Annotation of parser3/src/types/pa_vmail.C, revision 1.137
1.1 paf 1: /** @file
2: Parser: @b mail class.
3: relies on gmime library, by Jeffrey Stedfast <fejj@helixcode.com>
4:
1.135 moko 5: Copyright (c) 2001-2020 Art. Lebedev Studio (http://www.artlebedev.com)
1.1 paf 6: Author: Alexandr Petrosian <paf@design.ru>(http://paf.design.ru)
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"
20: #include "pa_uue.h"
21:
1.137 ! moko 22: volatile const char * IDENT_PA_VMAIL_C="$Id: pa_vmail.C,v 1.136 2020/12/26 23:09:08 moko Exp $" IDENT_PA_VMAIL_H;
1.98 moko 23:
1.3 paf 24: #ifdef WITH_MAILRECEIVE
1.4 paf 25: extern "C" {
1.99 moko 26: #include "gmime/gmime.h"
1.4 paf 27: }
1.45 paf 28:
29: #include "pa_charsets.h"
1.1 paf 30: #endif
31:
32: // defines
33:
34: #define RAW_NAME "raw"
35:
36: // internals
37:
38: enum PartType {
39: P_TEXT,
40: P_HTML,
41: P_FILE,
42: P_MESSAGE,
43: P_TYPES_COUNT
44: };
45:
1.99 moko 46: static const char* const part_name_begins[P_TYPES_COUNT] = {
47: "text",
48: "html",
49: "file",
50: "message"
51: };
1.45 paf 52:
53: // defines for statics
54:
55: #define FORMAT_NAME "format"
56: #define CHARSET_NAME "charset"
1.76 misha 57: #define CID_NAME "content-id"
1.45 paf 58:
1.112 moko 59: #define CONTENT_TRANSFER_ENCODING_NAME "content-transfer-encoding"
60: #define CONTENT_TRANSFER_ENCODING_CAPITALIZED "Content-Transfer-Encoding"
61:
1.45 paf 62: // statics
63:
64: static const String format_name(FORMAT_NAME);
65: static const String charset_name(CHARSET_NAME);
1.76 misha 66: static const String cid_name(CID_NAME);
1.1 paf 67:
1.112 moko 68: static const String content_transfer_encoding_name(CONTENT_TRANSFER_ENCODING_NAME);
69:
1.61 paf 70: // consts
71:
1.126 moko 72: 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 73:
1.1 paf 74: // VMail
75:
1.116 moko 76: extern Methoded* mail_class;
1.1 paf 77:
1.117 moko 78: VMail::VMail(): VStateless_class(mail_class) {}
1.1 paf 79:
1.3 paf 80: #ifdef WITH_MAILRECEIVE
1.1 paf 81:
1.48 paf 82: #define EXCEPTION_VALUE "x-exception"
83:
1.99 moko 84: static Charset* source_charset;
85:
1.107 moko 86: static const char *transcode(const char *value) {
87: if(value && !source_charset->isUTF8()){
1.121 moko 88: String::C transcoded=Charset::transcode(String::C(value, strlen(value)), pa_UTF8_charset, *source_charset);
1.107 moko 89: value=transcoded.str;
90: }
91: return value;
92: }
93:
1.99 moko 94: static void putReceived(HashStringValue& received, const char* name, Value* value, bool capitalizeName=false) {
95: if(name && value)
96: received.put(capitalizeName ? capitalize(pa_strdup(name)) : pa_strdup(name), value);
1.1 paf 97: }
98:
1.99 moko 99: static void putReceived(HashStringValue& received, const char* name, const char* value, bool capitalizeName=false) {
100: if(name && value)
1.108 moko 101: putReceived(received, name, new VString(*new String(pa_strdup(value), String::L_TAINTED)), capitalizeName);
1.1 paf 102: }
103:
1.99 moko 104: static void putReceived(HashStringValue& received, const char* name, time_t value) {
105: if(name)
106: received.put(pa_strdup(name), new VDate(value) );
1.1 paf 107: }
108:
1.99 moko 109: static void MimeHeaderField2received(const char* name, const char* value, gpointer data) {
110: HashStringValue* received=static_cast<HashStringValue*>(data);
111: putReceived(*received, name, value, true /*capitalizeName*/);
1.1 paf 112: }
113:
1.99 moko 114: static void parse(Request& r, GMimeMessage *message, HashStringValue& received);
1.1 paf 115:
116: #ifndef DOXYGEN
1.45 paf 117: struct MimePart2body_info {
118: Request* r;
119: HashStringValue* body;
1.1 paf 120: int partCounts[P_TYPES_COUNT];
121: };
122: #endif
1.99 moko 123:
1.128 moko 124: #if GMIME_MAJOR_VERSION > 2
125: typedef void (* GMimeHeaderForeachFunc) (const char *name, const char *value, gpointer user_data);
126:
1.130 moko 127: void g_mime_header_list_foreach (GMimeHeaderList *headers, GMimeHeaderForeachFunc func, gpointer user_data) {
