Annotation of parser3/src/types/pa_vmail.C, revision 1.140
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"
1.140 ! moko 20: #include "pa_vtable.h"
1.1 paf 21: #include "pa_uue.h"
22:
1.140 ! moko 23: volatile const char * IDENT_PA_VMAIL_C="$Id: pa_vmail.C,v 1.139 2021/10/19 16:16:36 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"
1.140 ! moko 36: #define TABLES_NAME "tables"
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)
1.108 moko 103: putReceived(received, name, new VString(*new String(pa_strdup(value), String::L_TAINTED)), 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;
! 113: HashStringValue *tables;
! 114:
! 115: Field2received_info(HashStringValue& part){
! 116: VHash* vhash(new VHash);
! 117: VHash* vtables(new VHash);
! 118: putReceived(part, RAW_NAME, vhash);
! 119: putReceived(part, TABLES_NAME, vtables);
! 120: hash=&vhash->hash();
! 121: tables=&vtables->hash();
! 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));
! 131: String *svalue = new String(pa_strdup(value), String::L_TAINTED);
! 132:
! 133: info->hash->put(name, new VString(*svalue));
! 134:
! 135: // tables
! 136: Value* vtable=info->tables->get(name);
! 137: if(!vtable) {
! 138: // first appearence
! 139: Table::columns_type columns(new ArrayString(1));
! 140: *columns+=new String("field");
! 141:
! 142: vtable=new VTable(new Table(columns));
! 143: info->tables->put(name, vtable);
! 144: }
! 145: Table& table=*vtable->get_table();
! 146:
! 147: // this string becomes next row
! 148: Table::element_type row(new ArrayString(1));
! 149: *row+=svalue;
! 150: table+=row;
! 151: }
1.1 paf 152: }
153:
1.99 moko 154: static void parse(Request& r, GMimeMessage *message, HashStringValue& received);
1.1 paf 155:
156: #ifndef DOXYGEN
1.45 paf 157: struct MimePart2body_info {
158: Request* r;
159: HashStringValue* body;
1.1 paf 160: int partCounts[P_TYPES_COUNT];
161: };
162: #endif
1.99 moko 163:
1.128 moko 164: #if GMIME_MAJOR_VERSION > 2
165: typedef void (* GMimeHeaderForeachFunc) (const char *name, const char *value, gpointer user_data);
166:
1.130 moko 167: void g_mime_header_list_foreach (GMimeHeaderList *headers, GMimeHeaderForeachFunc func, gpointer user_data) {
1.128 moko 168: int cnt = g_mime_header_list_get_count(headers);
169: for(int i = 0; i < cnt; i++){
170: GMimeHeader *header = g_mime_header_list_get_header_at(headers, i);
171: func (g_mime_header_get_name(header), g_mime_header_get_value(header), user_data);
172: }
173: }
174:
175: #define g_mime_part_get_content_object(arg) g_mime_part_get_content(arg)
1.129 moko 176: #define g_mime_filter_crlf_new(encode, dots) g_mime_filter_dos2unix_new(encode)
1.128 moko 177:
1.131 moko 178: #define G_MIME_CTYPE_PARAMS(action) { \
179: GMimeParamList *params=g_mime_content_type_get_parameters(type); \
180: int cnt = g_mime_param_list_length(params); \
181: for(int i = 0; i < cnt; i++){ \
182: GMimeParam *param = g_mime_param_list_get_parameter_at(params, i); \
183: action \
1.128 moko 184: }}
185:
186: #else
187:
1.131 moko 188: #define G_MIME_CTYPE_PARAMS(action) { \
189: const GMimeParam *param=g_mime_content_type_get_params(type); \
190: while(param) { \
191: action \
192: param=g_mime_param_next(param); \
1.128 moko 193: }}
194:
1.131 moko 195: #define g_mime_init() g_mime_init(0)
196:
1.129 moko 197: #define g_mime_parser_construct_message(msg,p) g_mime_parser_construct_message(msg)
1.128 moko 198:
199: #endif
1.99 moko 200:
201: static char *readStream(GMimeStream* gstream, size_t &length){
202: length=MAX_STRING;
203: char *result=(char*)pa_malloc_atomic(length+1);
204: char *ptr=result;
205:
206: while(true) {
207: size_t current_size=ptr-result;
208: ssize_t todo_size=length-current_size;
209: ssize_t received_size=g_mime_stream_read (gstream, ptr, todo_size);
210:
211: if(received_size<0)
