Annotation of parser3/src/types/pa_vmail.C, revision 1.148
1.1 paf 1: /** @file
2: Parser: @b mail class.
3: relies on gmime library, by Jeffrey Stedfast <fejj@helixcode.com>
4:
1.145 moko 5: Copyright (c) 2001-2024 Art. Lebedev Studio (http://www.artlebedev.com)
1.141 moko 6: Authors: Konstantin Morshnev <moko@design.ru>, Alexandr Petrosian <paf@design.ru>
1.1 paf 7: */
1.13 paf 8:
1.1 paf 9: #include "pa_sapi.h"
10: #include "pa_vmail.h"
11: #include "pa_vstring.h"
12: #include "pa_request.h"
13: #include "pa_common.h"
1.127 moko 14: #include "pa_base64.h"
1.1 paf 15: #include "pa_charset.h"
16: #include "pa_charsets.h"
1.118 moko 17: #include "pa_random.h"
1.1 paf 18: #include "pa_vdate.h"
19: #include "pa_vfile.h"
1.146 moko 20: #include "pa_varray.h"
1.1 paf 21: #include "pa_uue.h"
22:
1.148 ! moko 23: volatile const char * IDENT_PA_VMAIL_C="$Id: pa_vmail.C,v 1.147 2025/05/26 00:52:15 moko Exp $" IDENT_PA_VMAIL_H;
1.98 moko 24:
1.3 paf 25: #ifdef WITH_MAILRECEIVE
1.4 paf 26: extern "C" {
1.99 moko 27: #include "gmime/gmime.h"
1.4 paf 28: }
1.45 paf 29:
30: #include "pa_charsets.h"
1.1 paf 31: #endif
32:
33: // defines
34:
35: #define RAW_NAME "raw"
1.146 moko 36: #define ELEMENTS_NAME "elements"
1.1 paf 37:
38: // internals
39:
40: enum PartType {
41: P_TEXT,
42: P_HTML,
43: P_FILE,
44: P_MESSAGE,
45: P_TYPES_COUNT
46: };
47:
1.99 moko 48: static const char* const part_name_begins[P_TYPES_COUNT] = {
49: "text",
50: "html",
51: "file",
52: "message"
53: };
1.45 paf 54:
55: // defines for statics
56:
57: #define FORMAT_NAME "format"
58: #define CHARSET_NAME "charset"
1.76 misha 59: #define CID_NAME "content-id"
1.45 paf 60:
1.112 moko 61: #define CONTENT_TRANSFER_ENCODING_NAME "content-transfer-encoding"
62: #define CONTENT_TRANSFER_ENCODING_CAPITALIZED "Content-Transfer-Encoding"
63:
1.45 paf 64: // statics
65:
66: static const String format_name(FORMAT_NAME);
67: static const String charset_name(CHARSET_NAME);
1.76 misha 68: static const String cid_name(CID_NAME);
1.1 paf 69:
1.112 moko 70: static const String content_transfer_encoding_name(CONTENT_TRANSFER_ENCODING_NAME);
71:
1.61 paf 72: // consts
73:
1.126 moko 74: const int MAX_CHARS_IN_HEADER_LINE = (991 - 9 /* Subject: */ - 19 /* =?Windows-1251?Q?...?= */ - 128 /* just in case */) / 3 /* quote-printable */ - 5 /* maximum part of trancated UTF-8 char */;
1.61 paf 75:
1.1 paf 76: // VMail
77:
1.116 moko 78: extern Methoded* mail_class;
1.1 paf 79:
1.117 moko 80: VMail::VMail(): VStateless_class(mail_class) {}
1.1 paf 81:
1.3 paf 82: #ifdef WITH_MAILRECEIVE
1.1 paf 83:
1.48 paf 84: #define EXCEPTION_VALUE "x-exception"
85:
1.99 moko 86: static Charset* source_charset;
87:
1.107 moko 88: static const char *transcode(const char *value) {
89: if(value && !source_charset->isUTF8()){
1.121 moko 90: String::C transcoded=Charset::transcode(String::C(value, strlen(value)), pa_UTF8_charset, *source_charset);
1.107 moko 91: value=transcoded.str;
92: }
93: return value;
94: }
95:
1.99 moko 96: static void putReceived(HashStringValue& received, const char* name, Value* value, bool capitalizeName=false) {
97: if(name && value)
98: received.put(capitalizeName ? capitalize(pa_strdup(name)) : pa_strdup(name), value);
1.1 paf 99: }
100:
1.99 moko 101: static void putReceived(HashStringValue& received, const char* name, const char* value, bool capitalizeName=false) {
102: if(name && value)
1.148 ! moko 103: putReceived(received, name, new VString(pa_strdup(value)), capitalizeName);
1.1 paf 104: }
105:
1.99 moko 106: static void putReceived(HashStringValue& received, const char* name, time_t value) {
107: if(name)
108: received.put(pa_strdup(name), new VDate(value) );
1.1 paf 109: }
110:
1.140 moko 111: struct Field2received_info {
112: HashStringValue *hash;
1.146 moko 113: HashStringValue *elements;
1.140 moko 114:
115: Field2received_info(HashStringValue& part){
116: VHash* vhash(new VHash);
1.146 moko 117: VHash* velements(new VHash);
1.140 moko 118: putReceived(part, RAW_NAME, vhash);
1.146 moko 119: putReceived(part, ELEMENTS_NAME, velements);
1.140 moko 120: hash=&vhash->hash();
1.146 moko 121: elements=&velements->hash();
1.140 moko 122: }
123: };
124:
125:
1.99 moko 126: static void MimeHeaderField2received(const char* name, const char* value, gpointer data) {
