2 * Copyright © 2009 Carl Worth
4 * This program is free software: you can redistribute it and/or modify
5 * it under the terms of the GNU General Public License as published by
6 * the Free Software Foundation, either version 3 of the License, or
7 * (at your option) any later version.
9 * This program is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU General Public License
15 * along with this program. If not, see https://www.gnu.org/licenses/ .
17 * Author: Carl Worth <cworth@cworth.org>
20 #include "notmuch-private.h"
22 #include <gmime/gmime.h>
23 #include <gmime/gmime-filter.h>
33 int next_if_not_match;
36 /* Simple, linear state-transition diagram for the uuencode filter.
38 * If the character being processed is within the range of [a, b]
39 * for the current state then we transition next_if_match
40 * state. If not, we transition to the next_if_not_match state.
42 * The final two states are special in that they are the states in
43 * which we discard data. */
44 static const int first_uuencode_skipping_state = 11;
45 static const scanner_state_t uuencode_states[] = {
46 { 0, 'b', 'b', 1, 0 },
47 { 1, 'e', 'e', 2, 0 },
48 { 2, 'g', 'g', 3, 0 },
49 { 3, 'i', 'i', 4, 0 },
50 { 4, 'n', 'n', 5, 0 },
51 { 5, ' ', ' ', 6, 0 },
52 { 6, '0', '7', 7, 0 },
53 { 7, '0', '7', 8, 0 },
54 { 8, '0', '7', 9, 0 },
55 { 9, ' ', ' ', 10, 0 },
56 { 10, '\n', '\n', 11, 10 },
57 { 11, 'M', 'M', 12, 0 },
58 { 12, ' ', '`', 12, 11 }
61 /* The following table is intended to implement this DFA (in 'dot'
62 * format). Note that 2 and 3 are "hidden" states used to step through
63 * the possible out edges of state 1.
65 * digraph html_filter {
69 * 1 -> 5 [label="\""];
74 * 5 -> 1 [label="\""];
78 static const int first_html_skipping_state = 1;
79 static const scanner_state_t html_states[] = {
80 { 0, '<', '<', 1, 0 },
81 { 1, '\'', '\'', 4, 2 }, /* scanning for quote or > */
82 { 1, '"', '"', 5, 3 },
83 { 1, '>', '>', 0, 1 },
84 { 4, '\'', '\'', 1, 4 }, /* inside single quotes */
85 { 5, '"', '"', 1, 5 }, /* inside double quotes */
88 /* Oh, how I wish that gobject didn't require so much noisy boilerplate!
89 * (Though I have at least eliminated some of the stock set...) */
90 typedef struct _NotmuchFilterDiscardNonTerm NotmuchFilterDiscardNonTerm;
91 typedef struct _NotmuchFilterDiscardNonTermClass NotmuchFilterDiscardNonTermClass;
94 * NotmuchFilterDiscardNonTerm:
96 * @parent_object: parent #GMimeFilter
97 * @encode: encoding vs decoding
98 * @state: State of the parser
100 * A filter to discard uuencoded portions of an email.
102 * A uuencoded portion is identified as beginning with a line
105 * begin [0-7][0-7][0-7] .*
107 * After that detection, and beginning with the following line,
108 * characters will be discarded as long as the first character of each
109 * line begins with M and subsequent characters on the line are within
110 * the range of ASCII characters from ' ' to '`'.
