Skip to content

Latest commit

 

History

History
89 lines (76 loc) · 4.22 KB

git_cheatsheet.md

File metadata and controls

89 lines (76 loc) · 4.22 KB

Git Cheatsheet

  • The 4 levels of the git organization

    • Working directory (on local computer): work in text editor and/or an IDE (RStudio, Visual Studio, etc.)
    • Staging area: include/save changes to the next commit
    • Local repository: commit to project history
    • Remote repository on github.com: share code with collaborators and backup local branches

  • Inspecting a repository

    • State of working directory and staging area: git status
    • History of commits: git log --graph --full-history --oneline --decorate
    • List of branches (* indicates the active branch):
    • Local branches: git branch
    • Remote branches: git branch -r
    • List of remote connections: git remote -v

  • Finding stuff in a repository

    • Find all commits which have affected a file: git log -- *<part_of_file_name>*
    • Find the SHA of the last commit that affected a file git rev-list -n 1 HEAD -- <file_path>
    • Find all commits which have deleted files and list the deleted files: git log --diff-filter=D --summary

  • Remove commits from current (private, i.e., not published on remote repository) state of a branch (i.e., re-writing history)

    • From working directory, staged snapshot, and commit history: git reset --hard HEAD~1
    • From staged snapshot and commit history: git reset --mixed HEAD~1
    • From commit history: git reset --soft HEAD~1

  • Remove commits from current published branch (by creating a new commit, i.e., it does not re-write history): git revert HEAD~1

  • Restore file from a previous commit: git checkout <deleting_commit>~1 -- <file_path>

  • Amending commits

    • Amending the most recent commit message: (see SO for alternative scenarios)

      • Completely rewrite message from scratch: git commit --amend
      • Amend by starting from old message: git commit --amend -c HEAD

    • Tag an old commit retroactively: So that the tag's date/time corresponds to the commit's date/time by temporarily setting the tag's clock: fill in appropriate values for <branch>, <commit SHA1>, <version>, and <message>

      git checkout <branch>
git reset --hard <commit SHA1>
GIT_COMMITTER_DATE="$(git show --format=%aD  | head -1)" git tag -a <version> -m "<message>"
git push --tags
git pull

    • Moving commits from one branch to another: (see SO) example scenario:

      <SHA1> -> <SHA2> -> <SHA3> = branch <A>
                '-> <SHA4> -> <SHA5> -> <SHA6> -> <SHA7> = branch <B>
      • Apply commits <SHA6> to <SHA7> to head of branch <A> 
        git checkout <A>
git reset --hard <SHA7>
git rebase --onto <SHA3> <SHA5>
# resolve conflicts
git rebase --continue # respectively, --skip; repeat until done

  • Interruptions to coding

    'Stashing' saves uncommitted changes and resets/cleans the working directory, e.g., to switch branches, to pull into a dirty tree, to interrupt the workflow in general. Stashes are handled in the same way as commits by git commands, but they are not linked to a specific branch. Stashes are named <stash@{X}> where X is the number on the stack. For more details see here and here

    • Push a new stash onto stack: git stash (this will only stash files that are already tracked); to stash also untracked (i.e., new files): git stash --include-untracked
    • List stored stashes on stack: git stash list
    • Apply a stored stash: git stash apply will apply <stash@{0}>; apply stash with number X: git stash apply stash@{X}. Git gives merge conflict messages if a stash does not apply cleanly. Apply a stash and stage files as before: git stash apply --index
    • Remove a stash from the stack: git stash drop stash@{X}
    • Apply and remove a stash: git stash pop
    • Show what applying a stash would add/remove to : git diff <branch> stash@{X}