Merge Zanata repos

mergerepos.sh

#!/bin/zsh -ex

# https://zanata.atlassian.net/browse/ZNTA-1391

# FIXME handle extra branches on client and api

# determine directory containing this script
SCRIPT_DIR=$( dirname "$0:A" )

oldParentVersion=30-SNAPSHOT
newVersion=4.0.0-SNAPSHOT
workingDir=$PWD/zanata-megamerge

mkdir $workingDir; cd $workingDir
# FIXME depth
# --depth 1
git clone git@github.com:zanata/zanata-server.git server
git clone --mirror git@github.com:zanata/zanata-client.git client
git clone --mirror git@github.com:zanata/zanata-common.git common
git clone --mirror git@github.com:zanata/zanata-api.git api
git clone --mirror git@github.com:zanata/zanata-parent.git parent

setopt extendedglob # requires zsh
others=($(echo ^server))

function gitCloneBranches() {
  # http://stackoverflow.com/a/16563327/14379
  git branch --all | sed -n "/\/HEAD /d; /\/master$/d; /remotes/p;" | xargs -L1 git checkout --force --track
  git checkout master
}

function deletePRs() {
  git for-each-ref --format="%(refname)" refs/pull/ refs/reviewable/ | xargs --no-run-if-empty -n 1 git update-ref -d
}

cd server
deletePRs
gitCloneBranches
mkdir server
git rm .travis.yml
# NB we don't want to git mv .gitignore .gitattributes .travis-settings.xml
git mv ^server server/
git commit -nm "chore(ZNTA-1391) move server to subdir; disable Travis"

# fetch+merge all branches+tags from the other repos
for d in $others; do
  echo $d:
  git remote add --fetch --tags $d ../$d

  git checkout $d/master
  git checkout -b ZNTA-1391-${d}
  mkdir $d
  git rm .travis*

  # the glob pattern should pick up everything except $d and .git
  # (#qD) extended glob qualifier tells zsh to include dot files in the glob
  # NB the standard glob qualifier version ^(api|.git)(.D) negates zanata-common-api
  git mv ^($d|.git)(#qD) $d/
  git commit -nm "chore(ZNTA-1391) move $d to subdir; disable Travis"

  git checkout master
  # GIT_EDITOR=true: use default commit message
  GIT_EDITOR=true git merge ZNTA-1391-${d}
  git branch -d ZNTA-1391-${d}
  echo
done


# grab some other branches we want to keep
# from api:
git checkout api/ZNTA-1302-legacy-via-file && git checkout -b api-ZNTA-1302-legacy-via-file
git checkout ZNTA-1333-language-react || true
git checkout restore-version-4.0.0-SNAPSHOT || true
git checkout streaming-parser || true
# from client:
git checkout client/ZNTA-1302-legacy-via-file && git checkout -b client-ZNTA-1302-legacy-via-file
git checkout java-nio-file || true
git checkout maven-plugin-annotations || true
git checkout rawToFile || true
git checkout rhbz870876 || true
git checkout zanata-console || true
#from parent:
git checkout fix-warnings || true

git checkout master

# copy root pom.xml to server dir (which will become the new root dir):

cat <<EOF > pom.xml
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
  <modelVersion>4.0.0</modelVersion>
  <groupId>org.zanata</groupId>
  <artifactId>zanata-platform</artifactId>
  <name>Zanata Platform</name>
  <url>http://zanata.org/</url>
  <packaging>pom</packaging>
  <parent>
    <groupId>org.zanata</groupId>
    <artifactId>parent</artifactId>
    <version>${newVersion}</version>
    <relativePath>parent</relativePath>
  </parent>
  <modules>
    <module>parent</module>
    <module>api</module>
    <module>common</module>
    <module>client</module>
    <module>server</module>
  </modules>
</project>
EOF

# remove explicit version from aggregator poms
sed -i "/<version>${newVersion}<\/version>/d" */pom.xml

# rename zanata-parent to parent and point to the new parent pom
sed -i "s|<artifactId>zanata-parent</artifactId>|<artifactId>parent</artifactId>|" */pom.xml
sed -i "s|<relativePath>../zanata-parent</relativePath>|<relativePath>../parent</relativePath>|" */pom.xml

# update to match new parent's version instead of old
sed -i "s|<version>${oldParentVersion}</version>|<version>${newVersion}</version>|" */pom.xml

# check that Maven is happy
mvn -q validate

git add pom.xml */pom.xml
git commit -m "chore(ZNTA-1391) link poms to new parent pom, create root pom"

# make backup before BFG run
git clone . ../server-pre-clean
git-fix-emails
$SCRIPT_DIR/repoclean.sh

# tell zsh to stop logging every command
unsetopt -x

echo "don't forget to update the paths in .gitignore to match the new subdirectories"

echo "when you're ready:"
# need to push all refs, not just master
echo git push --all git@github.com:zanata/zanata-platform.git
echo git push --tags git@github.com:zanata/zanata-platform.git

echo "WARNING: don't push the shallow clone!"

echo the merged repo can be found here:
echo cd $workingDir/server

repoclean.sh

#!/bin/bash -e

echo Cleaning up repositories with BFG

echo Before expire/gc:
du -hs .git/

echo Running expire/gc:
git reflog expire --expire=now --all && git gc --prune=now --aggressive

echo Before cleanup:
du -hs .git/
#git-find-largest

echo Cleaning up with the BFG Repo-Cleaner:
bfg --strip-blobs-bigger-than 1M --delete-folders gwt-unitCache --delete-files '{bundle*.css,bundle.min.js,bundle.js,libs.css,libs*.js,templates*.js,*.js.map,*.pdf,*.swf,*.war}'

echo Running expire/gc:
git reflog expire --expire=now --all && git gc --prune=now --aggressive

echo After cleanup:
du -hs .git/
#git-find-largest

# remove these files if possible:
# All sizes are in kB's. The pack column is the size of the object, compressed, inside the pack file.
# size  pack  SHA                                       location
# 2872  703   62b29fc833d9c5c7423ce3b9b87cf83e3c83063e  zanata-adapter-glossary/src/test/resources/glossary/compendium-zh_TW.po
# 1828  471   d5ad8fd81f17c5b56dceaa9dd20f574bb6d6c5e5  zanata-war/src/main/webapp/app/bundle.js
# 1414  264   c1b1c8fcc6040bd4e3639c04967cf86c5f67c9b3  zanata-war/src/test/resources/tmx/fedora-readme-burning-isos.tmx
# 1318  179   d6a52093f3baed9d04be00ca0b8fb258c44832ad  zanata-adapter-xliff/src/test/resources/StringResource_de.xml
# 428   152   46ea179a7dca791987aa40a434e720daaeef0234  zanata-war/src/main/webapp/app/js/libs.js
# 338   31    00bd1183a902602675c1ca9b7805e3d583b9ad44  client/zanata-client-commands/src/test/resources/xliffDir2/StringResource_en-GB.xml
# 334   91    749900b6d88fdc2a10eb1b9c0a153d4e38634f06  zanata-war/src/main/webapp/profile/js/bundle.min.js
# 332   95    16285a69ef38a8ba6c4e6502a29215562cb94da0  zanata-war/src/main/webapp/app/js/app.js

Note that git-fix-emails is not in this gist, because it has hard-coded email addresses in it.

git-fix-emails used git fast-export --all piped to a sed script, piped to git fast-import --force to repair a couple of damaged author/committer headers we had in the repo, plus git filter-branch --force --env-filter '{embedded script}' --tag-name-filter cat -- --branches --tags to update some other email addresses.