1.133 moko 212: throw Exception(PARSER_RUNTIME, 0, "mail content stream read error");
1.99 moko 213: if(received_size==0)
214: break;
215: if(received_size==todo_size) {
216: length=length*2;
217: result=(char *)pa_realloc(result, length+1);
218: ptr=result+current_size+received_size;
219: } else {
220: ptr+=received_size;
221: }
222: }
223:
224: length=ptr-result;
225: result[length]='\0';
226: return result;
227: }
228:
229: static void MimePart2body(GMimeObject *parent, GMimeObject *part, gpointer data) {
1.45 paf 230: MimePart2body_info& info=*static_cast<MimePart2body_info *>(data);
1.1 paf 231:
1.99 moko 232: // skipping message/partial & frames
1.101 moko 233: if (GMIME_IS_MESSAGE_PARTIAL (part) || GMIME_IS_MULTIPART (part))
1.99 moko 234: return;
235:
236: if (GMimeContentType *type=g_mime_object_get_content_type(part)) {
237: PartType partType=P_FILE;
238:
1.107 moko 239: if (GMIME_IS_MESSAGE_PART(part)){
1.99 moko 240: partType=P_MESSAGE;
1.107 moko 241: } else {
242: const char *disposition=g_mime_object_get_disposition(part);
243: if(!disposition || strcmp(disposition, GMIME_DISPOSITION_ATTACHMENT)){
244: if(g_mime_content_type_is_type(type, "text", "plain"))
245: partType=P_TEXT;
246: else if(g_mime_content_type_is_type(type, "text", "html"))
247: partType=P_HTML;
248: }
249: }
1.99 moko 250:
1.1 paf 251: // partName
1.99 moko 252: int partNumber=++info.partCounts[partType];
253: const char *partName=part_name_begins[partType];
1.1 paf 254:
1.99 moko 255: char partNameNumbered[MAX_STRING];
256: snprintf(partNameNumbered, MAX_STRING, "%s%d", partName, partNumber);
1.1 paf 257:
1.99 moko 258: // $.partN[
259: VHash* vpartHash(new VHash);
260: if(partNumber==1)
261: putReceived(*info.body, partName, vpartHash);
262: putReceived(*info.body, partNameNumbered, vpartHash);
263:
264: HashStringValue& partHash=vpartHash->hash();
265:
266: // $.raw[
1.140 ! moko 267: Field2received_info f2r_info(partHash);
! 268: g_mime_header_list_foreach(part->headers, MimeHeaderField2received, &f2r_info);
1.99 moko 269:
270: // $.content-type[
271: VHash* vcontent_type(new VHash);
272: putReceived(partHash, "content-type", vcontent_type);
273:
274: // $.value[text/plain]
275: char value[MAX_STRING];
276: snprintf(value, MAX_STRING, "%s/%s", type->type ? type->type : "x-unknown", type->subtype ? type->subtype : "x-unknown");
277: putReceived(vcontent_type->hash(), VALUE_NAME, value);
278:
1.128 moko 279: // $.charset[windows-1251] && co
280: G_MIME_CTYPE_PARAMS(
1.107 moko 281: putReceived(vcontent_type->hash(), g_mime_param_get_name(param), transcode(g_mime_param_get_value(param)), true /*capitalizeName*/);
1.128 moko 282: );
1.99 moko 283:
284: if (GMIME_IS_MESSAGE_PART (part)) {
1.102 moko 285: /* message/rfc822, $.raw[] will be overwitten */
1.99 moko 286: GMimeMessage *message = g_mime_message_part_get_message ((GMimeMessagePart *) part);
287: parse(*info.r, message, partHash);
1.1 paf 288: } else {
1.101 moko 289: GMimePart *gpart = (GMimePart *)part;
290:
291: putReceived(partHash, "description", g_mime_part_get_content_description(gpart));
292: putReceived(partHash, "content-id", g_mime_part_get_content_id(gpart));
293: putReceived(partHash, "content-md5", g_mime_part_get_content_md5(gpart));
294: putReceived(partHash, "content-location", g_mime_part_get_content_location(gpart));
295:
1.1 paf 296: // $.value[string|file]
1.103 moko 297: if(GMimeDataWrapper* gcontent=g_mime_part_get_content_object(gpart)){
298: GMimeStream* gstream=g_mime_stream_filter_new(g_mime_data_wrapper_get_stream(gcontent));
299:
300: if(GMimeFilter* filter=g_mime_filter_basic_new(g_mime_part_get_content_encoding(gpart), false))
301: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
302:
303: size_t length;
1.99 moko 304:
1.103 moko 305: if(partType==P_FILE) {
306: char *content=readStream(gstream, length);