1.128 moko 128: int cnt = g_mime_header_list_get_count(headers);
129: for(int i = 0; i < cnt; i++){
130: GMimeHeader *header = g_mime_header_list_get_header_at(headers, i);
131: func (g_mime_header_get_name(header), g_mime_header_get_value(header), user_data);
132: }
133: }
134:
135: #define g_mime_part_get_content_object(arg) g_mime_part_get_content(arg)
1.129 moko 136: #define g_mime_filter_crlf_new(encode, dots) g_mime_filter_dos2unix_new(encode)
1.128 moko 137:
1.131 moko 138: #define G_MIME_CTYPE_PARAMS(action) { \
139: GMimeParamList *params=g_mime_content_type_get_parameters(type); \
140: int cnt = g_mime_param_list_length(params); \
141: for(int i = 0; i < cnt; i++){ \
142: GMimeParam *param = g_mime_param_list_get_parameter_at(params, i); \
143: action \
1.128 moko 144: }}
145:
146: #else
147:
1.131 moko 148: #define G_MIME_CTYPE_PARAMS(action) { \
149: const GMimeParam *param=g_mime_content_type_get_params(type); \
150: while(param) { \
151: action \
152: param=g_mime_param_next(param); \
1.128 moko 153: }}
154:
1.131 moko 155: #define g_mime_init() g_mime_init(0)
156:
1.129 moko 157: #define g_mime_parser_construct_message(msg,p) g_mime_parser_construct_message(msg)
1.128 moko 158:
159: #endif
1.99 moko 160:
161: static char *readStream(GMimeStream* gstream, size_t &length){
162: length=MAX_STRING;
163: char *result=(char*)pa_malloc_atomic(length+1);
164: char *ptr=result;
165:
166: while(true) {
167: size_t current_size=ptr-result;
168: ssize_t todo_size=length-current_size;
169: ssize_t received_size=g_mime_stream_read (gstream, ptr, todo_size);
170:
171: if(received_size<0)
1.133 moko 172: throw Exception(PARSER_RUNTIME, 0, "mail content stream read error");
1.99 moko 173: if(received_size==0)
174: break;
175: if(received_size==todo_size) {
176: length=length*2;
177: result=(char *)pa_realloc(result, length+1);
178: ptr=result+current_size+received_size;
179: } else {
180: ptr+=received_size;
181: }
182: }
183:
184: length=ptr-result;
185: result[length]='\0';
186: return result;
187: }
188:
189: static void MimePart2body(GMimeObject *parent, GMimeObject *part, gpointer data) {
1.45 paf 190: MimePart2body_info& info=*static_cast<MimePart2body_info *>(data);
1.1 paf 191:
1.99 moko 192: // skipping message/partial & frames
1.101 moko 193: if (GMIME_IS_MESSAGE_PARTIAL (part) || GMIME_IS_MULTIPART (part))
1.99 moko 194: return;
195:
196: if (GMimeContentType *type=g_mime_object_get_content_type(part)) {
197: PartType partType=P_FILE;
198:
1.107 moko 199: if (GMIME_IS_MESSAGE_PART(part)){
1.99 moko 200: partType=P_MESSAGE;
1.107 moko 201: } else {
202: const char *disposition=g_mime_object_get_disposition(part);
203: if(!disposition || strcmp(disposition, GMIME_DISPOSITION_ATTACHMENT)){
204: if(g_mime_content_type_is_type(type, "text", "plain"))
205: partType=P_TEXT;
206: else if(g_mime_content_type_is_type(type, "text", "html"))
207: partType=P_HTML;
208: }
209: }
1.99 moko 210:
1.1 paf 211: // partName
1.99 moko 212: int partNumber=++info.partCounts[partType];
213: const char *partName=part_name_begins[partType];
1.1 paf 214:
1.99 moko 215: char partNameNumbered[MAX_STRING];
216: snprintf(partNameNumbered, MAX_STRING, "%s%d", partName, partNumber);
1.1 paf 217:
1.99 moko 218: // $.partN[
219: VHash* vpartHash(new VHash);
220: if(partNumber==1)
221: putReceived(*info.body, partName, vpartHash);
222: putReceived(*info.body, partNameNumbered, vpartHash);
223:
224: HashStringValue& partHash=vpartHash->hash();
225:
226: // $.raw[
227: VHash* vraw(new VHash);
228: putReceived(partHash, RAW_NAME, vraw);
229: g_mime_header_list_foreach(part->headers, MimeHeaderField2received, &vraw->hash());
230:
231: // $.content-type[
232: VHash* vcontent_type(new VHash);
233: putReceived(partHash, "content-type", vcontent_type);
234:
235: // $.value[text/plain]
236: char value[MAX_STRING];
237: snprintf(value, MAX_STRING, "%s/%s", type->type ? type->type : "x-unknown", type->subtype ? type->subtype : "x-unknown");
