KMime

parsers.cpp
1/*
2 kmime_parsers.cpp
3
4 KMime, the KDE Internet mail/usenet news message library.
5 SPDX-FileCopyrightText: 2001 the KMime authors.
6 See file AUTHORS for details
7
8 SPDX-License-Identifier: LGPL-2.0-or-later
9*/
10#include "parsers_p.h"
11#include "util_p.h"
12
13#include <QMimeDatabase>
14#include <QRegularExpression>
15
16#include <cctype>
17
18using namespace KMime::Parser;
19
20namespace KMime
21{
22namespace Parser
23{
24
25MultiPart::MultiPart(const QByteArray &src, const QByteArray &boundary)
26 : m_src(src)
27 , m_boundary(boundary)
28{
29}
30
31bool MultiPart::parse()
32{
33 QByteArray b = "--" + m_boundary;
34 QByteArray part;
35 qsizetype pos1 = 0;
36 qsizetype pos2 = 0;
37 auto blen = b.length();
38
39 m_parts.clear();
40
41 //find the first valid boundary
42 while (true) {
43 if ((pos1 = m_src.indexOf(b, pos1)) == -1 || pos1 == 0 ||
44 m_src[pos1 - 1] == '\n') { //valid boundary found or no boundary at all
45 break;
46 }
47 pos1 += blen; //boundary found but not valid => skip it;
48 }
49
50 if (pos1 > -1) {
51 pos1 += blen;
52 if (m_src[pos1] == '-' && m_src[pos1 + 1] == '-') {
53 // the only valid boundary is the end-boundary
54 // this message is *really* broken
55 pos1 = -1; //we give up
56 } else if ((pos1 - blen) > 1) { //preamble present
57 m_preamble = m_src.left(pos1 - blen - 1);
58 }
59 }
60
61 while (pos1 > -1 && pos2 > -1) {
62
63 //skip the rest of the line for the first boundary - the message-part starts here
64 if ((pos1 = m_src.indexOf('\n', pos1)) > -1) {
65 //now search the next linebreak
66 //now find the next valid boundary
67 pos2 = ++pos1; //pos1 and pos2 point now to the beginning of the next line after the boundary
68 while (true) {
69 if ((pos2 = m_src.indexOf(b, pos2)) == -1 ||
70 m_src[pos2 - 1] == '\n') { //valid boundary or no more boundaries found
71 break;
72 }
73 pos2 += blen; //boundary is invalid => skip it;
74 }
75
76 if (pos2 == -1) { // no more boundaries found
77 part = m_src.mid(pos1, m_src.length() - pos1); //take the rest of the string
78 m_parts.append(part);
79 pos1 = -1;
80 pos2 = -1; //break;
81 } else {
82 part = m_src.mid(pos1, pos2 - pos1 - 1); // pos2 - 1 (\n) is part of the boundary (see RFC 2046, section 5.1.1)
83 m_parts.append(part);
84 pos2 += blen; //pos2 points now to the first character after the boundary
85 if (m_src[pos2] == '-' && m_src[pos2 + 1] == '-') { //end-boundary
86 pos1 = pos2 + 2; //pos1 points now to the character directly after the end-boundary
87
88 if ((pos1 = m_src.indexOf('\n', pos1)) > -1) { //skip the rest of this line
89 //everything after the end-boundary is considered as the epilouge
90 m_epilouge = m_src.mid(pos1 + 1, m_src.length() - pos1 - 1);
91 }
92 pos1 = -1;
93 pos2 = -1; //break
94 } else {
95 pos1 = pos2; //the search continues ...