1.107 moko 307: const char* content_filename=transcode(g_mime_part_get_filename(gpart));
1.103 moko 308: VFile* vfile(new VFile);
309: vfile->set_binary(true/*tainted*/, content, length, new String(content_filename), content_filename ? new VString(info.r->mime_type_of(content_filename)) : 0);
310: putReceived(partHash, VALUE_NAME, vfile);
311: } else {
312: // P_TEXT, P_HTML
313: if(Value *charset=vcontent_type->hash().get("Charset"))
314: if(GMimeFilter* filter=g_mime_filter_charset_new(charset->get_string()->cstr(), source_charset->NAME_CSTR()))
315: g_mime_stream_filter_add(GMIME_STREAM_FILTER(gstream), filter);
316:
317: char *content=readStream(gstream, length);
318: putReceived(partHash, VALUE_NAME,new VString(*new String(content)));
319: }
1.1 paf 320: }
321: }
322: }
323: }
324:
1.99 moko 325: static void parse(Request& r, GMimeMessage *message, HashStringValue& received) {
1.1 paf 326: try {
327: // firstly user-defined strings go
328: // user headers
329: {
330: // $.raw[
1.140 ! moko 331: Field2received_info f2r_info(received);
! 332: g_mime_header_list_foreach(g_mime_object_get_header_list(GMIME_OBJECT(message)), MimeHeaderField2received, &f2r_info);
1.99 moko 333: }
334:
335: // secondly standard headers
336: putReceived(received, "message-id", g_mime_message_get_message_id(message));
1.107 moko 337:
1.128 moko 338: #if GMIME_MAJOR_VERSION > 2
339: const char *msg_from=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_FROM), NULL, false);
340: const char *msg_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_TO), NULL, false);
341: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_CC), NULL, false);
342: const char *msg_reply_to=internet_address_list_to_string(g_mime_message_get_addresses (message, GMIME_ADDRESS_TYPE_REPLY_TO), NULL, false);
343:
344: GDateTime *gdate = g_mime_message_get_date(message);
345: time_t date = gdate ? g_date_time_to_unix(gdate) : 0;
346: #else
347: const char *msg_from=g_mime_message_get_sender(message);
1.107 moko 348: const char *msg_to=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_TO), false);
1.128 moko 349: const char *msg_cc=internet_address_list_to_string(g_mime_message_get_recipients(message, GMIME_RECIPIENT_TYPE_CC), false);
350: const char *msg_reply_to=g_mime_message_get_reply_to(message);
351:
352: time_t date;
353: g_mime_message_get_date(message, &date, 0);
354: #endif
355: putReceived(received, "from", transcode(msg_from));
1.107 moko 356: putReceived(received, "to", transcode(msg_to));
357: putReceived(received, "cc", transcode(msg_cc));
1.128 moko 358: putReceived(received, "reply-to", transcode(msg_reply_to));
1.107 moko 359: putReceived(received, "subject", transcode(g_mime_message_get_subject(message)));
360:
1.113 moko 361: // .date(time_t in UTC)
1.114 moko 362: putReceived(received, "date", date);
1.1 paf 363:
364: // .body[part/parts
1.55 paf 365: MimePart2body_info info={&r, &received, {0}};
1.99 moko 366: g_mime_message_foreach(message, MimePart2body, &info);
1.1 paf 367:
1.48 paf 368: } catch(const Exception& e) {
1.125 moko 369: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 370: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 371: } catch(...) {
1.125 moko 372: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.1 paf 373: }
374: }
375:
1.99 moko 376: void VMail::fill_received(Request& r) {
1.45 paf 377: if(r.request_info.mail_received) {
1.99 moko 378: source_charset=&r.charsets.source();
1.131 moko 379: g_mime_init();
1.1 paf 380: // create stream with CRLF filter
1.131 moko 381: #if GMIME_MAJOR_VERSION > 2
1.137 moko 382: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_pipe_new(STDIN_FILENO));
1.131 moko 383: #else
1.137 moko 384: GMimeStream *stream = g_mime_stream_filter_new(g_mime_stream_fs_new(STDIN_FILENO));
1.131 moko 385: #endif