238: putReceived(vcontent_type->hash(), VALUE_NAME, value);
239:
1.128 moko 240: // $.charset[windows-1251] && co
241: G_MIME_CTYPE_PARAMS(
1.107 moko 242: putReceived(vcontent_type->hash(), g_mime_param_get_name(param), transcode(g_mime_param_get_value(param)), true /*capitalizeName*/);
1.128 moko 243: );
1.99 moko 244:
245: if (GMIME_IS_MESSAGE_PART (part)) {
1.102 moko 246: /* message/rfc822, $.raw[] will be overwitten */
1.99 moko 247: GMimeMessage *message = g_mime_message_part_get_message ((GMimeMessagePart *) part);
248: parse(*info.r, message, partHash);
1.1 paf 249: } else {
1.101 moko 250: GMimePart *gpart = (GMimePart *)part;
251:
252: putReceived(partHash, "description", g_mime_part_get_content_description(gpart));
253: putReceived(partHash, "content-id", g_mime_part_get_content_id(gpart));
254: putReceived(partHash, "content-md5", g_mime_part_get_content_md5(gpart));
255: putReceived(partHash, "content-location", g_mime_part_get_content_location(gpart));
256:
1.1 paf 257: // $.value[string|file]
1.103 moko 258: if(GMimeDataWrapper* gcontent=g_mime_part_get_content_object(gpart)){
259: GMimeStream* gstream=g_mime_stream_filter_new(g_mime_data_wrapper_get_stream(gcontent));
260:
261: if(GMimeFilter* filter=g_mime_filter_basic_new(g_mime_part_get_content_encoding(gpart), false))
262: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
263:
264: size_t length;
1.99 moko 265:
1.103 moko 266: if(partType==P_FILE) {
267: char *content=readStream(gstream, length);
1.107 moko 268: const char* content_filename=transcode(g_mime_part_get_filename(gpart));
1.103 moko 269: VFile* vfile(new VFile);
270: vfile->set_binary(true/*tainted*/, content, length, new String(content_filename), content_filename ? new VString(info.r->mime_type_of(content_filename)) : 0);
271: putReceived(partHash, VALUE_NAME, vfile);
272: } else {
273: // P_TEXT, P_HTML
274: if(Value *charset=vcontent_type->hash().get("Charset"))
275: if(GMimeFilter* filter=g_mime_filter_charset_new(charset->get_string()->cstr(), source_charset->NAME_CSTR()))
276: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
277:
278: char *content=readStream(gstream, length);
279: putReceived(partHash, VALUE_NAME,new VString(*new String(content)));
280: }
1.1 paf 281: }
282: }
283: }
284: }
285:
1.99 moko 286: static void parse(Request& r, GMimeMessage *message, HashStringValue& received) {
1.1 paf 287: try {
288: // firstly user-defined strings go
289: // user headers
290: {
291: // $.raw[
1.99 moko 292: VHash* vraw(new VHash); putReceived( received, RAW_NAME, vraw);
293: g_mime_header_list_foreach(g_mime_object_get_header_list(GMIME_OBJECT(message)), MimeHeaderField2received, &vraw->hash());
294: }
295:
296: // secondly standard headers
297: putReceived(received, "message-id", g_mime_message_get_message_id(message));
1.107 moko 298:
1.128 moko 299: #if GMIME_MAJOR_VERSION > 2
300: const char *msg_from=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_FROM), NULL, false);
301: const char *msg_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_TO), NULL, false);
302: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_CC), NULL, false);
303: const char *msg_reply_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_REPLY_TO), NULL, false);
304:
305: GDateTime *gdate = g_mime_message_get_date(message);
306: time_t date = gdate ? g_date_time_to_unix(gdate) : 0;
307: #else
308: const char *msg_from=g_mime_message_get_sender(message);
1.107 moko 309: const char *msg_to=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_TO), false);
1.128 moko 310: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_CC), false);
311: const char *msg_reply_to=g_mime_message_get_reply_to(message);
312:
313: time_t date;
314: g_mime_message_get_date(message, &date, 0);
315: #endif
316: putReceived(received, "from", transcode(msg_from));
1.107 moko 317: putReceived(received, "to", transcode(msg_to));
