forked from client9/misspell
-
Notifications
You must be signed in to change notification settings - Fork 0
/
mime.go
210 lines (187 loc) · 5.73 KB
/
mime.go
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
package misspell
import (
"bytes"
"fmt"
"io"
"io/ioutil"
"net/http"
"os"
"path/filepath"
"strings"
)
// The number of possible binary formats is very large
// items that might be checked into a repo or be an
// artifact of a build. Additions welcome.
//
// Golang's internal table is very small and can't be
// relied on. Even then things like ".js" have a mime
// type of "application/javascipt" which isn't very helpful.
// "[x]" means we have sniff test and suffix test should be eliminated
var binary = map[string]bool{
".a": true, // [ ] archive
".bin": true, // [ ] binary
".bz2": true, // [ ] compression
".class": true, // [x] Java class file
".dll": true, // [ ] shared library
".exe": true, // [ ] binary
".gif": true, // [ ] image
".gpg": true, // [x] text, but really all base64
".gz": true, // [ ] compression
".ico": true, // [ ] image
".jar": true, // [x] archive
".jpeg": true, // [ ] image
".jpg": true, // [ ] image
".mp3": true, // [ ] audio
".mp4": true, // [ ] video
".mpeg": true, // [ ] video
".o": true, // [ ] object file
".pdf": true, // [x] pdf
".png": true, // [x] image
".pyc": true, // [ ] Python bytecode
".pyo": true, // [ ] Python bytecode
".so": true, // [x] shared library
".swp": true, // [ ] vim swap file
".tar": true, // [ ] archive
".tiff": true, // [ ] image
".woff": true, // [ ] font
".woff2": true, // [ ] font
".xz": true, // [ ] compression
".z": true, // [ ] compression
".zip": true, // [x] archive
}
// isBinaryFilename returns true if the file is likely to be binary
//
// Better heuristics could be done here, in particular a binary
// file is unlikely to be UTF-8 encoded. However this is cheap
// and will solve the immediate need of making sure common
// binary formats are not corrupted by mistake.
func isBinaryFilename(s string) bool {
return binary[strings.ToLower(filepath.Ext(s))]
}
var scm = map[string]bool{
".bzr": true,
".git": true,
".hg": true,
".svn": true,
"CVS": true,
}
// isSCMPath returns true if the path is likely part of a (private) SCM
// directory. E.g. ./git/something = true
func isSCMPath(s string) bool {
// hack for .git/COMMIT_EDITMSG and .git/TAG_EDITMSG
// normally we don't look at anything in .git
// but COMMIT_EDITMSG and TAG_EDITMSG are used as
// temp files for git commits. Allowing misspell to inspect
// these files allows for commit-msg hooks
// https://git-scm.com/book/en/v2/Customizing-Git-Git-Hooks
if strings.Contains(filepath.Base(s), "EDITMSG") {
return false
}
parts := strings.Split(filepath.Clean(s), string(filepath.Separator))
for _, dir := range parts {
if scm[dir] {
return true
}
}
return false
}
var magicHeaders = [][]byte{
// Issue #68
// PGP messages and signatures are "text" but really just
// blobs of base64-text and should not be misspell-checked
[]byte("-----BEGIN PGP MESSAGE-----"),
[]byte("-----BEGIN PGP SIGNATURE-----"),
// ELF
{0x7f, 0x45, 0x4c, 0x46},
// Postscript
{0x25, 0x21, 0x50, 0x53},
// PDF
{0x25, 0x50, 0x44, 0x46},
// Java class file
// https://en.wikipedia.org/wiki/Java_class_file
{0xCA, 0xFE, 0xBA, 0xBE},
// PNG
// https://en.wikipedia.org/wiki/Portable_Network_Graphics
{0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a},
// ZIP, JAR, ODF, OOXML
{0x50, 0x4B, 0x03, 0x04},
{0x50, 0x4B, 0x05, 0x06},
{0x50, 0x4B, 0x07, 0x08},
}
func isTextFile(raw []byte) bool {
for _, magic := range magicHeaders {
if bytes.HasPrefix(raw, magic) {
return false
}
}
// allow any text/ type with utf-8 encoding
// DetectContentType sometimes returns charset=utf-16 for XML stuff
// in which case ignore.
mime := http.DetectContentType(raw)
return strings.HasPrefix(mime, "text/") && strings.HasSuffix(mime, "charset=utf-8")
}
// ReadTextFile returns the contents of a file, first testing if it is a text file
// returns ("", nil) if not a text file
// returns ("", error) if error
// returns (string, nil) if text
//
// unfortunately, in worse case, this does
// 1 stat
// 1 open,read,close of 512 bytes
// 1 more stat,open, read everything, close (via ioutil.ReadAll)
// This could be kinder to the filesystem.
//
// This uses some heuristics of the file's extension (e.g. .zip, .txt) and
// uses a sniffer to determine if the file is text or not.
// Using file extensions isn't great, but probably
// good enough for real-world use.
// Golang's built in sniffer is problematic for differnet reasons. It's
// optimized for HTML, and is very limited in detection. It would be good
// to explicitly add some tests for ELF/DWARF formats to make sure we never
// corrupt binary files.
func ReadTextFile(filename string) (string, error) {
if isBinaryFilename(filename) {
return "", nil
}
if isSCMPath(filename) {
return "", nil
}
fstat, err := os.Stat(filename)
if err != nil {
return "", fmt.Errorf("Unable to stat %q: %s", filename, err)
}
// directory: nothing to do.
if fstat.IsDir() {
return "", nil
}
// avoid reading in multi-gig files
// if input is large, read the first 512 bytes to sniff type
// if not-text, then exit
isText := false
if fstat.Size() > 50000 {
fin, err := os.Open(filename)
if err != nil {
return "", fmt.Errorf("Unable to open large file %q: %s", filename, err)
}
defer fin.Close()
buf := make([]byte, 512)
_, err = io.ReadFull(fin, buf)
if err != nil {
return "", fmt.Errorf("Unable to read 512 bytes from %q: %s", filename, err)
}
if !isTextFile(buf) {
return "", nil
}
// set so we don't double check this file
isText = true
}
// read in whole file
raw, err := ioutil.ReadFile(filename)
if err != nil {
return "", fmt.Errorf("Unable to read all %q: %s", filename, err)
}
if !isText && !isTextFile(raw) {
return "", nil
}
return string(raw), nil
}