1.140 moko 127: Field2received_info* info=static_cast<Field2received_info*>(data);
128:
129: if(name && value){
130: name = capitalize(pa_strdup(name));
1.148 ! moko 131: VString *svalue = new VString(pa_strdup(value));
1.140 moko 132:
1.146 moko 133: info->hash->put(name, svalue);
1.140 moko 134:
1.146 moko 135: // elements
136: VArray* varray=(VArray*)(info->elements->get(name));
137: if(!varray)
138: info->elements->put(name, varray=new VArray());
139: varray->array()+=svalue;
1.140 moko 140: }
1.1 paf 141: }
142:
1.99 moko 143: static void parse(Request& r, GMimeMessage *message, HashStringValue& received);
1.1 paf 144:
145: #ifndef DOXYGEN
1.45 paf 146: struct MimePart2body_info {
147: Request* r;
148: HashStringValue* body;
1.1 paf 149: int partCounts[P_TYPES_COUNT];
150: };
151: #endif
1.99 moko 152:
1.128 moko 153: #if GMIME_MAJOR_VERSION > 2
154: typedef void (* GMimeHeaderForeachFunc) (const char *name, const char *value, gpointer user_data);
155:
1.130 moko 156: void g_mime_header_list_foreach (GMimeHeaderList *headers, GMimeHeaderForeachFunc func, gpointer user_data) {
1.128 moko 157: int cnt = g_mime_header_list_get_count(headers);
158: for(int i = 0; i < cnt; i++){
159: GMimeHeader *header = g_mime_header_list_get_header_at(headers, i);
160: func (g_mime_header_get_name(header), g_mime_header_get_value(header), user_data);
161: }
162: }
163:
164: #define g_mime_part_get_content_object(arg) g_mime_part_get_content(arg)
1.129 moko 165: #define g_mime_filter_crlf_new(encode, dots) g_mime_filter_dos2unix_new(encode)
1.128 moko 166:
1.131 moko 167: #define G_MIME_CTYPE_PARAMS(action) { \
168: GMimeParamList *params=g_mime_content_type_get_parameters(type); \
169: int cnt = g_mime_param_list_length(params); \
170: for(int i = 0; i < cnt; i++){ \
171: GMimeParam *param = g_mime_param_list_get_parameter_at(params, i); \
172: action \
1.128 moko 173: }}
174:
175: #else
176:
1.131 moko 177: #define G_MIME_CTYPE_PARAMS(action) { \
178: const GMimeParam *param=g_mime_content_type_get_params(type); \
179: while(param) { \
180: action \
181: param=g_mime_param_next(param); \
1.128 moko 182: }}
183:
1.131 moko 184: #define g_mime_init() g_mime_init(0)
185:
1.129 moko 186: #define g_mime_parser_construct_message(msg,p) g_mime_parser_construct_message(msg)
1.128 moko 187:
188: #endif
1.99 moko 189:
190: static char *readStream(GMimeStream* gstream, size_t &length){
191: length=MAX_STRING;
192: char *result=(char*)pa_malloc_atomic(length+1);
193: char *ptr=result;
194:
195: while(true) {
196: size_t current_size=ptr-result;
197: ssize_t todo_size=length-current_size;
198: ssize_t received_size=g_mime_stream_read (gstream, ptr, todo_size);
199:
200: if(received_size<0)
1.133 moko 201: throw Exception(PARSER_RUNTIME, 0, "mail content stream read error");
1.99 moko 202: if(received_size==0)
203: break;
204: if(received_size==todo_size) {
205: length=length*2;
206: result=(char *)pa_realloc(result, length+1);
207: ptr=result+current_size+received_size;
208: } else {
209: ptr+=received_size;
210: }
211: }
212:
213: length=ptr-result;
214: result[length]='\0';
215: return result;
216: }
217:
218: static void MimePart2body(GMimeObject *parent, GMimeObject *part, gpointer data) {
1.45 paf 219: MimePart2body_info& info=*static_cast<MimePart2body_info *>(data);
1.1 paf 220:
1.99 moko 221: // skipping message/partial & frames
1.101 moko 222: if (GMIME_IS_MESSAGE_PARTIAL (part) || GMIME_IS_MULTIPART (part))
1.99 moko 223: return;
224:
225: if (GMimeContentType *type=g_mime_object_get_content_type(part)) {
226: PartType partType=P_FILE;
227:
1.107 moko 228: if (GMIME_IS_MESSAGE_PART(part)){
1.99 moko 229: partType=P_MESSAGE;
1.107 moko 230: } else {
231: const char *disposition=g_mime_object_get_disposition(part);
232: if(!disposition || strcmp(disposition, GMIME_DISPOSITION_ATTACHMENT)){
233: if(g_mime_content_type_is_type(type, "text", "plain"))
234: partType=P_TEXT;
235: else if(g_mime_content_type_is_type(type, "text", "html"))
236: partType=P_HTML;
237: }
238: }
1.99 moko 239:
1.1 paf 240: // partName
1.99 moko 241: int partNumber=++info.partCounts[partType];
242: const char *partName=part_name_begins[partType];
1.1 paf 243:
1.99 moko 244: char partNameNumbered[MAX_STRING];
245: snprintf(partNameNumbered, MAX_STRING, "%s%d", partName, partNumber);
1.1 paf 246:
1.99 moko 247: // $.partN[
248: VHash* vpartHash(new VHash);
249: if(partNumber==1)
250: putReceived(*info.body, partName, vpartHash);
251: putReceived(*info.body, partNameNumbered, vpartHash);
252:
253: HashStringValue& partHash=vpartHash->hash();
254:
255: // $.raw[
1.140 moko 256: Field2received_info f2r_info(partHash);
257: g_mime_header_list_foreach(part->headers, MimeHeaderField2received, &f2r_info);
1.99 moko 258:
259: // $.content-type[
260: VHash* vcontent_type(new VHash);
261: putReceived(partHash, "content-type", vcontent_type);
262:
263: // $.value[text/plain]
264: char value[MAX_STRING];
265: snprintf(value, MAX_STRING, "%s/%s", type->type ? type->type : "x-unknown", type->subtype ? type->subtype : "x-unknown");
266: putReceived(vcontent_type->hash(), VALUE_NAME, value);
267:
1.128 moko 268: // $.charset[windows-1251] && co
269: G_MIME_CTYPE_PARAMS(
1.107 moko 270: putReceived(vcontent_type->hash(), g_mime_param_get_name(param), transcode(g_mime_param_get_value(param)), true /*capitalizeName*/);
1.128 moko 271: );
1.99 moko 272:
273: if (GMIME_IS_MESSAGE_PART (part)) {
1.102 moko 274: /* message/rfc822, $.raw[] will be overwitten */
1.99 moko 275: GMimeMessage *message = g_mime_message_part_get_message ((GMimeMessagePart *) part);
276: parse(*info.r, message, partHash);
1.1 paf 277: } else {
1.101 moko 278: GMimePart *gpart = (GMimePart *)part;
279:
280: putReceived(partHash, "description", g_mime_part_get_content_description(gpart));
281: putReceived(partHash, "content-id", g_mime_part_get_content_id(gpart));
282: putReceived(partHash, "content-md5", g_mime_part_get_content_md5(gpart));
283: putReceived(partHash, "content-location", g_mime_part_get_content_location(gpart));
284:
1.1 paf 285: // $.value[string|file]
1.103 moko 286: if(GMimeDataWrapper* gcontent=g_mime_part_get_content_object(gpart)){
287: GMimeStream* gstream=g_mime_stream_filter_new(g_mime_data_wrapper_get_stream(gcontent));
288:
289: if(GMimeFilter* filter=g_mime_filter_basic_new(g_mime_part_get_content_encoding(gpart), false))
290: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
291:
292: size_t length;
1.99 moko 293:
1.103 moko 294: if(partType==P_FILE) {
295: char *content=readStream(gstream, length);
1.107 moko 296: const char* content_filename=transcode(g_mime_part_get_filename(gpart));
1.103 moko 297: VFile* vfile(new VFile);
298: vfile->set_binary(true/*tainted*/, content, length, new String(content_filename), content_filename ? new VString(info.r->mime_type_of(content_filename)) : 0);
299: putReceived(partHash, VALUE_NAME, vfile);
300: } else {
301: // P_TEXT, P_HTML
302: if(Value *charset=vcontent_type->hash().get("Charset"))
303: if(GMimeFilter* filter=g_mime_filter_charset_new(charset->get_string()->cstr(), source_charset->NAME_CSTR()))
304: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
305:
306: char *content=readStream(gstream, length);
307: putReceived(partHash, VALUE_NAME,new VString(*new String(content)));
308: }
1.1 paf 309: }
310: }
311: }
312: }
313:
1.99 moko 314: static void parse(Request& r, GMimeMessage *message, HashStringValue& received) {
1.1 paf 315: try {
316: // firstly user-defined strings go
317: // user headers
318: {
319: // $.raw[
1.140 moko 320: Field2received_info f2r_info(received);
321: g_mime_header_list_foreach(g_mime_object_get_header_list(GMIME_OBJECT(message)), MimeHeaderField2received, &f2r_info);
1.99 moko 322: }
323:
324: // secondly standard headers
325: putReceived(received, "message-id", g_mime_message_get_message_id(message));
1.107 moko 326:
1.128 moko 327: #if GMIME_MAJOR_VERSION > 2
328: const char *msg_from=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_FROM), NULL, false);
329: const char *msg_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_TO), NULL, false);
330: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_CC), NULL, false);
331: const char *msg_reply_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_REPLY_TO), NULL, false);
332:
333: GDateTime *gdate = g_mime_message_get_date(message);
334: time_t date = gdate ? g_date_time_to_unix(gdate) : 0;
335: #else
336: const char *msg_from=g_mime_message_get_sender(message);
1.107 moko 337: const char *msg_to=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_TO), false);