318: putReceived(received, "cc", transcode(msg_cc));
1.128 moko 319: putReceived(received, "reply-to", transcode(msg_reply_to));
1.107 moko 320: putReceived(received, "subject", transcode(g_mime_message_get_subject(message)));
321:
1.113 moko 322: // .date(time_t in UTC)
1.114 moko 323: putReceived(received, "date", date);
1.1 paf 324:
325: // .body[part/parts
1.55 paf 326: MimePart2body_info info={&r, &received, {0}};
1.99 moko 327: g_mime_message_foreach(message, MimePart2body, &info);
1.1 paf 328:
1.48 paf 329: } catch(const Exception& e) {
1.125 moko 330: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 331: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 332: } catch(...) {
1.125 moko 333: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.1 paf 334: }
335: }
336:
1.99 moko 337: void VMail::fill_received(Request& r) {
1.45 paf 338: if(r.request_info.mail_received) {
1.99 moko 339: source_charset=&r.charsets.source();
1.131 moko 340: g_mime_init();
1.1 paf 341: // create stream with CRLF filter
1.131 moko 342: #if GMIME_MAJOR_VERSION > 2
1.137 ! moko 343: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_pipe_new(STDIN_FILENO));
1.131 moko 344: #else
1.137 ! moko 345: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_fs_new(STDIN_FILENO));
1.131 moko 346: #endif
1.137 ! moko 347: g_mime_stream_filter_add(GMIME_STREAM_FILTER(stream), g_mime_filter_crlf_new(false, false));
1.1 paf 348: try {
1.99 moko 349: // parse incoming message
1.129 moko 350: GMimeMessage *message=g_mime_parser_construct_message(g_mime_parser_new_with_stream(stream), NULL);
1.99 moko 351: parse(r, message, vreceived.hash());
1.101 moko 352: g_object_unref(GMIME_OBJECT(message));
1.48 paf 353: } catch(const Exception& e) {
354: HashStringValue& received=vreceived.hash();
1.125 moko 355: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 356: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 357: } catch(...) {
1.99 moko 358: // abnormal stream free
359: g_object_unref(stream);
1.48 paf 360: rethrow;
1.1 paf 361: }
1.99 moko 362: g_object_unref(stream);
363:
364: g_mime_shutdown();
1.1 paf 365: }
366: }
367:
1.99 moko 368: #else // WITH_MAILRECEIVE
369: void VMail::fill_received(Request&){}
370: #endif // WITH_MAILRECEIVE
371:
1.9 paf 372: typedef int (*string_contains_char_which_check)(int);
1.99 moko 373:
1.45 paf 374: static bool string_contains_char_which(const char* string, string_contains_char_which_check check) {
1.9 paf 375: while(char c=*string++) {
1.62 paf 376: if(check((unsigned char)c))
1.9 paf 377: return true;
378: }
379: return false;
380: }
1.99 moko 381:
1.32 paf 382: static char *trimBoth(char *s) {
383: // sanity check
384: if(!s)
385: return 0;
386:
387: // trim head whitespace
1.62 paf 388: while(*s && isspace((unsigned char)*s))
1.32 paf 389: s++;
390: // trim tail whitespace
391: char *tail=s+strlen(s);
392: if(tail>s) {
393: do {
394: --tail;
1.62 paf 395: if(isspace((unsigned char)*tail))
1.32 paf 396: *tail=0;
397: } while(tail>s);
398: }
399: // return it
400: return s;
401: }
1.99 moko 402:
1.45 paf 403: static void extractEmail(String& result, char *email) {
1.32 paf 404: email=trimBoth(email);
1.45 paf 405: result.append_help_length(email, 0, String::L_TAINTED);
1.9 paf 406:
407: /*
408: http://www.faqs.org/rfcs/rfc822.html
409:
410: addr-spec = local-part "@" domain ; global address
411:
412: local-part = word *("." word) ; uninterpreted case-preserved
413: word = atom / quoted-string
414:
415: domain = sub-domain *("." sub-domain)
416: sub-domain = domain-ref / domain-literal
417: domain-ref = atom ; symbolic reference
418:
1.92 misha 419: domain-literal << ignoring for now
1.9 paf 420: quoted-string in word << ignoring for now
421:
422: atom = 1*<any CHAR except specials, SPACE and CTLs> << the ONLY to check
423:
424: specials = "(" / ")" / "<" / ">" / "@" ; Must be in quoted-
1.92 misha 425: / "," / ";" / ":" / "\" / <"> ; string, to use
426: / "." / "[" / "]" ; within a word.