1.137 moko 386: g_mime_stream_filter_add(GMIME_STREAM_FILTER(stream), g_mime_filter_crlf_new(false, false));
1.1 paf 387: try {
1.99 moko 388: // parse incoming message
1.129 moko 389: GMimeMessage *message=g_mime_parser_construct_message(g_mime_parser_new_with_stream(stream), NULL);
1.138 moko 390: if(message){
391: parse(r, message, vreceived.hash());
392: g_object_unref(GMIME_OBJECT(message));
393: }
1.48 paf 394: } catch(const Exception& e) {
395: HashStringValue& received=vreceived.hash();
1.125 moko 396: putReceived(received, VALUE_NAME, "<exception occurred while parsing message>");
1.99 moko 397: putReceived(received, EXCEPTION_VALUE, e.comment());
1.1 paf 398: } catch(...) {
1.99 moko 399: // abnormal stream free
400: g_object_unref(stream);
1.48 paf 401: rethrow;
1.1 paf 402: }
1.99 moko 403: g_object_unref(stream);
404:
405: g_mime_shutdown();
1.1 paf 406: }
407: }
408:
1.99 moko 409: #else // WITH_MAILRECEIVE
410: void VMail::fill_received(Request&){}
411: #endif // WITH_MAILRECEIVE
412:
1.9 paf 413: typedef int (*string_contains_char_which_check)(int);
1.99 moko 414:
1.45 paf 415: static bool string_contains_char_which(const char* string, string_contains_char_which_check check) {
1.9 paf 416: while(char c=*string++) {
1.62 paf 417: if(check((unsigned char)c))
1.9 paf 418: return true;
419: }
420: return false;
421: }
1.99 moko 422:
1.32 paf 423: static char *trimBoth(char *s) {
424: // sanity check
425: if(!s)
426: return 0;
427:
428: // trim head whitespace
1.62 paf 429: while(*s && isspace((unsigned char)*s))
1.32 paf 430: s++;
431: // trim tail whitespace
432: char *tail=s+strlen(s);
433: if(tail>s) {
434: do {
435: --tail;
1.62 paf 436: if(isspace((unsigned char)*tail))
1.32 paf 437: *tail=0;
438: } while(tail>s);
439: }
440: // return it
441: return s;
442: }
1.99 moko 443:
1.45 paf 444: static void extractEmail(String& result, char *email) {
1.32 paf 445: email=trimBoth(email);
1.45 paf 446: result.append_help_length(email, 0, String::L_TAINTED);
1.9 paf 447:
448: /*
449: http://www.faqs.org/rfcs/rfc822.html
450:
451: addr-spec = local-part "@" domain ; global address
452:
453: local-part = word *("." word) ; uninterpreted case-preserved
454: word = atom / quoted-string
455:
456: domain = sub-domain *("." sub-domain)
457: sub-domain = domain-ref / domain-literal
458: domain-ref = atom ; symbolic reference
459:
1.92 misha 460: domain-literal << ignoring for now
1.9 paf 461: quoted-string in word << ignoring for now
462:
463: atom = 1*<any CHAR except specials, SPACE and CTLs> << the ONLY to check
464:
465: specials = "(" / ")" / "<" / ">" / "@" ; Must be in quoted-
1.92 misha 466: / "," / ";" / ":" / "\" / <"> ; string, to use
467: / "." / "[" / "]" ; within a word.
1.9 paf 468:
469: */
1.45 paf 470: const char* exception_type="email.format";
1.9 paf 471: if(strpbrk(email, "()<>,;:\\\"[]"/*specials minus @ and . */))
1.126 moko 472: throw Exception(exception_type, &result, "email contains bad characters (specials)");
1.9 paf 473: if(string_contains_char_which(email, (string_contains_char_which_check)isspace))
1.126 moko 474: throw Exception(exception_type, &result, "email contains bad characters (whitespace)");
1.9 paf 475: if(string_contains_char_which(email, (string_contains_char_which_check)iscntrl))
1.126 moko 476: throw Exception(exception_type, &result, "email contains bad characters (control)");
1.16 paf 477: if(result.is_empty())
1.126 moko 478: throw Exception(exception_type, 0, "email is empty");
1.39 paf 479: }
480:
1.45 paf 481: static const String& extractEmails(const String& string) {
482: char *emails=string.cstrm();
483: String& result=*new String;
1.39 paf 484: while(char *email=lsplit(&emails, ',')) {
485: rsplit(email, '>');
486: if(char *in_brackets=lsplit(email, '<'))
487: email=in_brackets;
488: if(!result.is_empty())
489: result<<",";