1.128 moko 338: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_CC), false);
339: const char *msg_reply_to=g_mime_message_get_reply_to(message);
340:
341: time_t date;
342: g_mime_message_get_date(message, &date, 0);
343: #endif
344: putReceived(received, "from", transcode(msg_from));
1.107 moko 345: putReceived(received, "to", transcode(msg_to));
346: putReceived(received, "cc", transcode(msg_cc));
1.128 moko 347: putReceived(received, "reply-to", transcode(msg_reply_to));
1.107 moko 348: putReceived(received, "subject", transcode(g_mime_message_get_subject(message)));
349:
1.113 moko 350: // .date(time_t in UTC)
1.114 moko 351: putReceived(received, "date", date);
1.1 paf 352:
353: // .body[part/parts
1.55 paf 354: MimePart2body_info info={&r, &received, {0}};
1.99 moko 355: g_mime_message_foreach(message, MimePart2body, &info);
1.1 paf 356:
1.48 paf 357: } catch(const Exception& e) {
1.125 moko 358: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 359: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 360: } catch(...) {
1.125 moko 361: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.1 paf 362: }
363: }
364:
1.99 moko 365: void VMail::fill_received(Request& r) {
1.45 paf 366: if(r.request_info.mail_received) {
1.99 moko 367: source_charset=&r.charsets.source();
1.131 moko 368: g_mime_init();
1.1 paf 369: // create stream with CRLF filter
1.131 moko 370: #if GMIME_MAJOR_VERSION > 2
1.137 moko 371: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_pipe_new(STDIN_FILENO));
1.131 moko 372: #else
1.137 moko 373: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_fs_new(STDIN_FILENO));
1.131 moko 374: #endif
1.137 moko 375: g_mime_stream_filter_add(GMIME_STREAM_FILTER(stream), g_mime_filter_crlf_new(false, false));
1.1 paf 376: try {
1.99 moko 377: // parse incoming message
1.129 moko 378: GMimeMessage *message=g_mime_parser_construct_message(g_mime_parser_new_with_stream(stream), NULL);
1.138 moko 379: if(message){
380: parse(r, message, vreceived.hash());
381: g_object_unref(GMIME_OBJECT(message));
382: }
1.48 paf 383: } catch(const Exception& e) {
384: HashStringValue& received=vreceived.hash();
1.125 moko 385: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 386: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 387: } catch(...) {
1.99 moko 388: // abnormal stream free
389: g_object_unref(stream);
1.48 paf 390: rethrow;
1.1 paf 391: }
1.99 moko 392: g_object_unref(stream);
393:
394: g_mime_shutdown();
1.1 paf 395: }
396: }
397:
1.99 moko 398: #else // WITH_MAILRECEIVE
399: void VMail::fill_received(Request&){}
400: #endif // WITH_MAILRECEIVE
401:
1.9 paf 402: typedef int (*string_contains_char_which_check)(int);
1.99 moko 403:
1.45 paf 404: static bool string_contains_char_which(const char* string, string_contains_char_which_check check) {
1.9 paf 405: while(char c=*string++) {
1.62 paf 406: if(check((unsigned char)c))
1.9 paf 407: return true;
408: }
409: return false;
410: }
1.99 moko 411:
1.32 paf 412: static char *trimBoth(char *s) {
413: // sanity check
414: if(!s)
415: return 0;
416:
417: // trim head whitespace
1.62 paf 418: while(*s && isspace((unsigned char)*s))
1.32 paf 419: s++;
420: // trim tail whitespace
421: char *tail=s+strlen(s);
422: if(tail>s) {
423: do {
424: --tail;
1.62 paf 425: if(isspace((unsigned char)*tail))
1.32 paf 426: *tail=0;
427: } while(tail>s);
428: }
429: // return it
430: return s;
431: }
1.99 moko 432:
1.45 paf 433: static void extractEmail(String& result, char *email) {
1.32 paf 434: email=trimBoth(email);
1.45 paf 435: result.append_help_length(email, 0, String::L_TAINTED);
1.9 paf 436:
437: /*
438: http://www.faqs.org/rfcs/rfc822.html
439:
440: addr-spec = local-part "@" domain ; global address
441:
442: local-part = word *("." word) ; uninterpreted case-preserved
443: word = atom / quoted-string
444:
445: domain = sub-domain *("." sub-domain)
446: sub-domain = domain-ref / domain-literal
447: domain-ref = atom ; symbolic reference
448:
1.92 misha 449: domain-literal << ignoring for now
1.9 paf 450: quoted-string in word << ignoring for now
451:
452: atom = 1*<any CHAR except specials, SPACE and CTLs> << the ONLY to check
453:
454: specials = "(" / ")" / "<" / ">" / "@" ; Must be in quoted-
1.92 misha 455: / "," / ";" / ":" / "\" / <"> ; string, to use
456: / "." / "[" / "]" ; within a word.
