forked from jhillyerd/enmime
-
Notifications
You must be signed in to change notification settings - Fork 0
/
part.go
428 lines (398 loc) · 13 KB
/
part.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
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
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
package enmime
import (
"bufio"
"bytes"
"encoding/base64"
"io"
"io/ioutil"
"mime/quotedprintable"
"net/textproto"
"strconv"
"strings"
"time"
"github.com/gogs/chardet"
"github.com/jhillyerd/enmime/internal/coding"
"github.com/jhillyerd/enmime/mediatype"
"github.com/pkg/errors"
)
const (
minCharsetConfidence = 85
minCharsetRuneLength = 100
)
// Part represents a node in the MIME multipart tree. The Content-Type, Disposition and File Name
// are parsed out of the header for easier access.
type Part struct {
PartID string // PartID labels this parts position within the tree.
Parent *Part // Parent of this part (can be nil.)
FirstChild *Part // FirstChild is the top most child of this part.
NextSibling *Part // NextSibling of this part.
Header textproto.MIMEHeader // Header for this Part.
Boundary string // Boundary marker used within this part.
ContentID string // ContentID header for cid URL scheme.
ContentType string // ContentType header without parameters.
ContentTypeParams map[string]string // Params, added to ContentType header.
Disposition string // Content-Disposition header without parameters.
FileName string // The file-name from disposition or type header.
FileModDate time.Time // The modification date of the file.
Charset string // The content charset encoding, may differ from charset in header.
OrigCharset string // The original content charset when a different charset was detected.
Errors []*Error // Errors encountered while parsing this part.
Content []byte // Content after decoding, UTF-8 conversion if applicable.
Epilogue []byte // Epilogue contains data following the closing boundary marker.
}
// NewPart creates a new Part object.
func NewPart(contentType string) *Part {
return &Part{
Header: make(textproto.MIMEHeader),
ContentType: contentType,
ContentTypeParams: make(map[string]string),
}
}
// AddChild adds a child part to either FirstChild or the end of the children NextSibling chain.
// The child may have siblings and children attached. This method will set the Parent field on
// child and all its siblings. Safe to call on nil.
func (p *Part) AddChild(child *Part) {
if p == child {
// Prevent paradox.
return
}
if p != nil {
if p.FirstChild == nil {
// Make it the first child.
p.FirstChild = child
} else {
// Append to sibling chain.
current := p.FirstChild
for current.NextSibling != nil {
current = current.NextSibling
}
if current == child {
// Prevent infinite loop.
return
}
current.NextSibling = child
}
}
// Update all new first-level children Parent pointers.
for c := child; c != nil; c = c.NextSibling {
if c == c.NextSibling {
// Prevent infinite loop.
return
}
c.Parent = p
}
}
// TextContent indicates whether the content is text based on its content type. This value
// determines what content transfer encoding scheme to use.
func (p *Part) TextContent() bool {
if p.ContentType == "" {
// RFC 2045: no CT is equivalent to "text/plain; charset=us-ascii"
return true
}
return strings.HasPrefix(p.ContentType, "text/") ||
strings.HasPrefix(p.ContentType, ctMultipartPrefix)
}
// setupHeaders reads the header, then populates the MIME header values for this Part.
func (p *Part) setupHeaders(r *bufio.Reader, defaultContentType string) error {
header, err := readHeader(r, p)
if err != nil {
return err
}
p.Header = header
ctype := header.Get(hnContentType)
if ctype == "" {
if defaultContentType == "" {
p.addWarning(ErrorMissingContentType, "MIME parts should have a Content-Type header")
return nil
}
ctype = defaultContentType
}
// Parse Content-Type header.
mtype, mparams, minvalidParams, err := mediatype.Parse(ctype)
if err != nil {
return err
}
for i := range minvalidParams {
p.addWarning(
ErrorMalformedHeader,
"Content-Type header has malformed parameter %q",
minvalidParams[i])
}
p.ContentType = mtype
// Set disposition, filename, charset if available.
p.setupContentHeaders(mparams)
p.Boundary = mparams[hpBoundary]
p.ContentID = coding.FromIDHeader(header.Get(hnContentID))
return nil
}
// setupContentHeaders uses Content-Type media params and Content-Disposition headers to populate
// the disposition, filename, and charset fields.
func (p *Part) setupContentHeaders(mediaParams map[string]string) {
// Determine content disposition, filename, character set.
