From 19f75cc16e5d01c827a081d46e4525a1cf516d75 Mon Sep 17 00:00:00 2001
From: "Kyle J. McKay" <mackyle@gmail.com>
Date: Sun, 28 Aug 2016 18:13:51 -0700
Subject: [PATCH] mirrors: implement refs namespace cleaning

Some hosting providers like to include all kinds of "extra" refs in the
refs namespace.  Things the project contributors never pushed and
generally are not allowed to push to.

Therefore provide a refs namespace "cleaning" service for mirrors which
is selected by default for new mirror project registrations and when
enabled causes the mirror to only mirror refs/heads/*, refs/tag/* and
refs/notes/* refs thereby "cleaning" the mirror.

Signed-off-by: Kyle J. McKay <mackyle@gmail.com>
---
 jobd/update.sh | 27 +++++++++++++++++++++++++++
 taskd/clone.sh | 10 +++++++++-
 2 files changed, 36 insertions(+), 1 deletion(-)

diff --git a/jobd/update.sh b/jobd/update.sh
index 41db367..2704c8c 100755
--- a/jobd/update.sh
+++ b/jobd/update.sh
@@ -271,6 +271,21 @@ case "$url" in
 		# Note the git config documentation is wrong
 		# transfer.unpackLimit, if set, overrides fetch.unpackLimit
 		git_add_config 'transfer.unpackLimit=1'
+		if ! is_gfi_mirror_url "$url"; then
+			lastwasclean=
+			[ "$(git config --bool girocco.lastupdateclean 2>/dev/null || :)" != "true" ] || lastwasclean=1
+			nextisclean=
+			[ "$(git config --bool girocco.cleanmirror 2>/dev/null || :)" != "true" ] || nextisclean=1
+			if [ "$nextisclean" != "$lastwasclean" ]; then
+				if [ -n "$nextisclean" ]; then
+					git config --replace-all remote.origin.fetch "+refs/heads/*:refs/heads/*"
+					git config --add remote.origin.fetch "+refs/tags/*:refs/tags/*"
+					git config --add remote.origin.fetch "+refs/notes/*:refs/notes/*"
+				else
+					git config --replace-all remote.origin.fetch "+refs/*:refs/*"
+				fi
+			fi
+		fi
 		# remember the starting time so we can easily detect new packs for fast-import mirrors
 		# we sleep for 1 second after creating .gfipack to make sure all packs are newer
 		if is_gfi_mirror_url "$url" && [ ! -e .gfipack ]; then
@@ -279,6 +294,18 @@ case "$url" in
 			sleep 1
 		fi
 		GIT_SSL_NO_VERIFY=1 bang git remote update $pruneopt
+		if ! is_gfi_mirror_url "$url" && [ "$nextisclean" != "$lastwasclean" ]; then
+			if [ -n "$nextisclean" ]; then
+				# We must manually purge the unclean refs now as even prune won't do it
+				git for-each-ref --format='%(refname)' |
+				while read -r aref; do
+					case "$aref" in refs/heads/*|refs/tags/*|refs/notes/*) :;; *)
+						git update-ref -d "$aref"
+					esac
+				done
+			fi
+			git config --bool girocco.lastupdateclean ${nextisclean:-0}
+		fi
 		if [ -e .gfipack ] && is_gfi_mirror_url "$url"; then
 			find objects/pack -type f -newer .gfipack -name "pack-$octet20.pack" -print >>gfi-packs
 			rm -f .gfipack
diff --git a/taskd/clone.sh b/taskd/clone.sh
index 2980320..f5309b7 100755
--- a/taskd/clone.sh
+++ b/taskd/clone.sh
@@ -315,7 +315,15 @@ case "$url" in
 		# not available until Git 1.7.5 and this way we guarantee we
 		# always get exactly the intended configuration and nothing else.
 		git config remote.origin.url "$url"
-		git config remote.origin.fetch "+refs/*:refs/*"
+		if ! is_gfi_mirror_url "$url" && [ "$(git config --bool girocco.cleanmirror 2>/dev/null || :)" = "true" ]; then
+			git config --replace-all remote.origin.fetch "+refs/heads/*:refs/heads/*"
+			git config --add remote.origin.fetch "+refs/tags/*:refs/tags/*"
+			git config --add remote.origin.fetch "+refs/notes/*:refs/notes/*"
+			git config --bool girocco.lastupdateclean true
+		else
+			git config --replace-all remote.origin.fetch "+refs/*:refs/*"
+			git config --bool girocco.lastupdateclean false
+		fi
 		# Set the correct HEAD symref by using ls-remote first
 		GIT_SSL_NO_VERIFY=1 GIT_TRACE_PACKET=1 git ls-remote origin >.refs-temp 2>.pkts-temp || \
 		{
-- 
2.11.4.GIT