![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to pa_vmail.C CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [parser3project] / parser3 / src / types|
Return to pa_vmail.C CVS log | Up to [parser3project] / parser3 / src / types |
1.1 paf 1: /** @file
2: Parser: @b mail class.
3: relies on gmime library, by Jeffrey Stedfast <fejj@helixcode.com>
4:
5: Copyright(c) 2001, 2002 ArtLebedev Group(http://www.artlebedev.com)
6: Author: Alexandr Petrosian <paf@design.ru>(http://paf.design.ru)
7: */
1.13 ! paf 8:
! 9: static const char* IDENT_VMAIL_C="$Id: zzz $";
1.1 paf 10:
11: #include "pa_sapi.h"
12: #include "pa_vmail.h"
13: #include "pa_vstring.h"
14: #include "pa_request.h"
15: #include "pa_common.h"
16: #include "pa_charset.h"
17: #include "pa_charsets.h"
18: #include "pa_vdate.h"
19: #include "pa_vfile.h"
20: #include "pa_uue.h"
21:
1.3 paf 22: #ifdef WITH_MAILRECEIVE
1.4 paf 23: extern "C" {
1.1 paf 24: #include "gmime.h"
1.4 paf 25: }
1.1 paf 26: #endif
27:
28: // defines
29:
30: #define RAW_NAME "raw"
31:
32: // internals
33:
34: enum PartType {
35: P_TEXT,
36: P_HTML,
37: P_FILE,
38: P_MESSAGE,
39:
40: P_TYPES_COUNT
41: };
42:
43: static const char * const part_name_starts[P_TYPES_COUNT]={"text", "html", "file", "message"};
44:
45: // VMail
46:
47: extern Methoded *mail_base_class;
48:
49: VMail::VMail(Pool& apool) : VStateless_class(apool, 0, mail_base_class),
50: vreceived(apool) {
51: }
52:
1.3 paf 53: #ifdef WITH_MAILRECEIVE
1.1 paf 54:
55: static const String& maybeUpperCase(Pool& pool, const String& src, bool toUpperCase) {
56: return toUpperCase?src.change_case(pool, String::CC_UPPER):src;
57: }
58:
59: static void UTF8toSource(Pool& pool, const char *source_body, size_t source_content_length,
60: const void *& dest_body, size_t& dest_content_length) {
61: if(source_body) {
62: if(!source_content_length)
63: source_content_length=strlen(source_body);
64: Charset::transcode(pool,
65: *utf8_charset, source_body, source_content_length,
66: pool.get_source_charset(), dest_body, dest_content_length);
67: } else {
68: dest_body=0;
69: dest_content_length=0;
70: }
71: }
72:
73: static void putReceived(Hash& received, const char *name, Value *value, bool nameToUpperCase=false) {
74: Pool& pool=received.pool();
75: if(name && value) {
76: received.put(
77: maybeUpperCase(pool, String::OnPool(pool, name, 0, true/*tainted*/), nameToUpperCase),
78: value);
79: }
80: }
81:
82: static void putReceived(Hash& received, const char *name, const char *value, size_t value_size=0, bool nameToUpperCase=false) {
83: if(value) {
84: Pool& pool=received.pool();
85:
86: const void *value_dest_body;
87: size_t value_dest_content_length;
88: UTF8toSource(pool, value, value_size, value_dest_body, value_dest_content_length);
89:
90: putReceived(received, name,
91: new(pool) VString(
92: String::OnPool(pool,(const char *)value_dest_body, value_dest_content_length, true/*tainted*/)));
93: }
94: }
95:
96: static void putReceived(Hash& received, const char *name, time_t value) {
97: Pool& pool=received.pool();
98: if(name)
99: received.put(String::OnPool(pool, name, 0, true/*tainted*/), new(pool) VDate(pool, value));
100: }
101:
102: static void MimeHeaderField2received(const char *name, const char *value, gpointer data) {
103: Hash& received=*static_cast<Hash *>(data);
104:
105: putReceived(received, name, value, 0, true/*nameInUpperCase*/);
106: }
107:
108: static void parse(GMimeStream *stream, Hash& received);
109:
110: #ifndef DOXYGEN
111: struct MimePart2bodyInfo {
112: Hash *body;
113: int partCounts[P_TYPES_COUNT];
114: };
115: #endif
116: static void MimePart2body(GMimePart *part,
117: gpointer data) {
118: MimePart2bodyInfo& i=*static_cast<MimePart2bodyInfo *>(data);
119: Pool& pool=i.body->pool();
120:
121: if(const GMimeContentType *type=g_mime_part_get_content_type(part)) {
122: if(g_mime_content_type_is_type(type, "multipart", "*"))
123: return; // skipping frames
124:
125: PartType partType;
126: if(g_mime_content_type_is_type(type, "text", "plain"))
127: partType=P_TEXT;
128: else if(g_mime_content_type_is_type(type, "text", "html"))
129: partType=P_HTML;
130: else if(g_mime_content_type_is_type(type, "message", "*"))
131: partType=P_MESSAGE;
132: else
133: partType=P_FILE;
134:
135: // partName
136: const char *partName;
137: char partNameBuf[MAX_STRING];
138: const char *partNameStart=part_name_starts[partType];
139: if(int partNo=i.partCounts[partType]++) {
140: snprintf(partNameBuf, MAX_STRING, "%s%d", partNameStart, partNo);
141: partName=partNameBuf;
142: } else
143: partName=partNameStart;
144:
145: // $.partX[
146: VHash& vpartX=*new(pool) VHash(pool); putReceived(*i.body, partName, &vpartX);
147: Hash& partX=vpartX.hash(0);
148: {
149: // $.raw[
150: VHash& vraw=*new(pool) VHash(pool); putReceived(partX, RAW_NAME, &vraw);
151: g_mime_header_foreach(part->headers, MimeHeaderField2received, &vraw.hash(0));
152: }
153: const char *content_filename=0;
154: {
155: // $.content-type[
156: VHash& vcontent_type=*new(pool) VHash(pool); putReceived(partX, "content-type", &vcontent_type);
157: Hash& content_type=vcontent_type.hash(0);
158: {
159: // $.value[text/plain]
160: char value[MAX_STRING];
161: snprintf(value, MAX_STRING, "%s/%s",
162: type->type?type->type:"x-unknown",
163: type->subtype?type->subtype:"x-unknown");
164: putReceived(content_type, VALUE_NAME, value);
165: }
166: GMimeParam *param=type->params;
167: while(param) {
168: // $.charset[windows-1251] && co
169: putReceived(content_type, param->name, param->value, true);
170: if(strcasecmp(param->name, "name")==0)
171: content_filename=param->value;
172: param=param->next;
173: }
174: }
175: // $.description
176: putReceived(partX, "description", part->description);
177: // $.content-id
178: putReceived(partX, "content-id", part->content_id);
179: // $.content-md5
180: putReceived(partX, "content-md5", part->content_md5);
181: // $.content-location
182: putReceived(partX, "content-location", part->content_location);
183:
184: // todo GMimePart:
185: // GMimePartEncodingType encoding;
186: // GMimeDisposition *disposition;
187: if(part->disposition) {
188: GMimeParam *param=part->disposition->params;
189: while(param) {
190: // $.charset[windows-1251] && co
191: if(strcasecmp(param->name, "filename")==0)
192: content_filename=param->value;
193: param=param->next;
194: }
195: }
196:
197: // MESSAGE
198: if(partType==P_MESSAGE) {
199: if(part->content)
200: if(GMimeStream *stream=part->content->stream)
201: parse(stream, partX);
202: } else {
203: // $.value[string|file]
204: size_t buf_len;
205: const void *buf=g_mime_part_get_content(part, &buf_len);
206: if(partType==P_FILE) {
207: VFile& vfile=*new(pool) VFile(pool);
208: vfile.set(true/*tainted*/, buf, buf_len, content_filename);
209: putReceived(partX, VALUE_NAME, &vfile);
210: } else {
211: // P_TEXT, P_HTML
212: putReceived(partX, VALUE_NAME,(const char*)buf, buf_len);
213: }
214: }
215: }
216: }
217:
218: static void parse(GMimeStream *stream, Hash& received) {
219: Pool& pool=received.pool();
220:
221: GMimeMessage *message=g_mime_parser_construct_message(stream);
222: try {
223: const GMimeMessageHeader *messageHeader=message->header;
224: if(!messageHeader)
225: return;
226:
227: // firstly user-defined strings go
228: // user headers
229: {
230: // $.raw[
1.5 paf 231: VHash& vraw=*new(pool) VHash(pool); putReceived(received, RAW_NAME, &vraw);
1.1 paf 232: g_mime_header_foreach(messageHeader->headers, MimeHeaderField2received, &vraw.hash(0));
233: }
234:
235: // maybe-todo-recipients
236: // x(messageHeader->recipients)
237:
238: // secondly standard headers&body go
239: // standard header
240: // .from
241: putReceived(received, "from", messageHeader->from);
242: // .reply-to
243: putReceived(received, "reply-to", messageHeader->reply_to);
244: // .to
245: // todo: messageHeader->recipients
246: // .subject
247: putReceived(received, "subject", messageHeader->subject);
248: // .date(date+gmt_offset)
249: int tt_offset =
250: ((messageHeader->gmt_offset / 100) *(60 * 60))
251: +(messageHeader->gmt_offset % 100) * 60;
252: putReceived(received, "date",
253: messageHeader->date // local sender
254: -tt_offset // move local sender to GMT sender
255: -(timezone+(daylight?60*60*sign(timezone):0)) // move GMT sender to our local time
256: );
257: // .message-id
258: putReceived(received, "message-id", messageHeader->message_id);
259:
260: // .body[part/parts
261: GMimePart *part=message->mime_part;
262: const GMimeContentType *type=g_mime_part_get_content_type(part);
263: MimePart2bodyInfo info={&received};
264: g_mime_part_foreach(part, MimePart2body, &info);
265:
266: // normal unref
267: g_mime_object_unref(GMIME_OBJECT(message));
268: } catch(...) {
269: // abnormal unref
270: g_mime_object_unref(GMIME_OBJECT(message));
271: }
272: }
273: #endif
274:
275:
276:
277: void VMail::fill_received(Request& request) {
278: // store letter to received
1.3 paf 279: #ifdef WITH_MAILRECEIVE
1.1 paf 280: if(request.info.mail_received) {
281: // init
282: g_mime_init(GMIME_INIT_FLAG_UTF8);
283:
284: // create stream with CRLF filter
285: GMimeStream *stream = g_mime_stream_fs_new(fileno(stdin));
286: GMimeStream *istream = g_mime_stream_filter_new_with_stream(stream);
287: GMimeFilter *filter = g_mime_filter_crlf_new(GMIME_FILTER_CRLF_DECODE, GMIME_FILTER_CRLF_MODE_CRLF_ONLY);
288: g_mime_stream_filter_add(GMIME_STREAM_FILTER(istream), filter);
289: g_mime_stream_unref(stream);
290: stream = istream;
291: try {
292: // parse incoming stream
293: parse(stream, vreceived.hash(0));
294: // normal stream free
295: g_mime_stream_unref(stream);
296: } catch(...) {
297: // abnormal stream free
298: g_mime_stream_unref(stream);
299: }
300: }
301: #endif
302: }
303:
304: #ifndef DOXYGEN
305: struct Store_message_element_info {
306: Charset *charset;
307: String *header;
1.11 paf 308: const String **from, **to;
309: const String *errors_to;
1.1 paf 310: Array *parts[P_TYPES_COUNT];
311: int parts_count;
312: bool has_content_type;
313: };
314: #endif
1.9 paf 315: typedef int (*string_contains_char_which_check)(int);
316: static bool string_contains_char_which(const char *string, string_contains_char_which_check check) {
317: while(char c=*string++) {
318: if(check(c))
319: return true;
320: }
321: return false;
322: }
323: const String& extractEmail(const String& string) {
324: Pool& pool=string.pool();
325:
326: char *email=string.cstr();
327: lsplit(email, '>'); lsplit(email, '\x0D');lsplit(email, '\x0A');
328: char *next=rsplit(email, '<');
329: if(next) email=next;
330:
331: String& result=*new(pool) String(pool);
332: result.APPEND_TAINTED(email, 0, string.origin().file, string.origin().line);
333:
334: /*
335: http://www.faqs.org/rfcs/rfc822.html
336:
337: addr-spec = local-part "@" domain ; global address
338:
339: local-part = word *("." word) ; uninterpreted case-preserved
340: word = atom / quoted-string
341:
342: domain = sub-domain *("." sub-domain)
343: sub-domain = domain-ref / domain-literal
344: domain-ref = atom ; symbolic reference
345:
346: domain-literal << ignoring for now
347: quoted-string in word << ignoring for now
348:
349: atom = 1*<any CHAR except specials, SPACE and CTLs> << the ONLY to check
350:
351: specials = "(" / ")" / "<" / ">" / "@" ; Must be in quoted-
352: / "," / ";" / ":" / "\" / <"> ; string, to use
353: / "." / "[" / "]" ; within a word.