disposition, dparams, _, err := mediatype.Parse(p.Header.Get(hnContentDisposition))
if err == nil {
// Disposition is optional
p.Disposition = disposition
p.FileName = coding.DecodeExtHeader(dparams[hpFilename])
}
if p.FileName == "" && mediaParams[hpName] != "" {
p.FileName = coding.DecodeExtHeader(mediaParams[hpName])
}
if p.FileName == "" && mediaParams[hpFile] != "" {
p.FileName = coding.DecodeExtHeader(mediaParams[hpFile])
}
if p.Charset == "" {
p.Charset = mediaParams[hpCharset]
}
if p.FileModDate.IsZero() {
p.FileModDate, _ = time.Parse(time.RFC822, mediaParams[hpModDate])
}
}
// convertFromDetectedCharset attempts to detect the character set for the given part, and returns
// an io.Reader that will convert from that charset to UTF-8. If the charset cannot be detected,
// this method adds a warning to the part and automatically falls back to using
// `convertFromStatedCharset` and returns the reader from that method.
func (p *Part) convertFromDetectedCharset(r io.Reader) (io.Reader, error) {
// Attempt to detect character set from part content.
var cd *chardet.Detector
switch p.ContentType {
case "text/html":
cd = chardet.NewHtmlDetector()
default:
cd = chardet.NewTextDetector()
}
buf, err := ioutil.ReadAll(r)
if err != nil {
return nil, errors.WithStack(err)
}
cs, err := cd.DetectBest(buf)
switch err {
case nil:
// Carry on
default:
p.addWarning(ErrorCharsetDeclaration, "charset could not be detected: %v", err)
}
// Restore r.
r = bytes.NewReader(buf)
if cs == nil || cs.Confidence < minCharsetConfidence || len(bytes.Runes(buf)) < minCharsetRuneLength {
// Low confidence or not enough characters, use declared character set.
return p.convertFromStatedCharset(r), nil
}
// Confidence exceeded our threshold, use detected character set.
if p.Charset != "" && !strings.EqualFold(cs.Charset, p.Charset) {
p.addWarning(ErrorCharsetDeclaration,
"declared charset %q, detected %q, confidence %d",
p.Charset, cs.Charset, cs.Confidence)
}
if reader, err := coding.NewCharsetReader(cs.Charset, r); err == nil {
r = reader
p.OrigCharset = p.Charset
p.Charset = cs.Charset
}
return r, nil
}
// convertFromStatedCharset returns a reader that will convert from the charset specified for the
// current `*Part` to UTF-8. In case of error, or an unhandled character set, a warning will be
// added to the `*Part` and the original io.Reader will be returned.
func (p *Part) convertFromStatedCharset(r io.Reader) io.Reader {
if p.Charset == "" {
// US-ASCII. Just read.
return r
}
reader, err := coding.NewCharsetReader(p.Charset, r)
if err != nil {
// Failed to get a conversion reader.
p.addWarning(ErrorCharsetConversion, "failed to get reader for charset %q: %v", p.Charset, err)
} else {
return reader
}
// Try to parse charset again here to see if we can salvage some badly formed
// ones like charset="charset=utf-8".
charsetp := strings.Split(p.Charset, "=")
if strings.ToLower(charsetp[0]) == "charset" && len(charsetp) > 1 {
p.Charset = charsetp[1]
reader, err = coding.NewCharsetReader(p.Charset, r)
if err != nil {
// Failed to get a conversion reader.
p.addWarning(ErrorCharsetConversion, "failed to get reader for charset %q: %v", p.Charset, err)
} else {
return reader
}
}
return r
}
// decodeContent performs transport decoding (base64, quoted-printable) and charset decoding,
// placing the result into Part.Content. IO errors will be returned immediately; other errors
// and warnings will be added to Part.Errors.
func (p *Part) decodeContent(r io.Reader) error {
// contentReader will point to the end of the content decoding pipeline.
contentReader := r
// b64cleaner aggregates errors, must maintain a reference to it to get them later.
var b64cleaner *coding.Base64Cleaner
// Build content decoding reader.
encoding := p.Header.Get(hnContentEncoding)
validEncoding := true
switch strings.ToLower(encoding) {
case cteQuotedPrintable:
contentReader = coding.NewQPCleaner(contentReader)
contentReader = quotedprintable.NewReader(contentReader)
case cteBase64:
b64cleaner = coding.NewBase64Cleaner(contentReader)
contentReader = base64.NewDecoder(base64.RawStdEncoding, b64cleaner)
case cte8Bit, cte7Bit, cteBinary, "":
// No decoding required.
default:
// Unknown encoding.
