tarsum: name collision fix

If a tar were constructed with duplicate file names, then depending on
the order, it could result in same tarsum.

Signed-off-by: Vincent Batts <vbatts@redhat.com>

tarsum/fileinfosums.go

@@ -0,0 +1,125 @@
+package tarsum
+
+import "sort"
+
+// This info will be accessed through interface so the actual name and sum cannot be medled with
+type FileInfoSumInterface interface {
+	// File name
+	Name() string
+	// Checksum of this particular file and its headers
+	Sum() string
+	// Position of file in the tar
+	Pos() int64
+}
+
+type fileInfoSum struct {
+	name string
+	sum  string
+	pos  int64
+}
+
+func (fis fileInfoSum) Name() string {
+	return fis.name
+}
+func (fis fileInfoSum) Sum() string {
+	return fis.sum
+}
+func (fis fileInfoSum) Pos() int64 {
+	return fis.pos
+}
+
+type FileInfoSums []FileInfoSumInterface
+
+// GetFile returns the first FileInfoSumInterface with a matching name
+func (fis FileInfoSums) GetFile(name string) FileInfoSumInterface {
+	for i := range fis {
+		if fis[i].Name() == name {
+			return fis[i]
+		}
+	}
+	return nil
+}
+
+// GetAllFile returns a FileInfoSums with all matching names
+func (fis FileInfoSums) GetAllFile(name string) FileInfoSums {
+	f := FileInfoSums{}
+	for i := range fis {
+		if fis[i].Name() == name {
+			f = append(f, fis[i])
+		}
+	}
+	return f
+}
+
+func contains(s []string, e string) bool {
+	for _, a := range s {
+		if a == e {
+			return true
+		}
+	}
+	return false
+}
+
+func (fis FileInfoSums) GetDuplicatePaths() (dups FileInfoSums) {
+	seen := make(map[string]int, len(fis)) // allocate earl. no need to grow this map.
+	for i := range fis {
+		f := fis[i]
+		if _, ok := seen[f.Name()]; ok {
+			dups = append(dups, f)
+		} else {
+			seen[f.Name()] = 0
+		}
+	}
+	return dups
+}
+
+func (fis FileInfoSums) Len() int      { return len(fis) }
+func (fis FileInfoSums) Swap(i, j int) { fis[i], fis[j] = fis[j], fis[i] }
+
+func (fis FileInfoSums) SortByPos() {
+	sort.Sort(byPos{fis})
+}
+
+func (fis FileInfoSums) SortByNames() {
+	sort.Sort(byName{fis})
+}
+
+func (fis FileInfoSums) SortBySums() {
+	dups := fis.GetDuplicatePaths()
+	if len(dups) > 0 {
+		sort.Sort(bySum{fis, dups})
+	} else {
+		sort.Sort(bySum{fis, nil})
+	}
+}
+
+// byName is a sort.Sort helper for sorting by file names.
+// If names are the same, order them by their appearance in the tar archive
+type byName struct{ FileInfoSums }
+
+func (bn byName) Less(i, j int) bool {
+	if bn.FileInfoSums[i].Name() == bn.FileInfoSums[j].Name() {
+		return bn.FileInfoSums[i].Pos() < bn.FileInfoSums[j].Pos()
+	}
+	return bn.FileInfoSums[i].Name() < bn.FileInfoSums[j].Name()
+}
+
+// bySum is a sort.Sort helper for sorting by the sums of all the fileinfos in the tar archive
+type bySum struct {
+	FileInfoSums
+	dups FileInfoSums
+}
+
+func (bs bySum) Less(i, j int) bool {
+	if bs.dups != nil && bs.FileInfoSums[i].Name() == bs.FileInfoSums[j].Name() {
+		return bs.FileInfoSums[i].Pos() < bs.FileInfoSums[j].Pos()
+	}
+	return bs.FileInfoSums[i].Sum() < bs.FileInfoSums[j].Sum()
+}
+
+// byPos is a sort.Sort helper for sorting by the sums of all the fileinfos by their original order
+type byPos struct{ FileInfoSums }
+
+func (bp byPos) Less(i, j int) bool {
+	return bp.FileInfoSums[i].Pos() < bp.FileInfoSums[j].Pos()
+}