354:
355: */
356: if(strpbrk(email, "()<>,;:\\\"[]"/*specials minus @ and . */))
357: throw Exception(0,
358: &result,
359: "email contains characters (specials)");
360: if(string_contains_char_which(email, (string_contains_char_which_check)isspace))
361: throw Exception(0,
362: &result,
363: "email contains characters (whitespace)");
364: if(string_contains_char_which(email, (string_contains_char_which_check)iscntrl))
365: throw Exception(0,
366: &result,
367: "email contains characters (control)");
368:
369: return result;
370: }
1.5 paf 371: static void store_message_element(const Hash::Key& raw_element_name, Hash::Val *aelement_value,
1.1 paf 372: void *info) {
373: Value& element_value=*static_cast<Value *>(aelement_value);
1.5 paf 374: const String& low_element_name=raw_element_name.change_case(raw_element_name.pool(), String::CC_LOWER);
1.1 paf 375: Store_message_element_info& i=*static_cast<Store_message_element_info *>(info);
376:
377: // exclude internals
1.5 paf 378: if(low_element_name==CHARSET_NAME
379: || low_element_name==VALUE_NAME
380: || low_element_name==RAW_NAME
381: || low_element_name=="date")
1.1 paf 382: return;
383:
384: // grep parts
385: for(int pt=0; pt<P_TYPES_COUNT; pt++) {
1.5 paf 386: if(low_element_name.starts_with(part_name_starts[pt])) {
1.1 paf 387: *i.parts[pt]+=&element_value;
388: i.parts_count++;
389: return;
390: }
391: }
392:
1.10 paf 393: // fetch some special headers
1.5 paf 394: if(i.from && low_element_name=="from")
1.9 paf 395: *i.from=&extractEmail(element_value.as_string());
1.5 paf 396: if(i.to && low_element_name=="to")
1.9 paf 397: *i.to=&extractEmail(element_value.as_string());
1.12 paf 398: if(low_element_name=="errors-to")
1.11 paf 399: i.errors_to=&extractEmail(element_value.as_string());
1.1 paf 400:
401: // append header line
402: *i.header <<
1.5 paf 403: raw_element_name << ":" <<
1.1 paf 404: attributed_meaning_to_string(element_value, String::UL_MAIL_HEADER).
405: cstr(String::UL_UNSPECIFIED, 0, i.charset, i.charset?i.charset->name().cstr():0) <<
406: "\n";
407:
408: // has content type?