1.9 paf 457:
458: */
1.45 paf 459: const char* exception_type="email.format";
1.9 paf 460: if(strpbrk(email, "()<>,;:\\\"[]"/*specials minus @ and . */))
1.126 moko 461: throw Exception(exception_type, &result, "email contains bad characters (specials)");
1.9 paf 462: if(string_contains_char_which(email, (string_contains_char_which_check)isspace))
1.126 moko 463: throw Exception(exception_type, &result, "email contains bad characters (whitespace)");
1.9 paf 464: if(string_contains_char_which(email, (string_contains_char_which_check)iscntrl))
1.126 moko 465: throw Exception(exception_type, &result, "email contains bad characters (control)");
1.16 paf 466: if(result.is_empty())
1.126 moko 467: throw Exception(exception_type, 0, "email is empty");
1.39 paf 468: }
469:
1.45 paf 470: static const String& extractEmails(const String& string) {
471: char *emails=string.cstrm();
472: String& result=*new String;
1.39 paf 473: while(char *email=lsplit(&emails, ',')) {
474: rsplit(email, '>');
475: if(char *in_brackets=lsplit(email, '<'))
476: email=in_brackets;
477: if(!result.is_empty())
478: result<<",";
1.45 paf 479: extractEmail(result, email);
1.39 paf 480: }
1.9 paf 481:
482: return result;
483: }
1.45 paf 484:
485: #ifndef DOXYGEN
486: struct Store_message_element_info {
487: Request_charsets& charsets;
488: String& header;
489: const String* & from;
1.60 paf 490: bool extract_to; String* & to;
1.45 paf 491: bool mime_version_specified;
1.143 moko 492: Array<Value*>* parts[P_TYPES_COUNT];
1.45 paf 493: int parts_count;
494: bool backward_compatibility;
495: Value* content_type;
1.66 paf 496: bool had_content_disposition;
1.45 paf 497:
1.126 moko 498: Store_message_element_info(Request_charsets& acharsets, String& aheader, const String* & afrom, bool aextract_to, String* & ato):
1.45 paf 499: charsets(acharsets),
500: header(aheader),
501: from(afrom),
1.60 paf 502: extract_to(aextract_to), to(ato),
1.45 paf 503: mime_version_specified(false),
504: parts_count(0),
1.66 paf 505: backward_compatibility(false), content_type(0),
1.133 moko 506: had_content_disposition(false)
507: {
508: for(int pt=0; pt<P_TYPES_COUNT; pt++)
1.143 moko 509: parts[pt]=new Array<Value*>(1);
1.45 paf 510: }
511: };
512: #endif
1.99 moko 513:
1.126 moko 514: size_t mail_header_utf8_substring(const char *mail, size_t sub_length, size_t length){
515: int error_offset;
516: if(int error_code=pa_pcre_valid_utf((unsigned char *)mail, sub_length, &error_offset)){
517: if(error_code<PCRE_UTF8_ERR6){ // Missing X byte at the end of the string errors
518: sub_length+=error_code; // adding X bytes
519: return sub_length < length ? sub_length : length;
520: }
521: }
522:
523: return sub_length;
524: }
525:
526: static void store_message_element(HashStringValue::key_type raw_element_name, HashStringValue::value_type element_value, Store_message_element_info *info) {
527: const String& low_element_name=String(raw_element_name, String::L_TAINTED).change_case(info->charsets.source(), String::CC_LOWER);
1.1 paf 528:
529: // exclude internals
1.52 paf 530: if(low_element_name==MAIL_OPTIONS_NAME
531: || low_element_name==CHARSET_NAME
1.5 paf 532: || low_element_name==VALUE_NAME
533: || low_element_name==RAW_NAME
1.146 moko 534: || low_element_name==ELEMENTS_NAME
1.64 paf 535: || low_element_name==FORMAT_NAME
1.76 misha 536: || low_element_name==NAME_NAME
1.94 misha 537: || low_element_name==CID_NAME
538: || low_element_name==MAIL_DEBUG_NAME)
1.1 paf 539: return;
540:
541: // grep parts
542: for(int pt=0; pt<P_TYPES_COUNT; pt++) {
1.45 paf 543: if(low_element_name.starts_with(part_name_begins[pt])) {
1.29 paf 544: // check that $.message# '#' is digit
1.45 paf 545: size_t start_len=strlen(part_name_begins[pt]);
546: if(low_element_name.length()>start_len) {
547: const char* at_num=low_element_name.mid(start_len, start_len+1).cstr();
1.63 paf 548: if(!isdigit((unsigned char)*at_num))
1.29 paf 549: continue;
550: }
1.45 paf 551: *info->parts[pt]+=element_value;
552: info->parts_count++;
1.1 paf 553: return;
554: }
555: }
556:
1.10 paf 557: // fetch some special headers
1.45 paf 558: if(low_element_name=="from")
559: info->from=&extractEmails(element_value->as_string());
1.90 misha 560: if(low_element_name==CONTENT_DISPOSITION)
1.66 paf 561: info->had_content_disposition=true;
1.60 paf 562: if(info->extract_to) { // defined only when SMTP used, see mail.C [collecting info for RCPT to-s]
1.39 paf 563: bool is_to=low_element_name=="to" ;
564: bool is_cc=low_element_name=="cc" ;
565: bool is_bcc=low_element_name=="bcc" ;
566: if(is_to||is_cc||is_bcc) {
1.45 paf 567: if(!info->to)
568: info->to=new String;
1.39 paf 569: else
1.45 paf 570: *info->to << ",";
571: *info->to << extractEmails(element_value->as_string());
1.39 paf 572: }
573:
574: if(is_bcc) // blinding it
1.45 paf 575: return;
1.39 paf 576: }
1.37 paf 577: if(low_element_name=="mime-version")
1.45 paf 578: info->mime_version_specified=true;
1.1 paf 579:
1.45 paf 580: // has content type?
