Add githook and an ant target to push it to metadata servers

See ukf-meta#244

build.xml

@@ -271,6 +271,7 @@
     <property name="mdx.dir" value="${basedir}/mdx"/>
     <property name="rules.dir" value="${mdx.dir}/_rules"/>
     <property name="utilities.dir" value="${basedir}/utilities"/>
+    <property name="githook.dir" value="${utilities.dir}/githooks"/>
 
     <!--
         Location of externally supplied tool bundles.
@@ -330,6 +331,7 @@
     -->
     <property name="mdq.cache" value="mdqcache.tar.gz"/>
     <property name="mdaggr.stats" value="ukfederation-stats.html"/>
+    <property name="post-receive-githook" value="post-receive"/>
 
 
     <!--
@@ -631,6 +633,7 @@
         git.data.merge.masterintoimmediate,
         git.data.merge.immediateintodeferred,
         git.data.allbranches.pushtoorigin,
+        scp.githook,
         publish.mdqcache,
         publish.md,
         publish.otherfiles,
@@ -2144,6 +2147,38 @@
         ***********************************************
     -->
 
+    <target name="scp.githook">
+        <!--
+            Metadata servers have a post-receive githook to fix up outputs of metadata aggregation.
+            Not needed for CDI because githook included in the container already.
+        -->
+        <echo>Pushing post-receive githook to MD dist.</echo>
+        <echo>-> MD-NE-01</echo>
+        <scp failonerror="true" remoteTodir="${md.user}@${md.dist.host-ne-01.name}:/tmp" keyfile="~/.ssh/id_rsa" knownhosts="~/.ssh/known_hosts">
+            <fileset dir="${githook.dir}">
+                <include name="${post-receive-githook}"/>
+            </fileset>
+        </scp>
+        <echo>-> MD-NE-02</echo>
+        <scp failonerror="true" remoteTodir="${md.user}@${md.dist.host-ne-02.name}:/tmp" keyfile="~/.ssh/id_rsa" knownhosts="~/.ssh/known_hosts">
+            <fileset dir="${githook.dir}">
+                <include name="${post-receive-githook}"/>
+            </fileset>
+        </scp>
+        <echo>-> MD-WE-01</echo>
+        <scp failonerror="true" remoteTodir="${md.user}@${md.dist.host-we-01.name}:/tmp" keyfile="~/.ssh/id_rsa" knownhosts="~/.ssh/known_hosts">
+            <fileset dir="${githook.dir}">
+                <include name="${post-receive-githook}"/>
+            </fileset>
+        </scp>
+        <echo>-> MD-WE-02</echo>
+        <scp failonerror="true" remoteTodir="${md.user}@${md.dist.host-we-02.name}:/tmp" keyfile="~/.ssh/id_rsa" knownhosts="~/.ssh/known_hosts">
+            <fileset dir="${githook.dir}">
+                <include name="${post-receive-githook}"/>
+            </fileset>
+        </scp>
+    </target>
+
     <target name="publish.md">
         <!--
             Push metadata files for the UK Federation to the MD dist servers

utilities/githooks/post-receive

@@ -0,0 +1,92 @@
+#!/bin/bash
+
+# Git provides the following as part of STDIN when invoking this script: <oldrev> <newrev> <refname>
+read oldrev newrev refname
+
+# Set the location of the git repo and the apache directories to serve content from
+gitdir=/var/git/ukf-products
+apacheaggrdir=/var/www/html/metadata.uou
+apachemdqdir=/var/www/html/mdq.uou/entities
+
+# Set the location of the temporary mdq cache dir
+mdqcachedir=/tmp/mdqcache
+
+# This Git repo has had the latest stuff pushed to it, but it hasn't checked it out yet. So let's do it.
+git --work-tree=$gitdir --git-dir=$gitdir/.git checkout -f
+
+# Make a gzipped version of each aggregate
+echo -n "Gzipping each aggregate... "
+for f in $gitdir/aggregates/*.xml
+do
+    gzip -9 < $f > $f.gz
+done
+echo "Done."
+
+# The MDQ cache should have been SCPed to /tmp and be sitting there happily.
+# First, we should untar it.
+echo -n "Untarring mdq cache... "
+rm -rf $mdqcachedir
+mkdir $mdqcachedir
+cd $mdqcachedir
+tar xzf /tmp/mdqcache.tar.gz
+echo "Done."
+
+# Make a gzipped version of each per-entity fragment; also create symlink
+# to the XML file and its .gz version named from the SHA1 hash of the entityId
+echo -n "Gzipping each fragment file, and symlinking to the file and the .gz... "
+cd $mdqcachedir
+for f in $mdqcachedir/*.xml
+do
+
+    # First we're going to figure out some stuff about the request and how it'll
+    # map to other versions of the name
+
+    # Convert the /full/path/and/filename.xml to just filename.xml
+    filename=${f##*/}
+
+    # And then filename.xml to just filename (i.e. the % encoded entityId)
+    entityidpercentencoded=${filename%.*}
+
+    # Un-%encode the entityId
+    entityid=$(echo $entityidpercentencoded | sed "s@+@ @g;s@%@\\\\x@g" | xargs -0 printf "%b")
+
+    # Calculate the sha1 hash of the entityId
+    entityidsha1=$(echo -n $entityid | openssl sha1 | awk '{print $2}')
+
+
+    # Now we're actually going to do something with that!
+
+    # Create the gzipped version of the file
+    gzip -9 < $filename > x_gz-$filename.gz
+
+    # Remove .xml from the filenames
+    mv -f $filename $entityidpercentencoded
+    mv -f x_gz-$filename.gz x_gz-$entityidpercentencoded.gz
+
+    # Create the symlinks to the XML file and the gzipped version
+    ln -s $entityidpercentencoded {sha1}$entityidsha1
+    ln -s x_gz-$entityidpercentencoded.gz x_gz-{sha1}$entityidsha1.gz
+
+done
+echo "Done."
+
+# Get the timestamp of the commit
+mtime=$(git --work-tree=$gitdir --git-dir=$gitdir/.git show $newrev --quiet --pretty=format:%ct)
+
+# Set the timestamp on each of the files to that of the commit
+echo -n "Setting the timestamp on each file to that of the commit... "
+find $gitdir -regextype posix-extended -regex '.*\.(xml|gz)' -exec touch -d @$mtime {} \;
+find $mdqcachedir -exec touch -d @$mtime {} \;
+echo "Done."
+
+# Put files into the correct directory
+echo -n "Rsyncing files to the appropriate apache directory... "
+rsync -at $gitdir/aggregates/*.{xml,gz} $apacheaggrdir
+rsync -at --delete $mdqcachedir/ $apachemdqdir
+echo "Done."
+
+# Remove the temporary files
+echo -n "Removing temporary files... "
+find $gitdir -name "*.gz" -exec rm -f {} \;
+rm -rf $mdqcachedir
+echo "Done."