96 }
97 }
98 }
99 }
100
101 return !m_parts.isEmpty();
102}
103
104//=============================================================================
105
106NonMimeParser::NonMimeParser(const QByteArray &src) :
107 m_src(src), m_partNr(-1), m_totalNr(-1)
108{
109}
110
111NonMimeParser::~NonMimeParser() = default;
112
113//==============================================================================
114
115[[nodiscard]] static qsizetype findUuencodeBeginMarker(const QByteArray &s, qsizetype startPos)
116{
117 auto idx = startPos;
118 while (true) {
119 idx = s.indexOf("begin ", idx);
120 if (idx < 0 || idx + 9 >= s.size()) {
121 break;
122 }
123 if (std::isdigit(s[idx + 6]) && std::isdigit(s[idx + 7]) && std::isdigit(s[idx + 8])) {
124 return idx;
125 }
126 idx += 6;
127 }
128 return -1;
129}
130
131UUEncoded::UUEncoded(const QByteArray &src, const QByteArray &head) :
132 NonMimeParser(src), m_head(head)
133{}
134
135bool UUEncoded::parse()
136{
137 qsizetype currentPos = 0;
138 bool success = true;
139 bool firstIteration = true;
140
141 while (success) {
142 qsizetype beginPos = currentPos;
143 qsizetype uuStart = currentPos;
144 qsizetype endPos = 0;
145 int lineCount = 0;
146 int MCount = 0;
147 qsizetype pos = 0;
148 qsizetype len = 0;
149 bool containsBegin = false;
150 bool containsEnd = false;
151 QByteArray tmp;
152 QByteArray fileName;
153
154 if ((beginPos = findUuencodeBeginMarker(m_src, currentPos)) > -1 &&
155 (beginPos == 0 || m_src.at(beginPos - 1) == '\n')) {
156 containsBegin = true;
157 uuStart = m_src.indexOf('\n', beginPos);
158 if (uuStart == -1) { //no more line breaks found, we give up
159 success = false;
160 break;
161 } else {
162 uuStart++; //points now at the beginning of the next line
163 }
164 } else {
165 beginPos = currentPos;
166 }
167
168 if (!containsBegin || (endPos = m_src.indexOf("\nend", (uuStart > 0) ? uuStart - 1 : 0)) == -1) {
169 endPos = m_src.length(); //no end found
170 } else {
171 containsEnd = true;
172 }
173
174 if ((containsBegin && containsEnd) || firstIteration) {
175
176 //printf("beginPos=%d , uuStart=%d , endPos=%d\n", beginPos, uuStart, endPos);
177 //all lines in a uuencoded text start with 'M'
178 for (auto idx = uuStart; idx < endPos; idx++) {
179 if (m_src.at(idx) == '\n') {
180 lineCount++;
181 if (idx + 1 < endPos && m_src.at(idx + 1) == 'M') {
182 idx++;
183 MCount++;
184 }
185
186 // partial version of the check below, for things that can be evaluated
187 // while this loop is still counting
188 if ((lineCount - MCount) > 10) {
189 success = false;
190 break;
191 }
192 }
193 }
194
195 //printf("lineCount=%d , MCount=%d\n", lineCount, MCount);
196 if (MCount == 0 || (lineCount - MCount) > 10 || ((!containsBegin || !containsEnd) && (MCount < 15))) {
197 // harder check for split-articles
198 success = false;
199 break; //too many "non-M-Lines" found, we give up
200 }
201
202 const auto subject = KMime::extractHeader(m_head, "Subject");
203 if ((!containsBegin || !containsEnd) && !subject.isNull()) {
204 // message may be split up => parse subject
205 const QRegularExpression subjectRegex(QStringLiteral("[0-9]+/[0-9]+"));
206 const auto match = subjectRegex.match(QLatin1StringView(subject));
207 pos = match.capturedStart(0);
208 len = match.capturedLength(0);
209 if (pos != -1) {
210 tmp = subject.mid(pos, len);
211 pos = tmp.indexOf('/');
212 m_partNr = tmp.left(pos).toInt();
213 m_totalNr = tmp.right(tmp.length() - pos - 1).toInt();
214 } else {
215 success = false;
216 break; //no "part-numbers" found in the subject, we give up
217 }
218 }
219
220 //everything before "begin" is text
221 if (beginPos > 0) {
222 m_text.append(m_src.mid(currentPos, beginPos - currentPos));
223 }
224
225 if (containsBegin) {
226 //everything between "begin ### " and the next LF is considered as the filename
