git-filter-branch examples and notes

git-commit-cleaner.py

"""

Creates pretty-looking commit messages for git. This was created to pretty
print the old-commit-id values for filter-branched-rewritten commits in
pyphantomjs from the phantomjs repository.

"""

import os
import sys

commit_id = os.environ["GIT_COMMIT"]
message   = sys.stdin.read()

if len(message) > 0:
    if message[-2] != "\n" and not "Last changes" in message:
        message += "\n"

    message += "original-commit-id: " + str(commit_id)

    message = message.replace("Last changes", "")

    if not "\n\noriginal-commit-id: " in message:
        message = message.replace("\noriginal-commit", "\n\noriginal-commit")

print message

git-filter-branch.sh

git filter-branch --index-filter "git ls-files | grep -v pilot/tag_extractor.py | xargs git rm --cached" --prune-empty
git filter-branch -f --tree-filter "if [[ -e pilot/tag_extractor.py ]]; then mv pilot/tag_extractor.py tag_extractor.py; fi;"


git filter-branch --commit-filter '
        if [ "$GIT_COMMITTER_NAME" = "--Nathan McCorkle" ];
        then
                GIT_COMMITTER_NAME="Nathan McCorkle";
                GIT_AUTHOR_NAME="Nathan McCorkle";
                GIT_COMMITTER_EMAIL="nmz787@gmail.com";
                GIT_AUTHOR_EMAIL="nmz787@gmail.com";
                git commit-tree "$@";
        else
                git commit-tree "$@";
        fi' HEAD

git filter-branch -f --index-filter "git rm -r -f --cached --ignore-unmatch $(ls -xd trainers/ text/ stats/ gfx/ audio/ battle/)" --prune-empty -- --all

git filter-branch -f --tree-filter "git rm -rf -f --cached --ignore-unmatch README.md compare.sh constants.asm main.asm INSTALL.md Makefile main.asm pokecrystal.asm wram.asm vblank.asm bittable2.asm" --prune-empty -- --all

git filter-branch --tag-name-filter cat --prune-empty --msg-filter 'python ~/scripts/commit-cleaner.py' -- --all

git filter-branch --index-filter "git update-index --remove README.md compare.sh constants.asm main.asm INSTALL.md Makefile main.asm pokecrystal.asm wram.asm vblank.asm bittable2.asm trainers/ text/ stats/ gfx/ audio/ battle/ music/" --

git filter-branch --index-filter "git rm -rf --cached --ignore-unmatch README.md compare.sh constants.asm main.asm INSTALL.md Makefile main.asm pokecrystal.asm wram.asm vblank.asm bittable2.asm trainers/ text/ stats/ gfx/ audio/ battle/ music/ README .hgignore textpre.awk data/" HEAD

git filter-branch --prune-empty

git filter-branch -f --commit-filter "git rm -r -f --cached --ignore-unmatch README.md compare.sh constants.asm main.asm INSTALL.md Makefile main.asm pokecrystal.asm wram.asm vblank.asm bittable2.asm trainers/ text/ stats/ gfx/ audio/ battle" --prune-empty -- --all


# move files out of a subdirectory
git filter-branch --tree-filter 'test -d extras/ && mv extras/* . || echo "Nope"' HEAD
git filter-branch --tree-filter 'test -d extras/ && mv extras crystal && (test -d preprocessor.py && mv preprocessor.py crystal/ || echo "hi") || echo "Nope"' HEAD

# prune empty merges ??
# http://git.661346.n2.nabble.com/Removing-useless-merge-commit-with-quot-filter-branch-quot-td7356544.html
git filter-branch -f --prune-empty --parent-filter ~/scripts/git-rewrite-parent.rb master

git filter-branch --commit-filter '
        if [ "$GIT_COMMITTER_NAME" = "padz" ];
        then
                GIT_COMMITTER_NAME="yenatch";
                GIT_AUTHOR_NAME="yenatch";
                GIT_COMMITTER_EMAIL="yenatch@gmail.com";
                GIT_AUTHOR_EMAIL="yenatch@gmail.com";
                git commit-tree "$@";
        else
                git commit-tree "$@";
        fi' HEAD

yo, use --subdirectory-filter. Also, can I have commit-cleaner.py pretty please?

Here's a good way to clean the repo up after filter-branch:

git remote rm origin
rm -rf .git/refs/original/ .git/refs/remotes/ .git/*_HEAD .git/logs/
git for-each-ref --format="%(refname)" refs/original/ | xargs -n1 --no-run-if-empty git update-ref -d

And then:

git -c gc.reflogExpire=0 -c gc.reflogExpireUnreachable=0 -c gc.rerereresolved=0 \
-c gc.rerereunresolved=0 -c gc.pruneExpire=now gc "$@" --aggressive

Is there some way to use filter--branch and replace a using a pattern. For example say I want to replace all issue number with just the word task(i.e. issue-\d{1,4} -> Task)

The real MVP! Thanks man!

Tip: you can write a complex shell script into a file rather than between a multiline '':

git filter-branch --commit-filter 'source $(PWD)/../../.git/commit-filter-foo.sh' <first-commit-hash>^..<last-commit-hash>

Note: PWD is <repo>/.git-rewrite/t/.

commit-filter-foo.sh in .git of the <repo> contains for example if [ "$GIT_COMMITTER_NAME" = ... from above.