1.45 paf 490: extractEmail(result, email);
1.39 paf 491: }
1.9 paf 492:
493: return result;
494: }
1.45 paf 495:
496: #ifndef DOXYGEN
497: struct Store_message_element_info {
498: Request_charsets& charsets;
499: String& header;
500: const String* & from;
1.60 paf 501: bool extract_to; String* & to;
1.45 paf 502: bool mime_version_specified;
503: ArrayValue* parts[P_TYPES_COUNT];
504: int parts_count;
505: bool backward_compatibility;
506: Value* content_type;
1.66 paf 507: bool had_content_disposition;
1.45 paf 508:
1.126 moko 509: Store_message_element_info(Request_charsets& acharsets, String& aheader, const String* & afrom, bool aextract_to, String* & ato):
1.45 paf 510: charsets(acharsets),
511: header(aheader),
512: from(afrom),
1.60 paf 513: extract_to(aextract_to), to(ato),
1.45 paf 514: mime_version_specified(false),
515: parts_count(0),
1.66 paf 516: backward_compatibility(false), content_type(0),
1.133 moko 517: had_content_disposition(false)
518: {
519: for(int pt=0; pt<P_TYPES_COUNT; pt++)
520: parts[pt]=new ArrayValue(1);
1.45 paf 521: }
522: };
523: #endif
1.99 moko 524:
1.126 moko 525: size_t mail_header_utf8_substring(const char *mail, size_t sub_length, size_t length){
526: int error_offset;
527: if(int error_code=pa_pcre_valid_utf((unsigned char *)mail, sub_length, &error_offset)){
528: if(error_code<PCRE_UTF8_ERR6){ // Missing X byte at the end of the string errors
529: sub_length+=error_code; // adding X bytes
530: return sub_length < length ? sub_length : length;
531: }
532: }
533:
534: return sub_length;
535: }
536:
537: static void store_message_element(HashStringValue::key_type raw_element_name, HashStringValue::value_type element_value, Store_message_element_info *info) {
538: const String& low_element_name=String(raw_element_name, String::L_TAINTED).change_case(info->charsets.source(), String::CC_LOWER);
1.1 paf 539:
540: // exclude internals
1.52 paf 541: if(low_element_name==MAIL_OPTIONS_NAME
542: || low_element_name==CHARSET_NAME
1.5 paf 543: || low_element_name==VALUE_NAME
544: || low_element_name==RAW_NAME
1.140 ! moko 545: || low_element_name==TABLES_NAME
1.64 paf 546: || low_element_name==FORMAT_NAME
1.76 misha 547: || low_element_name==NAME_NAME
1.94 misha 548: || low_element_name==CID_NAME
549: || low_element_name==MAIL_DEBUG_NAME)
1.1 paf 550: return;
551:
552: // grep parts
553: for(int pt=0; pt<P_TYPES_COUNT; pt++) {
1.45 paf 554: if(low_element_name.starts_with(part_name_begins[pt])) {
1.29 paf 555: // check that $.message# '#' is digit
1.45 paf 556: size_t start_len=strlen(part_name_begins[pt]);
557: if(low_element_name.length()>start_len) {
558: const char* at_num=low_element_name.mid(start_len, start_len+1).cstr();
1.63 paf 559: if(!isdigit((unsigned char)*at_num))
1.29 paf 560: continue;
561: }
1.45 paf 562: *info->parts[pt]+=element_value;
563: info->parts_count++;
1.1 paf 564: return;
565: }
566: }
567:
1.10 paf 568: // fetch some special headers
1.45 paf 569: if(low_element_name=="from")
570: info->from=&extractEmails(element_value->as_string());
1.90 misha 571: if(low_element_name==CONTENT_DISPOSITION)
1.66 paf 572: info->had_content_disposition=true;
1.60 paf 573: if(info->extract_to) { // defined only when SMTP used, see mail.C [collecting info for RCPT to-s]
1.39 paf 574: bool is_to=low_element_name=="to" ;
575: bool is_cc=low_element_name=="cc" ;
576: bool is_bcc=low_element_name=="bcc" ;
577: if(is_to||is_cc||is_bcc) {
1.45 paf 578: if(!info->to)
579: info->to=new String;
1.39 paf 580: else
1.45 paf 581: *info->to << ",";
582: *info->to << extractEmails(element_value->as_string());
1.39 paf 583: }
584:
585: if(is_bcc) // blinding it
1.45 paf 586: return;
1.39 paf 587: }
1.37 paf 588: if(low_element_name=="mime-version")
1.45 paf 589: info->mime_version_specified=true;
1.1 paf 590:
1.45 paf 591: // has content type?