1.5 paf 409: if(low_element_name==CONTENT_TYPE_NAME)
1.1 paf 410: i.has_content_type=true;
411: }
412:
413: static const String& file_value_to_string(Request& r, const String *source,
414: Value& send_value) {
415: Pool& pool=r.pool();
416: const VFile *vfile;
417: const String *file_name;
418: Value *vformat;
419: if(Hash *send_hash=send_value.get_hash(source)) { // hash
420: // $.value
421: if(Value *value=static_cast<Value *>(send_hash->get(*value_name)))
422: vfile=value->as_vfile(String::UL_AS_IS);
423: else
424: throw Exception("parser.runtime",
425: source,
426: "file part has no $value");
427:
428: // $.format
429: vformat=static_cast<Value *>(send_hash->get(*new(pool) String(pool, "format")));
430:
1.6 paf 431: // $.name
1.1 paf 432: if(Value *vfile_name=static_cast<Value *>(send_hash->get(
1.6 paf 433: *new(pool) String(pool, "name")))) // specified $name
1.1 paf 434: file_name=&vfile_name->as_string();
435: } else { // must be VFile
436: vfile=send_value.as_vfile(String::UL_AS_IS);
437: file_name=&static_cast<Value *>(vfile->fields().get(*name_name))->as_string();
438: vformat=0;
439: }
440: const char *file_name_cstr=file_name->cstr();
441:
442: String& result=*new(pool) String(pool);
443:
444: // content-type: application/octet-stream
445: result << "content-type: " << r.mime_type_of(file_name_cstr)
446: << "; name=\"" << file_name_cstr << "\"\n";
447: // content-disposition: attachment; filename="user_file_name"
448: result << "content-disposition: attachment; filename=\"" << file_name_cstr << "\"\n";
449:
450: const String *type=vformat?&vformat->as_string():0;
451: if(!type/*default = uue*/ || *type=="uue") {
452: pa_uuencode(result, file_name_cstr, *vfile);
453: } else // for now
454: throw Exception("parser.runtime",
455: type,
456: "unknown attachment encode format");
457:
458: return result;
459: }
460:
461: static const String& text_value_to_string(Request& r, const String *source,
462: PartType pt, Value& send_value,
463: Store_message_element_info& info) {
464: Pool& pool=r.pool();
465: String& result=*new(pool) String(pool);
466:
467: Value *text_value;
468: if(Hash *send_hash=send_value.get_hash(source)) {
469: // $.USER-HEADERS
470: info.has_content_type=false; // reset
471: send_hash->for_each(store_message_element, &info);
472: // $.value
473: text_value=static_cast<Value *>(send_hash->get(*value_name));
474: if(!text_value)
475: throw Exception("parser.runtime",
476: source,
477: "%s part has no $" VALUE_NAME, part_name_starts[pt]);
478: } else
479: text_value=&send_value;
480:
481: if(!info.has_content_type) {
482: result << "content-type: text/" << (pt==P_TEXT?"plain":"html");
483: if(info.charset)
484: result << "; charset=" << info.charset->name();
485: result << "\n";
486: }
487:
488: // header|body separator
489: result << "\n";
490:
491: // body
492: switch(pt) {
493: case P_TEXT:
494: result.append(text_value->as_string(), String::UL_AS_IS, true /* forced */);
495: break;
496: case P_HTML:
497: {
498: Temp_lang temp_lang(r, String::UL_HTML);
499: if(Junction *junction=text_value->get_junction()) {
500: // execution of found $.html{code} must be in context of ^send[...]
501: // setting code context, would execute in ^.send[>>context<<]
502: //junction->change_context(?.get_junction());
503: junction->root=r.root;
504: junction->rcontext=r.rcontext;
505: junction->wcontext=r.wcontext;
506:
507: result << r.process_to_string(*text_value);
508: } else
509: throw Exception("parser.runtime",
510: source,
511: "html part value must be code");
512:
513: break;
514: }
515: }
516:
517: return result;
518: };
519:
520: /// @todo files and messages in order (file, file2, ...)