1.9 paf 427:
428: */
1.45 paf 429: const char* exception_type="email.format";
1.9 paf 430: if(strpbrk(email, "()<>,;:\\\"[]"/*specials minus @ and . */))
1.126 moko 431: throw Exception(exception_type, &result, "email contains bad characters (specials)");
1.9 paf 432: if(string_contains_char_which(email, (string_contains_char_which_check)isspace))
1.126 moko 433: throw Exception(exception_type, &result, "email contains bad characters (whitespace)");
1.9 paf 434: if(string_contains_char_which(email, (string_contains_char_which_check)iscntrl))
1.126 moko 435: throw Exception(exception_type, &result, "email contains bad characters (control)");
1.16 paf 436: if(result.is_empty())
1.126 moko 437: throw Exception(exception_type, 0, "email is empty");
1.39 paf 438: }
439:
1.45 paf 440: static const String& extractEmails(const String& string) {
441: char *emails=string.cstrm();
442: String& result=*new String;
1.39 paf 443: while(char *email=lsplit(&emails, ',')) {
444: rsplit(email, '>');
445: if(char *in_brackets=lsplit(email, '<'))
446: email=in_brackets;
447: if(!result.is_empty())
448: result<<",";
1.45 paf 449: extractEmail(result, email);
1.39 paf 450: }
1.9 paf 451:
452: return result;
453: }
1.45 paf 454:
455: #ifndef DOXYGEN
456: struct Store_message_element_info {
457: Request_charsets& charsets;
458: String& header;
459: const String* & from;
1.60 paf 460: bool extract_to; String* & to;
1.45 paf 461: bool mime_version_specified;
462: ArrayValue* parts[P_TYPES_COUNT];
463: int parts_count;
464: bool backward_compatibility;
465: Value* content_type;
1.66 paf 466: bool had_content_disposition;
1.45 paf 467:
1.126 moko 468: Store_message_element_info(Request_charsets& acharsets, String& aheader, const String* & afrom, bool aextract_to, String* & ato):
1.45 paf 469: charsets(acharsets),
470: header(aheader),
471: from(afrom),
1.60 paf 472: extract_to(aextract_to), to(ato),
1.45 paf 473: mime_version_specified(false),
474: parts_count(0),
1.66 paf 475: backward_compatibility(false), content_type(0),
1.133 moko 476: had_content_disposition(false)
477: {
478: for(int pt=0; pt<P_TYPES_COUNT; pt++)
479: parts[pt]=new ArrayValue(1);
1.45 paf 480: }
481: };
482: #endif
1.99 moko 483:
1.126 moko 484: size_t mail_header_utf8_substring(const char *mail, size_t sub_length, size_t length){
485: int error_offset;
486: if(int error_code=pa_pcre_valid_utf((unsigned char *)mail, sub_length, &error_offset)){
487: if(error_code<PCRE_UTF8_ERR6){ // Missing X byte at the end of the string errors
488: sub_length+=error_code; // adding X bytes
489: return sub_length < length ? sub_length : length;
490: }
491: }
492:
493: return sub_length;
494: }
495:
496: static void store_message_element(HashStringValue::key_type raw_element_name, HashStringValue::value_type element_value, Store_message_element_info *info) {
497: const String& low_element_name=String(raw_element_name, String::L_TAINTED).change_case(info->charsets.source(), String::CC_LOWER);
1.1 paf 498:
499: // exclude internals
1.52 paf 500: if(low_element_name==MAIL_OPTIONS_NAME
501: || low_element_name==CHARSET_NAME
1.5 paf 502: || low_element_name==VALUE_NAME
503: || low_element_name==RAW_NAME
1.64 paf 504: || low_element_name==FORMAT_NAME
1.76 misha 505: || low_element_name==NAME_NAME
1.94 misha 506: || low_element_name==CID_NAME
507: || low_element_name==MAIL_DEBUG_NAME)
1.1 paf 508: return;
509:
510: // grep parts
511: for(int pt=0; pt<P_TYPES_COUNT; pt++) {
1.45 paf 512: if(low_element_name.starts_with(part_name_begins[pt])) {
1.29 paf 513: // check that $.message# '#' is digit
1.45 paf 514: size_t start_len=strlen(part_name_begins[pt]);
515: if(low_element_name.length()>start_len) {
516: const char* at_num=low_element_name.mid(start_len, start_len+1).cstr();
1.63 paf 517: if(!isdigit((unsigned char)*at_num))
1.29 paf 518: continue;
519: }
1.45 paf 520: *info->parts[pt]+=element_value;
521: info->parts_count++;
1.1 paf 522: return;
523: }
524: }
525:
1.10 paf 526: // fetch some special headers
1.45 paf 527: if(low_element_name=="from")
528: info->from=&extractEmails(element_value->as_string());
1.90 misha 529: if(low_element_name==CONTENT_DISPOSITION)
1.66 paf 530: info->had_content_disposition=true;
1.60 paf 531: if(info->extract_to) { // defined only when SMTP used, see mail.C [collecting info for RCPT to-s]
1.39 paf 532: bool is_to=low_element_name=="to" ;
533: bool is_cc=low_element_name=="cc" ;
534: bool is_bcc=low_element_name=="bcc" ;
535: if(is_to||is_cc||is_bcc) {
1.45 paf 536: if(!info->to)
537: info->to=new String;
1.39 paf 538: else
1.45 paf 539: *info->to << ",";
540: *info->to << extractEmails(element_value->as_string());
1.39 paf 541: }
542:
543: if(is_bcc) // blinding it
1.45 paf 544: return;
1.39 paf 545: }
1.37 paf 546: if(low_element_name=="mime-version")
1.45 paf 547: info->mime_version_specified=true;
1.1 paf 548:
1.45 paf 549: // has content type?