592: if(low_element_name==CONTENT_TYPE_NAME) {
593: info->content_type=element_value;
594: if(info->backward_compatibility)
595: return;
1.39 paf 596: }
1.1 paf 597:
1.45 paf 598: // preparing header line
1.123 moko 599: const String& source_line=attributed_meaning_to_string(*element_value, String::L_AS_IS);
600:
1.66 paf 601: if(source_line.is_empty())
602: return; // we don't need empty headers here [used in clearing content-disposition]
603:
1.45 paf 604: const char* source_line_cstr=source_line.cstr();
1.126 moko 605: String::C mail=Charset::transcode(String::C(source_line_cstr, source_line.length()), info->charsets.source(), info->charsets.mail());
1.92 misha 606:
1.45 paf 607: String& mail_line=*new String;
1.126 moko 608: if(low_element_name=="to" || low_element_name=="cc" || low_element_name=="bcc")
1.73 paf 609: {
610: // never wrap address lines, mailer can not handle wrapped properly
611: mail_line.append_strdup(mail.str, mail.length, String::L_MAIL_HEADER);
612: } else {
613: while(mail.length) {
1.126 moko 614: bool too_long=mail.length > MAX_CHARS_IN_HEADER_LINE;
615: 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 616:
617: mail_line.append_strdup(mail.str, length, String::L_MAIL_HEADER);
1.126 moko 618:
1.73 paf 619: mail.length-=length;
1.126 moko 620: mail.str+=length;
1.73 paf 621:
622: if(too_long)
623: mail_line << "\n "; // break header and continue it on next line
1.126 moko 624: }
1.73 paf 625: }
1.45 paf 626:
627: // append header line
628: info->header
1.91 misha 629: << capitalize(raw_element_name.cstr())
1.96 misha 630: << ": " << mail_line.untaint_cstr(String::L_AS_IS, 0, &info->charsets)
1.45 paf 631: << "\n";
632: }
633:
634: static const String& file_value_to_string(Request& r, Value* send_value) {
1.64 paf 635: String& result=*new String;
636:
1.45 paf 637: VFile* vfile;
638: const String* file_name=0;
639: Value* vformat=0;
1.76 misha 640: Value* vcid=0;
1.66 paf 641: const String* dummy_from;
642: String* dummy_to;
1.95 misha 643: Store_message_element_info info(r.charsets, result, dummy_from, false, dummy_to);
644:
1.120 moko 645: HashStringValue *send_hash=send_value->get_hash();
646: if(send_hash && !send_value->as("file")) { // hash
1.74 paf 647: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.64 paf 648:
1.1 paf 649: // $.value
1.45 paf 650: if(Value* value=send_hash->get(value_name))
1.139 moko 651: vfile=value->as_vfile();
1.1 paf 652: else
1.120 moko 653: throw Exception(PARSER_RUNTIME, 0, "file part has no $value");
1.1 paf 654:
655: // $.format
1.45 paf 656: vformat=send_hash->get(format_name);
1.1 paf 657:
1.76 misha 658: // $.content-id
659: vcid=send_hash->get(cid_name);
660:
1.6 paf 661: // $.name
1.45 paf 662: if(Value* vfile_name=send_hash->get(name_name)) // $name specified
1.1 paf 663: file_name=&vfile_name->as_string();
1.28 paf 664: } else // must be VFile then
1.139 moko 665: vfile=send_value->as_vfile();
1.28 paf 666:
667: if(!file_name)
1.45 paf 668: file_name=&vfile->fields().get(name_name)->as_string();
1.28 paf 669:
1.95 misha 670: const char* file_name_cstr;
671: const char* quoted_file_name_cstr;
672: {
673: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.97 misha 674: file_name_cstr=file_name->untaint_and_transcode_cstr(String::L_FILE_SPEC, &charsets);
1.95 misha 675: quoted_file_name_cstr=String(file_name_cstr).taint_cstr(String::L_MAIL_HEADER, 0, &charsets);
676: }
1.1 paf 677:
1.95 misha 678: // Content-Type: application/octet-stream
679: result
680: << HTTP_CONTENT_TYPE_CAPITALIZED ": "
681: << r.mime_type_of(file_name_cstr)
682: << "; name=\""
683: << quoted_file_name_cstr
684: << "\"\n";
1.66 paf 685:
1.95 misha 686: if(!info.had_content_disposition) // $.Content-Disposition wasn't specified by user
1.79 misha 687: result
1.91 misha 688: << CONTENT_DISPOSITION_CAPITALIZED ": "
1.79 misha 689: << ( vcid ? CONTENT_DISPOSITION_INLINE : CONTENT_DISPOSITION_ATTACHMENT )
690: << "; "
1.95 misha 691: << CONTENT_DISPOSITION_FILENAME_NAME"=\"" << quoted_file_name_cstr << "\"\n";
1.1 paf 692:
1.79 misha 693: if(vcid)
1.93 misha 694: result
695: << "Content-Id: <"
696: << vcid->as_string()