521: const String& VMail::message_hash_to_string(Request& r, const String *source,
522: Hash *message_hash, int level,
523: const String **from, const String **to) {
524: if(!message_hash)
525: throw Exception("parser.runtime",
526: source,
527: "message must be hash");
528:
529: String& result=*NEW String(pool());
530:
531: Charset *charset;
532: if(Value *vrecodecharset_name=static_cast<Value *>(message_hash->get(*charset_name)))
533: charset=&charsets->get_charset(vrecodecharset_name->as_string());
534: else
535: charset=&pool().get_source_charset();
536:
537: Store_message_element_info info={
538: charset,
539: &result,
540: from, to
541: };
542: {
543: if(from)
544: *from=0;
545: if(to)
546: *to=0;
547: for(int pt=0; pt<P_TYPES_COUNT; pt++)
548: info.parts[pt]=NEW Array(pool());
549: message_hash->for_each(store_message_element, &info);
1.10 paf 550: if(!info.errors_to)
551: result << "errors-to: postmaster\n"; // errors-to: default
1.1 paf 552: }
553:
554: int textCount=info.parts[P_TEXT]->size();
555: if(textCount>1)
556: throw Exception("parser.runtime",
557: source,
558: "multiple text parts not supported, use file part");
559: int htmlCount=info.parts[P_HTML]->size();
560: if(htmlCount>1)
561: throw Exception("parser.runtime",
562: source,
563: "multiple html parts not supported, use file part");
564:
565:
566: bool multipart=info.parts_count>1;
567: bool alternative=textCount && htmlCount;
568: // header
569: char *boundary=0;
570: if(multipart) {
571: boundary=(char *)malloc(MAX_NUMBER);
572: snprintf(boundary, MAX_NUMBER-5/*lEvEl*/, "lEvEl%d", level);
573: // multi-part
574: result << "content-type: multipart/mixed; boundary=\"" << boundary << "\"\n";
575: result << "\n"
576: "This is a multi-part message in MIME format.";
577: }
578:
579: // alternative or not
580: {
581: if(alternative) {
582: result << "\n\n--" << boundary << "\n"; // intermediate boundary
583: result << "content-type: multipart/alternative; boundary=\"ALT" << boundary << "\"\n";
584: }
585: for(int i=0; i<2; i++) {
586: PartType pt=i==0?P_TEXT:P_HTML;
587: if(info.parts[pt]->size()) {
588: if(alternative)
589: result << "\n\n--ALT" << boundary << "\n"; // intermediate boundary
590: else if(boundary)
591: result << "\n\n--" << boundary << "\n"; // intermediate boundary
592: result << text_value_to_string(r, source, pt,
593: *static_cast<Value *>(info.parts[pt]->get(0)), info);
594: }
595: }
596: if(alternative)
597: result << "\n\n--ALT" << boundary << "--\n";
598: }
599:
600: // files
601: {
602: Array& files=*info.parts[P_FILE];
603: for(int i=0; i<files.size(); i++) {
604: if(boundary)
605: result << "\n\n--" << boundary << "\n"; // intermediate boundary
606: result << file_value_to_string(r, source, *static_cast<Value *>(files.get(i)));
607: }
608: }
609:
610: // messages
611: {
612: Array& messages=*info.parts[P_MESSAGE];
613: for(int i=0; i<messages.size(); i++) {
614: if(boundary)
615: result << "\n\n--" << boundary << "\n"; // intermediate boundary
616:
617: result << message_hash_to_string(r, source,
618: static_cast<Value *>(messages.get(i))->get_hash(source),
619: level+1);
620: }
621: }
622:
623: // tailer
624: if(boundary)
625: result << "\n\n--" << boundary << "--\n"; // finish boundary
626:
627: // return
628: return result;
629: }
630:
631:
632: Value *VMail::get_element(const String& aname) {
633: // $fields
1.3 paf 634: #ifdef WITH_MAILRECEIVE
1.1 paf 635: if(aname==MAIL_RECEIVED_ELEMENT_NAME)
636: return &vreceived;
637: #endif
638:
639: // $CLASS,$method
640: if(Value *result=VStateless_class::get_element(aname))
641: return result;
642:
643: return 0;
644: }
645:
1.3 paf 646: #if defined(WITH_MAILRECEIVE) && defined(_MSC_VER)
1.1 paf 647: # define GNOME_LIBS "/parser3project/win32mailreceive/win32/gnome"
648: # pragma comment(lib, GNOME_LIBS "/glib/lib/libglib-1.3-11.lib")
649: # ifdef _DEBUG
650: # pragma comment(lib, GNOME_LIBS "/gmime-x.x.x/Debug/libgmime.lib")
651: # else
652: # pragma comment(lib, GNOME_LIBS "/gmime-x.x.x/Release/libgmime.lib")
653: # endif
654: #endif