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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
|
// SPDX-License-Identifier: GPL-3.0-or-later
// Package scan walks a directory tree and reports the files krino will
// consider sorting, and why the rest were skipped.
package scan
import (
"errors"
"fmt"
"io/fs"
"os"
"path/filepath"
"sort"
"time"
"krino/internal/ignore"
)
// File is a regular file found by Walk.
type File struct {
Path string // absolute, cleaned
Rel string // slash-separated, relative to the root
Name string // base name
Size int64
ModTime time.Time
Mode fs.FileMode
}
// Reason is why an entry was not returned as a File.
type Reason int
const (
Ignored Reason = iota // matched an ignore pattern
Busy // a sibling NAME<busy-suffix> exists
TooNew // modified less than MinAge ago
Symlink // a symbolic link (never followed)
NotRegular // fifo, socket, device
Unreadable // a directory that could not be read
TooBig // larger than MaxSize
)
// String names a Reason the way it should read in a report.
func (r Reason) String() string {
switch r {
case Ignored:
return "ignored"
case Busy:
return "busy"
case TooNew:
return "too new"
case Symlink:
return "symlink"
case NotRegular:
return "not a regular file"
case Unreadable:
return "unreadable"
case TooBig:
return "too big"
default:
return fmt.Sprintf("Reason(%d)", int(r))
}
}
// Skipped is one entry Walk did not return as a File, and why.
type Skipped struct {
Rel string
Reason Reason
}
// Options controls how Walk traverses a directory.
type Options struct {
Recursive bool
MaxDepth int // 0: unlimited; 1: the root's own entries only
Ignore *ignore.Matcher // nil: nothing ignored
Exclude []string // absolute directories never entered
Busy []string // suffixes, e.g. ".part"
MinAge time.Duration
MaxSize int64 // bytes; 0: no limit
Now time.Time
}
// Result is everything Walk found under a root.
type Result struct {
Files []File // sorted by Rel
Skipped []Skipped // sorted by Rel
}
// Walk lists the files under root that krino will consider, and reports why
// the rest were skipped. root must exist, be a directory, and be readable;
// an unreadable subdirectory found during the walk is reported as
// Unreadable and does not abort the scan.
func Walk(root string, opt Options) (*Result, error) {
root, err := filepath.Abs(root)
if err != nil {
return nil, err
}
info, err := os.Stat(root)
if err != nil {
return nil, err
}
if !info.IsDir() {
return nil, fmt.Errorf("%s is not a directory", root)
}
entries, err := os.ReadDir(root)
if err != nil {
return nil, err
}
exclude := make(map[string]bool, len(opt.Exclude))
for _, e := range opt.Exclude {
exclude[filepath.Clean(e)] = true
}
w := &walker{root: root, opt: opt, exclude: exclude}
if err := w.walk(root, "", 1, entries); err != nil {
return nil, err
}
sort.Slice(w.result.Files, func(i, j int) bool { return w.result.Files[i].Rel < w.result.Files[j].Rel })
sort.Slice(w.result.Skipped, func(i, j int) bool { return w.result.Skipped[i].Rel < w.result.Skipped[j].Rel })
return &w.result, nil
}
// walker accumulates the Result across recursive calls.
type walker struct {
opt Options
root string
exclude map[string]bool
result Result
}
// walk applies the skip checks to entries, the already-read contents of dir
// (at relDir relative to the root, dir's own entries at depth). A
// subdirectory it recurses into is read here, right before recursing, so a
// ReadDir failure on it can be reported as Unreadable and skipped without
// aborting the rest of the walk; only a failure reading dir itself (passed
// in by the caller) would need to propagate, and only Walk's own read of
// the root works that way.
func (w *walker) walk(dir, relDir string, depth int, entries []os.DirEntry) error {
// The set of names in this directory, built once, for the busy check.
names := make(map[string]bool, len(entries))
for _, e := range entries {
names[e.Name()] = true
}
for _, e := range entries {
name := e.Name()
rel := name
if relDir != "" {
rel = relDir + "/" + name
}
path := filepath.Join(dir, name)
// A symlink is never followed, whatever it points to; DirEntry's
// Type is Lstat-like and does not resolve it.
if e.Type()&fs.ModeSymlink != 0 {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Symlink})
continue
}
if e.IsDir() {
// Directories themselves are not reported, with two
// exceptions: the symlink case above, and an ignored
// directory (C2) — reported once for the directory itself,
// not for each file inside it, since pruning it without
// descending is the whole point; without this, its contents
// would appear in no count and in no -v listing at all.
if !w.opt.Recursive {
continue
}
if w.opt.Ignore != nil && w.opt.Ignore.Match(rel, true) {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Ignored})
continue
}
if w.exclude[filepath.Clean(path)] {
continue
}
if w.opt.MaxDepth > 0 && depth+1 > w.opt.MaxDepth {
continue
}
subEntries, err := os.ReadDir(path)
if err != nil {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Unreadable})
continue
}
if err := w.walk(path, rel, depth+1, subEntries); err != nil {
return err
}
continue
}
info, err := e.Info()
if err != nil {
if errors.Is(err, fs.ErrNotExist) {
continue // vanished mid-walk (a download finishing, say): silently skipped
}
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Unreadable})
continue
}
if !info.Mode().IsRegular() {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: NotRegular})
continue
}
if w.opt.Ignore != nil && w.opt.Ignore.Match(rel, false) {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Ignored})
continue
}
if busy := w.isBusy(name, names); busy {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: Busy})
continue
}
if w.opt.Now.Sub(info.ModTime()) < w.opt.MinAge {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: TooNew})
continue
}
if w.opt.MaxSize > 0 && info.Size() > w.opt.MaxSize {
w.result.Skipped = append(w.result.Skipped, Skipped{Rel: rel, Reason: TooBig})
continue
}
w.result.Files = append(w.result.Files, File{
Path: path,
Rel: rel,
Name: name,
Size: info.Size(),
ModTime: info.ModTime(),
Mode: info.Mode(),
})
}
return nil
}
// isBusy reports whether name+suffix, for any configured Busy suffix, is
// among names — the sibling of an in-progress download.
func (w *walker) isBusy(name string, names map[string]bool) bool {
for _, suffix := range w.opt.Busy {
if names[name+suffix] {
return true
}
}
return false
}
|