550: if(low_element_name==CONTENT_TYPE_NAME) {
551: info->content_type=element_value;
552: if(info->backward_compatibility)
553: return;
1.39 paf 554: }
1.1 paf 555:
1.45 paf 556: // preparing header line
1.123 moko 557: const String& source_line=attributed_meaning_to_string(*element_value, String::L_AS_IS);
558:
1.66 paf 559: if(source_line.is_empty())
560: return; // we don't need empty headers here [used in clearing content-disposition]
561:
1.45 paf 562: const char* source_line_cstr=source_line.cstr();
1.126 moko 563: String::C mail=Charset::transcode(String::C(source_line_cstr, source_line.length()), info->charsets.source(), info->charsets.mail());
1.92 misha 564:
1.45 paf 565: String& mail_line=*new String;
1.126 moko 566: if(low_element_name=="to" || low_element_name=="cc" || low_element_name=="bcc")
1.73 paf 567: {
568: // never wrap address lines, mailer can not handle wrapped properly
569: mail_line.append_strdup(mail.str, mail.length, String::L_MAIL_HEADER);
570: } else {
571: while(mail.length) {
1.126 moko 572: bool too_long=mail.length > MAX_CHARS_IN_HEADER_LINE;
573: 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 574:
575: mail_line.append_strdup(mail.str, length, String::L_MAIL_HEADER);
1.126 moko 576:
1.73 paf 577: mail.length-=length;
1.126 moko 578: mail.str+=length;
1.73 paf 579:
580: if(too_long)
581: mail_line << "\n "; // break header and continue it on next line
1.126 moko 582: }
1.73 paf 583: }
1.45 paf 584:
585: // append header line
586: info->header
1.91 misha 587: << capitalize(raw_element_name.cstr())
1.96 misha 588: << ": " << mail_line.untaint_cstr(String::L_AS_IS, 0, &info->charsets)
1.45 paf 589: << "\n";
590: }
591:
592: static const String& file_value_to_string(Request& r, Value* send_value) {
1.64 paf 593: String& result=*new String;
594:
1.45 paf 595: VFile* vfile;
596: const String* file_name=0;
597: Value* vformat=0;
1.76 misha 598: Value* vcid=0;
1.66 paf 599: const String* dummy_from;
600: String* dummy_to;
1.95 misha 601: Store_message_element_info info(r.charsets, result, dummy_from, false, dummy_to);
602:
1.120 moko 603: HashStringValue *send_hash=send_value->get_hash();
604: if(send_hash && !send_value->as("file")) { // hash
1.74 paf 605: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.64 paf 606:
1.1 paf 607: // $.value
1.45 paf 608: if(Value* value=send_hash->get(value_name))
1.85 misha 609: vfile=value->as_vfile(String::L_AS_IS);
1.1 paf 610: else
1.120 moko 611: throw Exception(PARSER_RUNTIME, 0, "file part has no $value");
1.1 paf 612:
613: // $.format
1.45 paf 614: vformat=send_hash->get(format_name);
1.1 paf 615:
1.76 misha 616: // $.content-id
617: vcid=send_hash->get(cid_name);
618:
1.6 paf 619: // $.name
1.45 paf 620: if(Value* vfile_name=send_hash->get(name_name)) // $name specified
1.1 paf 621: file_name=&vfile_name->as_string();
1.28 paf 622: } else // must be VFile then
1.45 paf 623: vfile=send_value->as_vfile(String::L_AS_IS);
1.28 paf 624:
625: if(!file_name)
1.45 paf 626: file_name=&vfile->fields().get(name_name)->as_string();
1.28 paf 627:
1.95 misha 628: const char* file_name_cstr;
629: const char* quoted_file_name_cstr;
630: {
631: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.97 misha 632: file_name_cstr=file_name->untaint_and_transcode_cstr(String::L_FILE_SPEC, &charsets);
1.95 misha 633: quoted_file_name_cstr=String(file_name_cstr).taint_cstr(String::L_MAIL_HEADER, 0, &charsets);
634: }
1.1 paf 635:
1.95 misha 636: // Content-Type: application/octet-stream
637: result
638: << HTTP_CONTENT_TYPE_CAPITALIZED ": "
639: << r.mime_type_of(file_name_cstr)
640: << "; name=\""
641: << quoted_file_name_cstr
642: << "\"\n";
1.66 paf 643:
1.95 misha 644: if(!info.had_content_disposition) // $.Content-Disposition wasn't specified by user
1.79 misha 645: result
1.91 misha 646: << CONTENT_DISPOSITION_CAPITALIZED ": "
1.79 misha 647: << ( vcid ? CONTENT_DISPOSITION_INLINE : CONTENT_DISPOSITION_ATTACHMENT )
648: << "; "
1.95 misha 649: << CONTENT_DISPOSITION_FILENAME_NAME"=\"" << quoted_file_name_cstr << "\"\n";
1.1 paf 650:
1.79 misha 651: if(vcid)
1.93 misha 652: result
653: << "Content-Id: <"
654: << vcid->as_string()
655: << ">\n"; // @todo: value must be escaped as %hh
1.79 misha 656:
1.45 paf 657: const String* type=vformat?&vformat->as_string():0;
1.93 misha 658: if(!type/*default*/ || *type=="base64") {
659: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": base64\n\n";
660: result << pa_base64_encode(vfile->value_ptr(), vfile->value_size());
1.75 paf 661: } else {
1.93 misha 662: if(*type=="uue") {
663: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": x-uuencode\n\n";
1.95 misha 664: result << pa_uuencode((const unsigned char*)vfile->value_ptr(), vfile->value_size(), file_name_cstr);
665: } else
1.133 moko 666: throw Exception(PARSER_RUNTIME, type, "unknown attachment encode format");
1.75 paf 667: }
1.95 misha 668:
1.1 paf 669: return result;
670: }
671:
1.133 moko 672: static const String& text_value_to_string(Request& r, PartType pt, Value* send_value, Store_message_element_info& info) {
1.45 paf 673: String& result=*new String;
1.1 paf 674:
1.45 paf 675: Value* text_value;
1.81 misha 676: Value* content_transfer_encoding=0;
1.45 paf 677: if(HashStringValue* send_hash=send_value->get_hash()) {
1.1 paf 678: // $.USER-HEADERS
1.92 misha 679: info.content_type=0;
680: info.backward_compatibility=false; // reset
1.74 paf 681: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.1 paf 682: // $.value
1.45 paf 683: text_value=send_hash->get(value_name);
1.1 paf 684: if(!text_value)
1.133 moko 685: throw Exception(PARSER_RUNTIME, 0, "%s part has no $" VALUE_NAME, part_name_begins[pt]);
1.81 misha 686: content_transfer_encoding=send_hash->get(content_transfer_encoding_name);
1.1 paf 687: } else
1.45 paf 688: text_value=send_value;
1.1 paf 689:
1.45 paf 690: if(!info.content_type) {
691: result
1.91 misha 692: << HTTP_CONTENT_TYPE_CAPITALIZED ": text/" << (pt==P_TEXT?"plain":"html")
1.45 paf 693: << "; charset=" << info.charsets.mail().NAME()
694: << "\n";
1.1 paf 695: }
1.81 misha 696: if(!content_transfer_encoding)
1.91 misha 697: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED << ": 8bit\n";
1.1 paf 698:
699: // header|body separator
700: result << "\n";
701:
702: // body
1.45 paf 703: const String* body;
1.122 moko 704: String::Language body_lang=String::L_AS_IS;
705:
1.1 paf 706: switch(pt) {
707: case P_TEXT:
1.92 misha 708: {
709: body=&text_value->as_string();
710: break;
711: }
1.122 moko 712: case P_HTML:
1.1 paf 713: {
1.122 moko 714: body_lang=String::Language(String::L_HTML | String::L_OPTIMIZE_BIT);
1.55 paf 715: if(text_value->get_junction())
1.41 paf 716: body=&r.process_to_string(*text_value);
1.92 misha 717: else
1.122 moko 718: throw Exception(PARSER_RUNTIME, 0, "html part value must be code");
1.1 paf 719: break;
720: }
1.53 paf 721: default:
1.122 moko 722: throw Exception(0, 0, "unhandled part type #%d", pt);
1.41 paf 723: }
1.122 moko 724:
1.41 paf 725: if(body) {
1.69 paf 726: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.122 moko 727: const char* body_cstr=body->untaint_and_transcode_cstr(body_lang, &charsets);
1.95 misha 728: result.append_know_length(body_cstr, strlen(body_cstr), String::L_CLEAN);
1.1 paf 729: }
730:
731: return result;
1.111 moko 732: }
1.1 paf 733:
734: /// @todo files and messages in order (file, file2, ...)