227 fileName = m_src.mid(beginPos + 10, uuStart - beginPos - 11);
228 } else {
229 fileName = "";
230 }
231 m_filenames.append(fileName);
232 //everything between "begin" and "end" is uuencoded
233 m_bins.append(m_src.mid(uuStart, endPos - uuStart + 1));
234 QMimeDatabase db;
235 m_mimeTypes.append(db.mimeTypeForFile(QString::fromUtf8(fileName), QMimeDatabase::MatchExtension).name().toUtf8());
236 firstIteration = false;
237
238 auto next = m_src.indexOf('\n', endPos + 1);
239 if (next == -1) { //no more line breaks found, we give up
240 success = false;
241 break;
242 } else {
243 next++; //points now at the beginning of the next line
244 }
245 currentPos = next;
246
247 } else {
248 success = false;
249 }
250 }
251
252 // append trailing text part of the article (only
253 if (!m_bins.isEmpty() || isPartial()) {
254 m_text.append(m_src.right(m_src.length() - currentPos));
255 return true;
256 }
257
258 return false;
259}
260
261//==============================================================================
262
263YENCEncoded::YENCEncoded(const QByteArray &src) :
264 NonMimeParser(src)
265{
266}
267
268bool YENCEncoded::yencMeta(QByteArray &src, const QByteArray &name, int *value)
269{
270 bool found = false;
271 QByteArray sought = name + '=';
272
273 auto iPos = src.indexOf(sought);
274 if (iPos > -1) {
275 auto pos1 = src.indexOf(' ', iPos);
276 auto pos2 = src.indexOf('\r', iPos);
277 auto pos3 = src.indexOf('\t', iPos);
278 auto pos4 = src.indexOf('\n', iPos);
279 if (pos2 >= 0 && (pos1 < 0 || pos1 > pos2)) {
280 pos1 = pos2;
281 }
282 if (pos3 >= 0 && (pos1 < 0 || pos1 > pos3)) {
283 pos1 = pos3;
284 }
285 if (pos4 >= 0 && (pos1 < 0 || pos1 > pos4)) {
286 pos1 = pos4;
287 }
288 iPos = src.lastIndexOf('=', pos1) + 1;
289 if (iPos < pos1) {
290 char c = src.at(iPos);
291 if (c >= '0' && c <= '9') {
292 found = true;
293 *value = src.mid(iPos, pos1 - iPos).toInt();
294 }
295 }
296 }
297 return found;
298}
299
300bool YENCEncoded::parse()
301{
302 qsizetype currentPos = 0;
303 bool success = true;
304 while (success) {
305 qsizetype beginPos = currentPos;
306 qsizetype yencStart = currentPos;
307 bool containsPart = false;
308 QByteArray fileName;
309
310 if ((beginPos = m_src.indexOf("=ybegin ", currentPos)) > -1 &&
311 (beginPos == 0 || m_src.at(beginPos - 1) == '\n')) {
312 yencStart = m_src.indexOf('\n', beginPos);
313 if (yencStart == -1) { // no more line breaks found, give up
314 success = false;
315 break;
316 } else {
317 yencStart++;
318 if (m_src.indexOf("=ypart", yencStart) == yencStart) {
319 containsPart = true;
320 yencStart = m_src.indexOf('\n', yencStart);
321 if (yencStart == -1) {
322 success = false;
323 break;
324 }
325 yencStart++;
326 }
327 }
328 // Try to identify yenc meta data
329
330 // Filenames can contain any embedded chars until end of line
331 QByteArray meta = m_src.mid(beginPos, yencStart - beginPos);
332 qsizetype namePos = meta.indexOf("name=");
333 if (namePos == -1) {
334 success = false;
335 break;
336 }
337 qsizetype eolPos = meta.indexOf('\r', namePos);
338 if (eolPos == -1) {
339 eolPos = meta.indexOf('\n', namePos);
340 }
341 if (eolPos == -1) {
342 success = false;
343 break;
344 }
345 fileName = meta.mid(namePos + 5, eolPos - (namePos + 5));
346
347 // Other metadata is integer
348 int yencLine;
349 if (!yencMeta(meta, "line", &yencLine)) {
350 success = false;
351 break;
352 }
353 int yencSize;
354 if (!yencMeta(meta, "size", &yencSize)) {
355 success = false;
356 break;
357 }
358
359 int partBegin;
360 int partEnd;
361 if (containsPart) {
362 if (!yencMeta(meta, "part", &m_partNr)) {
363 success = false;
364 break;
365 }
366 if (!yencMeta(meta, "begin", &partBegin) ||
367 !yencMeta(meta, "end", &partEnd)) {
368 success = false;
369 break;