697: << ">\n"; // @todo: value must be escaped as %hh
1.79 misha 698:
1.45 paf 699: const String* type=vformat?&vformat->as_string():0;
1.93 misha 700: if(!type/*default*/ || *type=="base64") {
701: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": base64\n\n";
702: result << pa_base64_encode(vfile->value_ptr(), vfile->value_size());
1.75 paf 703: } else {
1.93 misha 704: if(*type=="uue") {
705: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED ": x-uuencode\n\n";
1.95 misha 706: result << pa_uuencode((const unsigned char*)vfile->value_ptr(), vfile->value_size(), file_name_cstr);
707: } else
1.133 moko 708: throw Exception(PARSER_RUNTIME, type, "unknown attachment encode format");
1.75 paf 709: }
1.95 misha 710:
1.1 paf 711: return result;
712: }
713:
1.133 moko 714: static const String& text_value_to_string(Request& r, PartType pt, Value* send_value, Store_message_element_info& info) {
1.45 paf 715: String& result=*new String;
1.1 paf 716:
1.45 paf 717: Value* text_value;
1.81 misha 718: Value* content_transfer_encoding=0;
1.45 paf 719: if(HashStringValue* send_hash=send_value->get_hash()) {
1.1 paf 720: // $.USER-HEADERS
1.92 misha 721: info.content_type=0;
722: info.backward_compatibility=false; // reset
1.74 paf 723: send_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.1 paf 724: // $.value
1.45 paf 725: text_value=send_hash->get(value_name);
1.1 paf 726: if(!text_value)
1.133 moko 727: throw Exception(PARSER_RUNTIME, 0, "%s part has no $" VALUE_NAME, part_name_begins[pt]);
1.81 misha 728: content_transfer_encoding=send_hash->get(content_transfer_encoding_name);
1.1 paf 729: } else
1.45 paf 730: text_value=send_value;
1.1 paf 731:
1.45 paf 732: if(!info.content_type) {
733: result
1.91 misha 734: << HTTP_CONTENT_TYPE_CAPITALIZED ": text/" << (pt==P_TEXT?"plain":"html")
1.45 paf 735: << "; charset=" << info.charsets.mail().NAME()
736: << "\n";
1.1 paf 737: }
1.81 misha 738: if(!content_transfer_encoding)
1.91 misha 739: result << CONTENT_TRANSFER_ENCODING_CAPITALIZED << ": 8bit\n";
1.1 paf 740:
741: // header|body separator
742: result << "\n";
743:
744: // body
1.45 paf 745: const String* body;
1.122 moko 746: String::Language body_lang=String::L_AS_IS;
747:
1.1 paf 748: switch(pt) {
749: case P_TEXT:
1.92 misha 750: {
751: body=&text_value->as_string();
752: break;
753: }
1.122 moko 754: case P_HTML:
1.1 paf 755: {
1.122 moko 756: body_lang=String::Language(String::L_HTML | String::L_OPTIMIZE_BIT);
1.55 paf 757: if(text_value->get_junction())
1.41 paf 758: body=&r.process_to_string(*text_value);
1.92 misha 759: else
1.122 moko 760: throw Exception(PARSER_RUNTIME, 0, "html part value must be code");
1.1 paf 761: break;
762: }
1.53 paf 763: default:
1.122 moko 764: throw Exception(0, 0, "unhandled part type #%d", pt);
1.41 paf 765: }
1.122 moko 766:
1.41 paf 767: if(body) {
1.69 paf 768: Request_charsets charsets(r.charsets.source(), r.charsets.mail()/*uri!*/, r.charsets.mail());
1.122 moko 769: const char* body_cstr=body->untaint_and_transcode_cstr(body_lang, &charsets);
1.95 misha 770: result.append_know_length(body_cstr, strlen(body_cstr), String::L_CLEAN);
1.1 paf 771: }
772:
773: return result;
1.111 moko 774: }
1.1 paf 775:
776: /// @todo files and messages in order (file, file2, ...)
1.134 moko 777: const String& VMail::message_hash_to_string(Request& r, HashStringValue* message_hash, const String* & from, bool extract_to, String* & to) {
1.34 paf 778:
1.1 paf 779: if(!message_hash)
1.133 moko 780: throw Exception(PARSER_RUNTIME, 0, "message must be hash");
1.1 paf 781:
1.45 paf 782: String& result=*new String;
1.1 paf 783:
1.45 paf 784: if(Value* vrecodecharset_name=message_hash->get(charset_name))
1.121 moko 785: r.charsets.set_mail(pa_charsets.get(vrecodecharset_name->as_string()));
1.1 paf 786: else
1.45 paf 787: r.charsets.set_mail(r.charsets.source());
788: // no big deal that we leave it set. they wont miss this point which would reset it
1.1 paf 789:
1.133 moko 790: Store_message_element_info info(r.charsets, result, from, extract_to, to);
791:
1.1 paf 792: {
1.133 moko 793: // for backward compatibilyty $.body+$.content-type -> $.text[$.value[] $.content-type[]]