581: if(low_element_name==CONTENT_TYPE_NAME) {
582: info->content_type=element_value;
583: if(info->backward_compatibility)
584: return;
1.39 paf 585: }
1.1 paf 586:
1.45 paf 587: // preparing header line
1.123 moko 588: const String& source_line=attributed_meaning_to_string(*element_value, String::L_AS_IS);
589:
1.66 paf 590: if(source_line.is_empty())
591: return; // we don't need empty headers here [used in clearing content-disposition]
592:
1.45 paf 593: const char* source_line_cstr=source_line.cstr();
1.126 moko 594: String::C mail=Charset::transcode(String::C(source_line_cstr, source_line.length()), info->charsets.source(), info->charsets.mail());
1.92 misha 595:
1.45 paf 596: String& mail_line=*new String;
1.126 moko 597: if(low_element_name=="to" || low_element_name=="cc" || low_element_name=="bcc")
1.73 paf 598: {
1.144 moko 599: // never wrap address lines, mailer cannot handle wrapped properly
1.73 paf 600: mail_line.append_strdup(mail.str, mail.length, String::L_MAIL_HEADER);
601: } else {
602: while(mail.length) {
1.126 moko 603: bool too_long=mail.length > MAX_CHARS_IN_HEADER_LINE;
604: size_t length=too_long ? info->charsets.mail().isUTF8() ? mail_header_utf8_substring(mail.str, MAX_CHARS_IN_HEADER_LINE, mail.length) : MAX_CHARS_IN_HEADER_LINE : mail.length;
1.73 paf 605:
606: mail_line.append_strdup(mail.str, length, String::L_MAIL_HEADER);
1.126 moko 607:
1.73 paf 608: mail.length-=length;
1.126 moko 609: mail.str+=length;
1.73 paf 610:
611: if(too_long)
612: mail_line << "\n "; // break header and continue it on next line
1.126 moko 613: }
1.73 paf 614: }
1.45 paf 615:
616: // append header line
617: info->header
1.91 misha 618: << capitalize(raw_element_name.cstr())
1.96 misha 619: << ": " << mail_line.untaint_cstr(String::L_AS_IS, 0, &info->charsets)
1.45 paf 620: << "\n";
621: }
622:
623: static const String& file_value_to_string(Request& r, Value* send_value) {
1.64 paf 624: String& result=*new String;
625:
1.45 paf 626: VFile* vfile;
627: const String* file_name=0;
628: Value* vformat=0;
1.76 misha 629: Value* vcid=0;
1.66 paf 630: const String* dummy_from;
631: String* dummy_to;
1.95 misha 632: Store_message_element_info info(r.charsets, result, dummy_from, false, dummy_to);
633:
1.120 moko 634: HashStringValue *send_hash=send_value->get_hash();
1.142 moko 635: if(send_hash && !dynamic_cast<VFile*>(send_value)) { // hash
1.74 paf 636: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.64 paf 637:
1.1 paf 638: // $.value
1.45 paf 639: if(Value* value=send_hash->get(value_name))
1.139 moko 640: vfile=value->as_vfile();
1.1 paf 641: else
1.120 moko 642: throw Exception(PARSER_RUNTIME, 0, "file part has no $value");
1.1 paf 643:
644: // $.format
1.45 paf 645: vformat=send_hash->get(format_name);
1.1 paf 646:
1.76 misha 647: // $.content-id
648: vcid=send_hash->get(cid_name);
649:
1.6 paf 650: // $.name
1.45 paf 651: if(Value* vfile_name=send_hash->get(name_name)) // $name specified
1.1 paf 652: file_name=&vfile_name->as_string();
1.28 paf 653: } else // must be VFile then
1.139 moko 654: vfile=send_value->as_vfile();
1.28 paf 655:
656: if(!file_name)
1.45 paf 657: file_name=&vfile->fields().get(name_name)->as_string();
1.28 paf 658:
1.95 misha 659: const char* file_name_cstr;
660: const char* quoted_file_name_cstr;
661: {
662: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.97 misha 663: file_name_cstr=file_name->untaint_and_transcode_cstr(String::L_FILE_SPEC, &charsets);
1.95 misha 664: quoted_file_name_cstr=String(file_name_cstr).taint_cstr(String::L_MAIL_HEADER, 0, &charsets);
665: }
1.1 paf 666:
1.95 misha 667: // Content-Type: application/octet-stream
668: result
669: << HTTP_CONTENT_TYPE_CAPITALIZED ": "
670: << r.mime_type_of(file_name_cstr)
671: << "; name=\""