validEncoding = false
p.addWarning(
ErrorContentEncoding,
"Unrecognized Content-Transfer-Encoding type %q",
encoding)
}
// Build charset decoding reader.
if validEncoding && strings.HasPrefix(p.ContentType, "text/") {
var err error
contentReader, err = p.convertFromDetectedCharset(contentReader)
if err != nil {
return p.base64CorruptInputCheck(err)
}
}
// Decode and store content.
content, err := ioutil.ReadAll(contentReader)
if err != nil {
return p.base64CorruptInputCheck(errors.WithStack(err))
}
p.Content = content
// Collect base64 errors.
if b64cleaner != nil {
for _, err := range b64cleaner.Errors {
p.addWarning(ErrorMalformedBase64, err.Error())
}
}
// Set empty content-type error.
if p.ContentType == "" {
p.addWarning(
ErrorMissingContentType, "content-type is empty for part id: %s", p.PartID)
}
return nil
}
// base64CorruptInputCheck will avoid fatal failure on corrupt base64 input
//
// This is a switch on errors.Cause(err).(type) for base64.CorruptInputError
func (p *Part) base64CorruptInputCheck(err error) error {
switch errors.Cause(err).(type) {
case base64.CorruptInputError:
p.Content = nil
p.Errors = append(p.Errors, &Error{
Name: ErrorMalformedBase64,
Detail: err.Error(),
Severe: true,
})
return nil
default:
return err
}
}
// Clone returns a clone of the current Part.
func (p *Part) Clone(parent *Part) *Part {
if p == nil {
return nil
}
newPart := &Part{
PartID: p.PartID,
Header: p.Header,
Parent: parent,
Boundary: p.Boundary,
ContentID: p.ContentID,
ContentType: p.ContentType,
Disposition: p.Disposition,
FileName: p.FileName,
Charset: p.Charset,
Errors: p.Errors,
Content: p.Content,
Epilogue: p.Epilogue,
}
newPart.FirstChild = p.FirstChild.Clone(newPart)
newPart.NextSibling = p.NextSibling.Clone(parent)
return newPart
}
// ReadParts reads a MIME document from the provided reader and parses it into tree of Part objects.
func ReadParts(r io.Reader) (*Part, error) {
br := bufio.NewReader(r)
root := &Part{PartID: "0"}
// Read header; top-level default CT is text/plain us-ascii according to RFC 822.
err := root.setupHeaders(br, `text/plain; charset="us-ascii"`)
if err != nil {
return nil, err
}
if strings.HasPrefix(root.ContentType, ctMultipartPrefix) {
// Content is multipart, parse it.
err = parseParts(root, br)
if err != nil {
return nil, err
}
} else {
// Content is text or data, decode it.
if err := root.decodeContent(br); err != nil {
return nil, err
}
}
return root, nil
}
// parseParts recursively parses a MIME multipart document and sets each Parts PartID.
func parseParts(parent *Part, reader *bufio.Reader) error {
firstRecursion := parent.Parent == nil
// Loop over MIME boundaries.
br := newBoundaryReader(reader, parent.Boundary)
for indexPartID := 1; true; indexPartID++ {
next, err := br.Next()
if err != nil && errors.Cause(err) != io.EOF {
return err
}
if br.unbounded {
parent.addWarning(ErrorMissingBoundary, "Boundary %q was not closed correctly",
parent.Boundary)
}
if !next {
break
}
p := &Part{}
// Set this Part's PartID, indicating its position within the MIME Part tree.
if firstRecursion {
p.PartID = strconv.Itoa(indexPartID)
} else {
p.PartID = parent.PartID + "." + strconv.Itoa(indexPartID)
}
// Look for part header.
bbr := bufio.NewReader(br)
if err = p.setupHeaders(bbr, ""); err != nil {
return err
}
// Insert this Part into the MIME tree.
parent.AddChild(p)
if p.Boundary == "" {
// Content is text or data, decode it.
if err = p.decodeContent(bbr); err != nil {
return err
}
} else {
// Content is another multipart.
err = parseParts(p, bbr)
if err != nil {
return err
}
}
}
// Store any content following the closing boundary marker into the epilogue.
epilogue, err := ioutil.ReadAll(reader)
if err != nil {
return errors.WithStack(err)
}
parent.Epilogue = epilogue
// If a Part is "multipart/" Content-Type, it will have .0 appended to its PartID
// i.e. it is the root of its MIME Part subtree.
if !firstRecursion {
parent.PartID += ".0"
}
return nil
}