1.45 paf 794:
795: Value* body=message_hash->get("body");
796: if(body) {
797: message_hash->remove("body");
798: info.backward_compatibility=true;
1.133 moko 799: }
1.74 paf 800: message_hash->for_each<Store_message_element_info*>(store_message_element, &info);
1.45 paf 801:
802: if(body) {
803: VHash& text_part=*new VHash();
804: HashStringValue& hash=text_part.hash();
805: hash.put(value_name, body);
806: if(info.content_type)
807: hash.put(content_type_name, info.content_type);
808:
809: *info.parts[P_TEXT]+=&text_part;
810: info.parts_count++;
811: }
812:
1.37 paf 813: if(!info.mime_version_specified)
814: result << "MIME-Version: 1.0\n"; // MIME-Version: default
1.1 paf 815: }
816:
1.45 paf 817: int textCount=info.parts[P_TEXT]->count();
1.1 paf 818: if(textCount>1)
1.136 moko 819: throw Exception(PARSER_RUNTIME, 0, "multiple text parts are not supported, use file part");
1.45 paf 820: int htmlCount=info.parts[P_HTML]->count();
1.1 paf 821: if(htmlCount>1)
1.136 moko 822: throw Exception(PARSER_RUNTIME, 0, "multiple html parts are not supported, use file part");
1.1 paf 823:
824: bool multipart=info.parts_count>1;
825: bool alternative=textCount && htmlCount;
826: // header
827: char *boundary=0;
828: if(multipart) {
1.118 moko 829: boundary=get_uuid_boundary();
1.76 misha 830:
831: bool is_inline = false;
832: {
833: ArrayValue& files=*info.parts[P_FILE];
834: for(size_t i=0; i<files.count(); i++) {
1.82 misha 835: HashStringValue* file;
836: if((file=files.get(i)->get_hash()) && file->get(cid_name)){
1.76 misha 837: is_inline = true;
838: break;
1.77 misha 839: }
1.76 misha 840: }
841: }
1.133 moko 842:
1.91 misha 843: result << HTTP_CONTENT_TYPE_CAPITALIZED ": " << ( is_inline ? HTTP_CONTENT_TYPE_MULTIPART_RELATED : HTTP_CONTENT_TYPE_MULTIPART_MIXED ) << ";";
1.76 misha 844:
1.1 paf 845: // multi-part
1.133 moko 846: result << " boundary=\"" << boundary << "\"\n\nThis is a multi-part message in MIME format.";
1.1 paf 847: }
848:
849: // alternative or not
850: {
851: if(alternative) {
1.45 paf 852: result << "\n\n--" << boundary << "\n" // intermediate boundary
1.91 misha 853: HTTP_CONTENT_TYPE_CAPITALIZED ": multipart/alternative; boundary=\"ALT" << boundary << "\"\n";
1.1 paf 854: }
855: for(int i=0; i<2; i++) {
856: PartType pt=i==0?P_TEXT:P_HTML;
1.45 paf 857: if(info.parts[pt]->count()) {
1.1 paf 858: if(alternative)
859: result << "\n\n--ALT" << boundary << "\n"; // intermediate boundary
860: else if(boundary)
861: result << "\n\n--" << boundary << "\n"; // intermediate boundary
1.45 paf 862: result << text_value_to_string(r, pt, info.parts[pt]->get(0), info);
1.1 paf 863: }
864: }
865: if(alternative)
866: result << "\n\n--ALT" << boundary << "--\n";
867: }
868:
869: // messages
870: {
1.45 paf 871: ArrayValue& messages=*info.parts[P_MESSAGE];
872: for(size_t i=0; i<messages.count(); i++) {
1.1 paf 873: if(boundary)
874: result << "\n\n--" << boundary << "\n"; // intermediate boundary
875:
1.45 paf 876: const String* dummy_from;
877: String* dummy_to;
1.134 moko 878: result << message_hash_to_string(r, messages.get(i)->get_hash(), dummy_from, false, dummy_to);
1.1 paf 879: }
880: }
881:
1.132 moko 882: // files go last
883: {
884: ArrayValue& files=*info.parts[P_FILE];
885: for(size_t i=0; i<files.count(); i++) {
886: if(boundary)
887: result << "\n\n--" << boundary << "\n"; // intermediate boundary
888: result << file_value_to_string(r, files.get(i));
889: }
890: }
891:
1.1 paf 892: // tailer
893: if(boundary)
894: result << "\n\n--" << boundary << "--\n"; // finish boundary
895:
896: // return
897: return result;
898: }
899:
900:
1.88 misha 901: Value* VMail::get_element(const String& aname) {
1.1 paf 902: // $fields
1.3 paf 903: #ifdef WITH_MAILRECEIVE
1.1 paf 904: if(aname==MAIL_RECEIVED_ELEMENT_NAME)
1.48 paf 905: return &vreceived;
1.1 paf 906: #endif
907:
1.110 misha 908: // $method
1.88 misha 909: if(Value* result=VStateless_class::get_element(aname))
1.1 paf 910: return result;
911:
1.109 moko 912: return bark("%s field not found", &aname);
1.1 paf 913: }
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