From 52c79a819e41e3655327eab5d90aa3867734cc38 Mon Sep 17 00:00:00 2001
From: Jonathan Boulle <jonathanboulle@gmail.com>
Date: Fri, 5 Dec 2014 17:40:46 -0800
Subject: [PATCH 1/2] pkg/tarheader: support hard links when building tar
 headers

This introduces an inode "cache" that can be used when populating a tar
header (pkg/tarheader.Populate), so that any inodes already seen can be
linked to instead of inserting an entire copy of the file into the tar.
---
 app-container/actool/build.go |  9 ++++++++-
 pkg/tarheader/pop_darwin.go   |  2 +-
 pkg/tarheader/pop_linux.go    |  2 +-
 pkg/tarheader/pop_posix.go    | 10 +++++++++-
 pkg/tarheader/tarheader.go    |  6 +++---
 5 files changed, 22 insertions(+), 7 deletions(-)

diff --git a/app-container/actool/build.go b/app-container/actool/build.go
index a1730cc..5600639 100644
--- a/app-container/actool/build.go
+++ b/app-container/actool/build.go
@@ -37,6 +37,8 @@ func init() {
 }
 
 func buildWalker(root string, aw aci.ArchiveWriter, rootfs bool) filepath.WalkFunc {
+	// cache of inode -> filepath, used to leverage hard links in the archive
+	inos := map[uint64]string{}
 	return func(path string, info os.FileInfo, err error) error {
 		if err != nil {
 			return err
@@ -81,7 +83,12 @@ func buildWalker(root string, aw aci.ArchiveWriter, rootfs bool) filepath.WalkFu
 		// modify the Name field of the returned header to provide the
 		// full path name of the file.
 		hdr.Name = relpath
-		tarheader.Populate(hdr, info)
+		tarheader.Populate(hdr, info, inos)
+		// If the file is a hard link we don't need the contents
+		if hdr.Typeflag == tar.TypeLink {
+			hdr.Size = 0
+			file = nil
+		}
 		aw.AddFile(relpath, hdr, file)
 
 		return nil
diff --git a/pkg/tarheader/pop_darwin.go b/pkg/tarheader/pop_darwin.go
index 99144db..dd0176a 100644
--- a/pkg/tarheader/pop_darwin.go
+++ b/pkg/tarheader/pop_darwin.go
@@ -13,7 +13,7 @@ func init() {
 	populateHeaderStat = append(populateHeaderStat, populateHeaderCtime)
 }
 
-func populateHeaderCtime(h *tar.Header, fi os.FileInfo) {
+func populateHeaderCtime(h *tar.Header, fi os.FileInfo, _ map[uint64]string) {
 	st, ok := fi.Sys().(*syscall.Stat_t)
 	if !ok {
 		return
diff --git a/pkg/tarheader/pop_linux.go b/pkg/tarheader/pop_linux.go
index aec5c73..16c0c7e 100644
--- a/pkg/tarheader/pop_linux.go
+++ b/pkg/tarheader/pop_linux.go
@@ -11,7 +11,7 @@ func init() {
 	populateHeaderStat = append(populateHeaderStat, populateHeaderCtime)
 }
 
-func populateHeaderCtime(h *tar.Header, fi os.FileInfo) {
+func populateHeaderCtime(h *tar.Header, fi os.FileInfo, _ map[uint64]string) {
 	st, ok := fi.Sys().(*syscall.Stat_t)
 	if !ok {
 		return
diff --git a/pkg/tarheader/pop_posix.go b/pkg/tarheader/pop_posix.go
index 99dcc42..aadefa5 100644
--- a/pkg/tarheader/pop_posix.go
+++ b/pkg/tarheader/pop_posix.go
@@ -10,11 +10,19 @@ func init() {
 	populateHeaderStat = append(populateHeaderStat, populateHeaderUnix)
 }
 
-func populateHeaderUnix(h *tar.Header, fi os.FileInfo) {
+func populateHeaderUnix(h *tar.Header, fi os.FileInfo, seen map[uint64]string) {
 	st, ok := fi.Sys().(*syscall.Stat_t)
 	if !ok {
 		return
 	}
 	h.Uid = int(st.Uid)
 	h.Gid = int(st.Gid)
+	// If we have already seen this inode, generate a hardlink
+	p, ok := seen[st.Ino]
+	if ok {
+		h.Linkname = p
+		h.Typeflag = tar.TypeLink
+	} else {
+		seen[st.Ino] = h.Name
+	}
 }
diff --git a/pkg/tarheader/tarheader.go b/pkg/tarheader/tarheader.go
index f570ac0..d74ec90 100644
--- a/pkg/tarheader/tarheader.go
+++ b/pkg/tarheader/tarheader.go
@@ -5,10 +5,10 @@ import (
 	"os"
 )
 
-var populateHeaderStat []func(h *tar.Header, fi os.FileInfo)
+var populateHeaderStat []func(h *tar.Header, fi os.FileInfo, seen map[uint64]string)
 
-func Populate(h *tar.Header, fi os.FileInfo) {
+func Populate(h *tar.Header, fi os.FileInfo, seen map[uint64]string) {
 	for _, pop := range populateHeaderStat {
-		pop(h, fi)
+		pop(h, fi, seen)
 	}
 }

From 69e3f6a4fb57b58cc3d94882bcb3b37274d7c9e9 Mon Sep 17 00:00:00 2001
From: Maciej Pasternacki <maciej@3ofcoins.net>
Date: Sat, 6 Dec 2014 19:11:08 +0000
Subject: [PATCH 2/2] pkg/tarheader: cast inode up to uint64

Some systems (FreeBSD included) have uint32 as syscall.Stat_t.Ino
---
 pkg/tarheader/pop_posix.go | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/pkg/tarheader/pop_posix.go b/pkg/tarheader/pop_posix.go
index aadefa5..b454d7f 100644
--- a/pkg/tarheader/pop_posix.go
+++ b/pkg/tarheader/pop_posix.go
@@ -18,11 +18,11 @@ func populateHeaderUnix(h *tar.Header, fi os.FileInfo, seen map[uint64]string) {
 	h.Uid = int(st.Uid)
 	h.Gid = int(st.Gid)
 	// If we have already seen this inode, generate a hardlink
-	p, ok := seen[st.Ino]
+	p, ok := seen[uint64(st.Ino)]
 	if ok {
 		h.Linkname = p
 		h.Typeflag = tar.TypeLink
 	} else {
-		seen[st.Ino] = h.Name
+		seen[uint64(st.Ino)] = h.Name
 	}
 }