672: << quoted_file_name_cstr
673: << "\"\n";
1.66 paf 674:
1.95 misha 675: if(!info.had_content_disposition) // $.Content-Disposition wasn't specified by user
1.79 misha 676: result
1.91 misha 677: << CONTENT_DISPOSITION_CAPITALIZED ": "
1.79 misha 678: << ( vcid ? CONTENT_DISPOSITION_INLINE : CONTENT_DISPOSITION_ATTACHMENT )
679: << "; "
1.95 misha 680: << CONTENT_DISPOSITION_FILENAME_NAME"=\"" << quoted_file_name_cstr << "\"\n";
1.1 paf 681:
1.79 misha 682: if(vcid)
1.93 misha 683: result
684: << "Content-Id: <"
685: << vcid->as_string()
686: << ">\n"; // @todo: value must be escaped as %hh
1.79 misha 687:
1.45 paf 688: const String* type=vformat?&vformat->as_string():0;
1.93 misha 689: if(!type/*default*/ || *type=="base64") {
690: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": base64\n\n";
691: result << pa_base64_encode(vfile->value_ptr(), vfile->value_size());
1.75 paf 692: } else {
1.93 misha 693: if(*type=="uue") {
694: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": x-uuencode\n\n";
1.95 misha 695: result << pa_uuencode((const unsigned char*)vfile->value_ptr(), vfile->value_size(), file_name_cstr);
696: } else
1.133 moko 697: throw Exception(PARSER_RUNTIME, type, "unknown attachment encode format");
1.75 paf 698: }
1.95 misha 699:
1.1 paf 700: return result;
701: }
702:
1.133 moko 703: static const String& text_value_to_string(Request& r, PartType pt, Value* send_value, Store_message_element_info& info) {
1.45 paf 704: String& result=*new String;
1.1 paf 705:
1.45 paf 706: Value* text_value;
1.81 misha 707: Value* content_transfer_encoding=0;
1.45 paf 708: if(HashStringValue* send_hash=send_value->get_hash()) {
1.1 paf 709: // $.USER-HEADERS
1.92 misha 710: info.content_type=0;
711: info.backward_compatibility=false; // reset
1.74 paf 712: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.1 paf 713: // $.value
1.45 paf 714: text_value=send_hash->get(value_name);
1.1 paf 715: if(!text_value)
1.133 moko 716: throw Exception(PARSER_RUNTIME, 0, "%s part has no $" VALUE_NAME, part_name_begins[pt]);
1.81 misha 717: content_transfer_encoding=send_hash->get(content_transfer_encoding_name);
1.1 paf 718: } else
1.45 paf 719: text_value=send_value;
1.1 paf 720:
1.45 paf 721: if(!info.content_type) {
722: result
1.91 misha 723: << HTTP_CONTENT_TYPE_CAPITALIZED ": text/" << (pt==P_TEXT?"plain":"html")
1.45 paf 724: << "; charset=" << info.charsets.mail().NAME()
725: << "\n";
1.1 paf 726: }
1.81 misha 727: if(!content_transfer_encoding)
1.91 misha 728: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED << ": 8bit\n";
1.1 paf 729:
730: // header|body separator
731: result << "\n";
732:
733: // body
1.45 paf 734: const String* body;
1.122 moko 735: String::Language body_lang=String::L_AS_IS;
736:
1.1 paf 737: switch(pt) {
738: case P_TEXT:
1.92 misha 739: {
740: body=&text_value->as_string();
741: break;
742: }
1.122 moko 743: case P_HTML:
1.1 paf 744: {
1.122 moko 745: body_lang=String::Language(String::L_HTML | String::L_OPTIMIZE_BIT);
1.55 paf 746: if(text_value->get_junction())
1.41 paf 747: body=&r.process_to_string(*text_value);
1.92 misha 748: else
1.122 moko 749: throw Exception(PARSER_RUNTIME, 0, "html part value must be code");
1.1 paf 750: break;
751: }
1.53 paf 752: default:
1.122 moko 753: throw Exception(0, 0, "unhandled part type #%d", pt);
1.41 paf 754: }
1.122 moko 755:
1.41 paf 756: if(body) {
1.69 paf 757: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.122 moko 758: const char* body_cstr=body->untaint_and_transcode_cstr(body_lang, &charsets);
1.95 misha 759: result.append_know_length(body_cstr, strlen(body_cstr), String::L_CLEAN);
1.1 paf 760: }
761:
762: return result;
1.111 moko 763: }
1.1 paf 764:
765: /// @todo files and messages in order (file, file2, ...)
