input. This creates the object in your local Git object database (you don’t need a full repo—just Git installed):

  # Replace the echo command with your actual stream (e.g., cat /dev/stdin, curl output, etc.)
  CONTENT_STREAM=$(echo "This is my streamed file content")
  BLOB_HASH=$(echo "$CONTENT_STREAM" | git hash-object -w --stdin)
  

• Next, create a tree object to link the blob to a specific file path:

  TREE_HASH=$(echo "100644 blob $BLOB_HASH    my-streamed-file.txt" | git mktree)
  

• Then, create a commit object pointing to this tree, using the remote branch’s latest commit as the parent:

  # Fetch the latest commit hash of the remote branch (e.g., main)
  REMOTE_LATEST=$(git ls-remote origin main | cut -f1)
  COMMIT_HASH=$(echo "Add streamed file via Git low-level commands" | git commit-tree $TREE_HASH -p $REMOTE_LATEST)
  

• Finally, push the new commit to the remote branch:

  git push origin $COMMIT_HASH:main
  

Pros: Works with any Git remote (not just hosted platforms), no dependency on external APIs.
Cons: Requires basic understanding of Git’s object model, more manual steps, and uses Git’s local object database temporarily.

2. Use your Git host's API to upload content directly #

Most major Git hosting platforms (GitHub, GitLab, Bitbucket) offer REST APIs that let you create or update files in a repo without any local Git setup. This is the simplest approach for most use cases.

For example, using GitHub’s API to upload a streamed file:

# Replace placeholders with your details
OWNER="your-username"
REPO="your-repo"
FILE_PATH="my-streamed-file.txt"
MESSAGE="Add streamed file via GitHub API"
CONTENT=$(echo "This is my streamed content" | base64) # API expects base64-encoded content

curl -X PUT \
  https://api.github.com/repos/$OWNER/$REPO/contents/$FILE_PATH \
  -H "Authorization: token YOUR_GITHUB_PAT" \
  -d '{
    "message": "'"$MESSAGE"'",
    "content": "'"$CONTENT"'"
  }'

Pros: No local Git required, easy to automate, handles commit creation and branching logic for you.
Cons: Platform-specific (you’ll need to adjust syntax for GitLab/Bitbucket APIs), requires an authentication token.

Quick Note #

Neither method is true "streaming" in the sense of sending data incrementally with zero local storage, but both avoid storing the file in a local working directory or full repo. If you’re in an environment without Git installed, the API approach is your best bet.

内容的提问来源于stack exchange，提问作者Alexander Mills