1.134 moko 735: const String& VMail::message_hash_to_string(Request& r, HashStringValue* message_hash, const String* & from, bool extract_to, String* & to) {
1.34 paf 736:
1.1 paf 737: if(!message_hash)
1.133 moko 738: throw Exception(PARSER_RUNTIME, 0, "message must be hash");
1.1 paf 739:
1.45 paf 740: String& result=*new String;
1.1 paf 741:
1.45 paf 742: if(Value* vrecodecharset_name=message_hash->get(charset_name))
1.121 moko 743: r.charsets.set_mail(pa_charsets.get(vrecodecharset_name->as_string()));
1.1 paf 744: else
1.45 paf 745: r.charsets.set_mail(r.charsets.source());
746: // no big deal that we leave it set. they wont miss this point which would reset it
1.1 paf 747:
1.133 moko 748: Store_message_element_info info(r.charsets, result, from, extract_to, to);
749:
1.1 paf 750: {
1.133 moko 751: // for backward compatibilyty $.body+$.content-type -> $.text[$.value[] $.content-type[]]
1.45 paf 752:
753: Value* body=message_hash->get("body");
754: if(body) {
755: message_hash->remove("body");
756: info.backward_compatibility=true;
1.133 moko 757: }
1.74 paf 758: message_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.45 paf 759:
760: if(body) {
761: VHash& text_part=*new VHash();
762: HashStringValue& hash=text_part.hash();
763: hash.put(value_name, body);
764: if(info.content_type)
765: hash.put(content_type_name, info.content_type);
766:
767: *info.parts[P_TEXT]+=&text_part;
768: info.parts_count++;
769: }
770:
1.37 paf 771: if(!info.mime_version_specified)
772: result << "MIME-Version: 1.0\n"; // MIME-Version: default
1.1 paf 773: }
774:
1.45 paf 775: int textCount=info.parts[P_TEXT]->count();
1.1 paf 776: if(textCount>1)
1.136 moko 777: throw Exception(PARSER_RUNTIME, 0, "multiple text parts are not supported, use file part");
1.45 paf 778: int htmlCount=info.parts[P_HTML]->count();
1.1 paf 779: if(htmlCount>1)
1.136 moko 780: throw Exception(PARSER_RUNTIME, 0, "multiple html parts are not supported, use file part");
1.1 paf 781:
782: bool multipart=info.parts_count>1;
783: bool alternative=textCount && htmlCount;
784: // header
785: char *boundary=0;
786: if(multipart) {
1.118 moko 787: boundary=get_uuid_boundary();
1.76 misha 788:
789: bool is_inline = false;
790: {
791: ArrayValue& files=*info.parts[P_FILE];
792: for(size_t i=0; i<files.count(); i++) {
1.82 misha 793: HashStringValue* file;
794: if((file=files.get(i)->get_hash()) && file->get(cid_name)){
1.76 misha 795: is_inline = true;
796: break;
1.77 misha 797: }
1.76 misha 798: }
799: }
1.133 moko 800:
1.91 misha 801: result << HTTP_CONTENT_TYPE_CAPITALIZED ": " << ( is_inline ? HTTP_CONTENT_TYPE_MULTIPART_RELATED : HTTP_CONTENT_TYPE_MULTIPART_MIXED ) << ";";
1.76 misha 802:
1.1 paf 803: // multi-part
1.133 moko 804: result << " boundary=\"" << boundary << "\"\n\nThis is a multi-part message in MIME format.";
1.1 paf 805: }
806:
807: // alternative or not
808: {
809: if(alternative) {
1.45 paf 810: result << "\n\n--" << boundary << "\n" // intermediate boundary
1.91 misha 811: HTTP_CONTENT_TYPE_CAPITALIZED ": multipart/alternative; boundary=\"ALT" << boundary << "\"\n";
1.1 paf 812: }
813: for(int i=0; i<2; i++) {
814: PartType pt=i==0?P_TEXT:P_HTML;
1.45 paf 815: if(info.parts[pt]->count()) {
1.1 paf 816: if(alternative)
817: result << "\n\n--ALT" << boundary << "\n"; // intermediate boundary
818: else if(boundary)
819: result << "\n\n--" << boundary << "\n"; // intermediate boundary
1.45 paf 820: result << text_value_to_string(r, pt, info.parts[pt]->get(0), info);
1.1 paf 821: }
822: }
823: if(alternative)
824: result << "\n\n--ALT" << boundary << "--\n";
825: }
826:
827: // messages
828: {
1.45 paf 829: ArrayValue& messages=*info.parts[P_MESSAGE];
830: for(size_t i=0; i<messages.count(); i++) {
1.1 paf 831: if(boundary)
832: result << "\n\n--" << boundary << "\n"; // intermediate boundary
833:
1.45 paf 834: const String* dummy_from;
835: String* dummy_to;
1.134 moko 836: result << message_hash_to_string(r, messages.get(i)->get_hash(), dummy_from, false, dummy_to);
1.1 paf 837: }
838: }
839:
1.132 moko 840: // files go last
841: {
842: ArrayValue& files=*info.parts[P_FILE];
843: for(size_t i=0; i<files.count(); i++) {
844: if(boundary)
845: result << "\n\n--" << boundary << "\n"; // intermediate boundary
846: result << file_value_to_string(r, files.get(i));
847: }
848: }
849:
1.1 paf 850: // tailer
851: if(boundary)
852: result << "\n\n--" << boundary << "--\n"; // finish boundary
853:
854: // return
855: return result;
856: }
857:
858:
1.88 misha 859: Value* VMail::get_element(const String& aname) {
1.1 paf 860: // $fields
1.3 paf 861: #ifdef WITH_MAILRECEIVE
1.1 paf 862: if(aname==MAIL_RECEIVED_ELEMENT_NAME)
1.48 paf 863: return &vreceived;
1.1 paf 864: #endif
865:
1.110 misha 866: // $method
1.88 misha 867: if(Value* result=VStateless_class::get_element(aname))
1.1 paf 868: return result;
869:
1.109 moko 870: return bark("%s field not found", &aname);
1.1 paf 871: }
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