1.134 moko 766: const String& VMail::message_hash_to_string(Request& r, HashStringValue* message_hash, const String* & from, bool extract_to, String* & to) {
1.34 paf 767:
1.1 paf 768: if(!message_hash)
1.133 moko 769: throw Exception(PARSER_RUNTIME, 0, "message must be hash");
1.1 paf 770:
1.45 paf 771: String& result=*new String;
1.1 paf 772:
1.45 paf 773: if(Value* vrecodecharset_name=message_hash->get(charset_name))
1.121 moko 774: r.charsets.set_mail(pa_charsets.get(vrecodecharset_name->as_string()));
1.1 paf 775: else
1.45 paf 776: r.charsets.set_mail(r.charsets.source());
777: // no big deal that we leave it set. they wont miss this point which would reset it
1.1 paf 778:
1.133 moko 779: Store_message_element_info info(r.charsets, result, from, extract_to, to);
780:
1.1 paf 781: {
1.133 moko 782: // for backward compatibilyty $.body+$.content-type -> $.text[$.value[] $.content-type[]]
1.45 paf 783:
784: Value* body=message_hash->get("body");
785: if(body) {
1.147 moko 786: message_hash->remove(String::Body("body"));
1.45 paf 787: info.backward_compatibility=true;
1.133 moko 788: }
1.74 paf 789: message_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.45 paf 790:
791: if(body) {
792: VHash& text_part=*new VHash();
793: HashStringValue& hash=text_part.hash();
794: hash.put(value_name, body);
795: if(info.content_type)
796: hash.put(content_type_name, info.content_type);
797:
798: *info.parts[P_TEXT]+=&text_part;
799: info.parts_count++;
800: }
801:
1.37 paf 802: if(!info.mime_version_specified)
803: result << "MIME-Version: 1.0\n"; // MIME-Version: default
1.1 paf 804: }
805:
1.45 paf 806: int textCount=info.parts[P_TEXT]->count();
1.1 paf 807: if(textCount>1)
1.136 moko 808: throw Exception(PARSER_RUNTIME, 0, "multiple text parts are not supported, use file part");
1.45 paf 809: int htmlCount=info.parts[P_HTML]->count();
1.1 paf 810: if(htmlCount>1)
1.136 moko 811: throw Exception(PARSER_RUNTIME, 0, "multiple html parts are not supported, use file part");
1.1 paf 812:
813: bool multipart=info.parts_count>1;
814: bool alternative=textCount && htmlCount;
815: // header
816: char *boundary=0;
817: if(multipart) {
1.118 moko 818: boundary=get_uuid_boundary();
1.76 misha 819:
820: bool is_inline = false;
821: {
1.143 moko 822: Array<Value*>& files=*info.parts[P_FILE];
1.76 misha 823: for(size_t i=0; i<files.count(); i++) {
1.82 misha 824: HashStringValue* file;
825: if((file=files.get(i)->get_hash()) && file->get(cid_name)){
1.76 misha 826: is_inline = true;
827: break;
1.77 misha 828: }
1.76 misha 829: }
830: }
1.133 moko 831:
1.91 misha 832: result << HTTP_CONTENT_TYPE_CAPITALIZED ": " << ( is_inline ? HTTP_CONTENT_TYPE_MULTIPART_RELATED : HTTP_CONTENT_TYPE_MULTIPART_MIXED ) << ";";
1.76 misha 833:
1.1 paf 834: // multi-part
1.133 moko 835: result << " boundary=\"" << boundary << "\"\n\nThis is a multi-part message in MIME format.";
1.1 paf 836: }
837:
838: // alternative or not
839: {
840: if(alternative) {
1.45 paf 841: result << "\n\n--" << boundary << "\n" // intermediate boundary
1.91 misha 842: HTTP_CONTENT_TYPE_CAPITALIZED ": multipart/alternative; boundary=\"ALT" << boundary << "\"\n";
1.1 paf 843: }
844: for(int i=0; i<2; i++) {
845: PartType pt=i==0?P_TEXT:P_HTML;
1.45 paf 846: if(info.parts[pt]->count()) {
1.1 paf 847: if(alternative)
848: result << "\n\n--ALT" << boundary << "\n"; // intermediate boundary
849: else if(boundary)
850: result << "\n\n--" << boundary << "\n"; // intermediate boundary
1.45 paf 851: result << text_value_to_string(r, pt, info.parts[pt]->get(0), info);
1.1 paf 852: }
853: }
854: if(alternative)
855: result << "\n\n--ALT" << boundary << "--\n";
856: }
857:
858: // messages
859: {
1.143 moko 860: Array<Value*>& messages=*info.parts[P_MESSAGE];
1.45 paf 861: for(size_t i=0; i<messages.count(); i++) {
1.1 paf 862: if(boundary)
863: result << "\n\n--" << boundary << "\n"; // intermediate boundary
864:
1.45 paf 865: const String* dummy_from;
866: String* dummy_to;
1.134 moko 867: result << message_hash_to_string(r, messages.get(i)->get_hash(), dummy_from, false, dummy_to);
1.1 paf 868: }
869: }
870:
1.132 moko 871: // files go last
872: {
1.143 moko 873: Array<Value*>& files=*info.parts[P_FILE];
1.132 moko 874: for(size_t i=0; i<files.count(); i++) {
875: if(boundary)
876: result << "\n\n--" << boundary << "\n"; // intermediate boundary
877: result << file_value_to_string(r, files.get(i));
878: }
879: }
880:
1.1 paf 881: // tailer
882: if(boundary)
883: result << "\n\n--" << boundary << "--\n"; // finish boundary
884:
885: // return
886: return result;
887: }
888:
889:
1.88 misha 890: Value* VMail::get_element(const String& aname) {
1.1 paf 891: // $fields
1.3 paf 892: #ifdef WITH_MAILRECEIVE
1.1 paf 893: if(aname==MAIL_RECEIVED_ELEMENT_NAME)
1.48 paf 894: return &vreceived;
1.1 paf 895: #endif
896:
1.110 misha 897: // $method
1.88 misha 898: if(Value* result=VStateless_class::get_element(aname))
1.1 paf 899: return result;
900:
1.109 moko 901: return bark("%s field not found", &aname);
1.1 paf 902: }
E-mail:
![[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