370 }
371 if (!yencMeta(meta, "total", &m_totalNr)) {
372 m_totalNr = m_partNr + 1;
373 }
374 if (yencSize == partEnd - partBegin + 1) {
375 m_totalNr = 1;
376 } else {
377 yencSize = partEnd - partBegin + 1;
378 }
379 }
380
381 // We have a valid yenc header; now we extract the binary data
382 int totalSize = 0;
383 qsizetype pos = yencStart;
384 qsizetype len = m_src.length();
385 bool lineStart = true;
386 int lineLength = 0;
387 bool containsEnd = false;
388 QByteArray binary;
389 binary.resize(yencSize);
390 while (pos < len) {
391 int ch = m_src.at(pos);
392 if (ch < 0) {
393 ch += 256;
394 }
395 if (ch == '\r') {
396 if (lineLength != yencLine && totalSize != yencSize) {
397 break;
398 }
399 pos++;
400 } else if (ch == '\n') {
401 lineStart = true;
402 lineLength = 0;
403 pos++;
404 } else {
405 if (ch == '=') {
406 if (pos + 1 < len) {
407 ch = m_src.at(pos + 1);
408 if (lineStart && ch == 'y') {
409 containsEnd = true;
410 break;
411 }
412 pos += 2;
413 ch -= 64 + 42;
414 if (ch < 0) {
415 ch += 256;
416 }
417 if (totalSize >= yencSize) {
418 break;
419 }
420 binary[totalSize++] = ch;
421 lineLength++;
422 } else {
423 break;
424 }
425 } else {
426 ch -= 42;
427 if (ch < 0) {
428 ch += 256;
429 }
430 if (totalSize >= yencSize) {
431 break;
432 }
433 binary[totalSize++] = ch;
434 lineLength++;
435 pos++;
436 }
437 lineStart = false;
438 }
439 }
440
441 if (!containsEnd) {
442 success = false;
443 break;
444 }
445 if (totalSize != yencSize) {
446 success = false;
447 break;
448 }
449
450 // pos now points to =yend; get end data
451 eolPos = m_src.indexOf('\n', pos);
452 if (eolPos == -1) {
453 success = false;
454 break;
455 }
456 meta = m_src.mid(pos, eolPos - pos);
457 if (!yencMeta(meta, "size", &totalSize)) {
458 success = false;
459 break;
460 }
461 if (totalSize != yencSize) {
462 success = false;
463 break;
464 }
465
466 m_filenames.append(fileName);
467 QMimeDatabase db;
468 m_mimeTypes.append(db.mimeTypeForFile(QString::fromUtf8(fileName), QMimeDatabase::MatchExtension).name().toUtf8());
469 m_bins.append(binary);
470
471 //everything before "begin" is text
472 if (beginPos > 0) {
473 m_text.append(m_src.mid(currentPos, beginPos - currentPos));
474 }
475 currentPos = eolPos + 1;
476
477 } else {
478 success = false;
479 }
480 }
481
482 // append trailing text part of the article
483 if (!m_bins.isEmpty()) {
484 m_text.append(m_src.right(m_src.length() - currentPos));
485 return true;
486 }
487
488 return false;
489}
490
491} // namespace Parser
492
493} // namespace KMime
KFuzzyMatcher::match
KCOREADDONS_EXPORT Result match(QStringView pattern, QStringView str)
KStandardActions::name
QString name(StandardAction id)
KStandardShortcut::next
const QList< QKeySequence > & next()
QByteArray::append
QByteArray & append(QByteArrayView data)
QByteArray::at
char at(qsizetype i) const const
QByteArray::indexOf
qsizetype indexOf(QByteArrayView bv, qsizetype from) const const
QByteArray::lastIndexOf
qsizetype lastIndexOf(QByteArrayView bv) const const
QByteArray::left
QByteArray left(qsizetype len) const const
QByteArray::length
qsizetype length() const const
QByteArray::mid
QByteArray mid(qsizetype pos, qsizetype len) const const
QByteArray::resize
void resize(qsizetype newSize, char c)
QByteArray::right
QByteArray right(qsizetype len) const const
QByteArray::toInt
int toInt(bool *ok, int base) const const
QMimeDatabase::mimeTypeForFile
QMimeType mimeTypeForFile(const QFileInfo &fileInfo, MatchMode mode) const const
QString::fromUtf8
QString fromUtf8(QByteArrayView str)
This file is part of the KDE documentation.
Documentation copyright © 1996-2024 The KDE developers.
Generated on Mon Nov 18 2024 12:18:08 by doxygen 1.12.0 written by Dimitri van Heesch, © 1997-2006

KDE's Doxygen guidelines are available online.