1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
|
package git
import (
"fmt"
"github.com/go-git/go-git/v5/plumbing"
"github.com/go-git/go-git/v5/plumbing/filemode"
"github.com/go-git/go-git/v5/plumbing/object"
"github.com/go-git/go-git/v5/storage"
)
type objectWalker struct {
Storer storage.Storer
// seen is the set of objects seen in the repo.
// seen map can become huge if walking over large
// repos. Thus using struct{} as the value type.
seen map[plumbing.Hash]struct{}
}
func newObjectWalker(s storage.Storer) *objectWalker {
return &objectWalker{s, map[plumbing.Hash]struct{}{}}
}
// walkAllRefs walks all (hash) references from the repo.
func (p *objectWalker) walkAllRefs() error {
// Walk over all the references in the repo.
it, err := p.Storer.IterReferences()
if err != nil {
return err
}
defer it.Close()
err = it.ForEach(func(ref *plumbing.Reference) error {
// Exit this iteration early for non-hash references.
if ref.Type() != plumbing.HashReference {
return nil
}
return p.walkObjectTree(ref.Hash())
})
return err
}
func (p *objectWalker) isSeen(hash plumbing.Hash) bool {
_, seen := p.seen[hash]
return seen
}
func (p *objectWalker) add(hash plumbing.Hash) {
p.seen[hash] = struct{}{}
}
// walkObjectTree walks over all objects and remembers references
// to them in the objectWalker. This is used instead of the revlist
// walks because memory usage is tight with huge repos.
func (p *objectWalker) walkObjectTree(hash plumbing.Hash) error {
// Check if we have already seen, and mark this object
if p.isSeen(hash) {
return nil
}
p.add(hash)
// Fetch the object.
obj, err := object.GetObject(p.Storer, hash)
if err != nil {
return fmt.Errorf("getting object %s failed: %v", hash, err)
}
// Walk all children depending on object type.
switch obj := obj.(type) {
case *object.Commit:
err = p.walkObjectTree(obj.TreeHash)
if err != nil {
return err
}
for _, h := range obj.ParentHashes {
err = p.walkObjectTree(h)
if err != nil {
return err
}
}
case *object.Tree:
for i := range obj.Entries {
// Shortcut for blob objects:
// 'or' the lower bits of a mode and check that it
// it matches a filemode.Executable. The type information
// is in the higher bits, but this is the cleanest way
// to handle plain files with different modes.
// Other non-tree objects are somewhat rare, so they
// are not special-cased.
if obj.Entries[i].Mode|0755 == filemode.Executable {
p.add(obj.Entries[i].Hash)
continue
}
// Normal walk for sub-trees (and symlinks etc).
err = p.walkObjectTree(obj.Entries[i].Hash)
if err != nil {
return err
}
}
case *object.Tag:
return p.walkObjectTree(obj.Target)
default:
// Error out on unhandled object types.
return fmt.Errorf("unknown object %X %s %T", obj.ID(), obj.Type(), obj)
}
return nil
}
|