112 * This is not a perfect UUencode filter. It's possible to have a
113 * message that will legitimately match that pattern, (so that some
114 * legitimate content is discarded). And for most UUencoded files, the
115 * final line of encoded data (the line not starting with M) will be
118 struct _NotmuchFilterDiscardNonTerm {
119 GMimeFilter parent_object;
120 GMimeContentType *content_type;
122 int first_skipping_state;
123 const scanner_state_t *states;
126 struct _NotmuchFilterDiscardNonTermClass {
127 GMimeFilterClass parent_class;
130 static GMimeFilter *notmuch_filter_discard_non_term_new (GMimeContentType *content);
132 static void notmuch_filter_discard_non_term_finalize (GObject *object);
134 static GMimeFilter *filter_copy (GMimeFilter *filter);
135 static void filter_filter (GMimeFilter *filter, char *in, size_t len, size_t prespace,
136 char **out, size_t *outlen, size_t *outprespace);
137 static void filter_complete (GMimeFilter *filter, char *in, size_t len, size_t prespace,
138 char **out, size_t *outlen, size_t *outprespace);
139 static void filter_reset (GMimeFilter *filter);
142 static GMimeFilterClass *parent_class = NULL;
145 notmuch_filter_discard_non_term_class_init (NotmuchFilterDiscardNonTermClass *klass,
146 unused (void *class_data))
148 GObjectClass *object_class = G_OBJECT_CLASS (klass);
149 GMimeFilterClass *filter_class = GMIME_FILTER_CLASS (klass);
151 parent_class = (GMimeFilterClass *) g_type_class_ref (GMIME_TYPE_FILTER);
153 object_class->finalize = notmuch_filter_discard_non_term_finalize;
155 filter_class->copy = filter_copy;
156 filter_class->filter = filter_filter;
157 filter_class->complete = filter_complete;
158 filter_class->reset = filter_reset;
162 notmuch_filter_discard_non_term_finalize (GObject *object)
164 G_OBJECT_CLASS (parent_class)->finalize (object);
168 filter_copy (GMimeFilter *gmime_filter)
170 NotmuchFilterDiscardNonTerm *filter = (NotmuchFilterDiscardNonTerm *) gmime_filter;
172 return notmuch_filter_discard_non_term_new (filter->content_type);
176 filter_filter (GMimeFilter *gmime_filter, char *inbuf, size_t inlen, size_t prespace,
177 char **outbuf, size_t *outlen, size_t *outprespace)
179 NotmuchFilterDiscardNonTerm *filter = (NotmuchFilterDiscardNonTerm *) gmime_filter;
180 const scanner_state_t *states = filter->states;
181 const char *inptr = inbuf;
182 const char *inend = inbuf + inlen;
189 g_mime_filter_set_size (gmime_filter, inlen, false);
190 outptr = gmime_filter->outbuf;
192 next = filter->state;
193 while (inptr < inend) {
194 /* Each state is defined by a contiguous set of rows of the
195 * state table marked by a common value for '.state'. The
196 * state numbers must be equal to the index of the first row
197 * in a given state; thus the loop condition here looks for a
198 * jump to a first row of a state, which is a real transition
199 * in the underlying DFA.
202 if (*inptr >= states[next].a && *inptr <= states[next].b) {
203 next = states[next].next_if_match;
205 next = states[next].next_if_not_match;
208 } while (next != states[next].state);
210 if (filter->state < filter->first_skipping_state)
213 filter->state = next;
217 *outlen = outptr - gmime_filter->outbuf;
218 *outprespace = gmime_filter->outpre;
219 *outbuf = gmime_filter->outbuf;
223 filter_complete (GMimeFilter *filter, char *inbuf, size_t inlen, size_t prespace,
224 char **outbuf, size_t *outlen, size_t *outprespace)
227 filter_filter (filter, inbuf, inlen, prespace, outbuf, outlen, outprespace);
231 filter_reset (GMimeFilter *gmime_filter)
233 NotmuchFilterDiscardNonTerm *filter = (NotmuchFilterDiscardNonTerm *) gmime_filter;
239 * notmuch_filter_discard_non_term_new:
241 * Returns: a new #NotmuchFilterDiscardNonTerm filter.
244 notmuch_filter_discard_non_term_new (GMimeContentType *content_type)
246 static GType type = 0;
247 NotmuchFilterDiscardNonTerm *filter;
250 static const GTypeInfo info = {
251 .class_size = sizeof (NotmuchFilterDiscardNonTermClass),
253 .base_finalize = NULL,
254 .class_init = (GClassInitFunc) notmuch_filter_discard_non_term_class_init,
255 .class_finalize = NULL,
257 .instance_size = sizeof (NotmuchFilterDiscardNonTerm),
259 .instance_init = NULL,
263 type = g_type_register_static (GMIME_TYPE_FILTER, "NotmuchFilterDiscardNonTerm", &info, (GTypeFlags) 0);
266 filter = (NotmuchFilterDiscardNonTerm *) g_object_new (type, NULL);
267 filter->content_type = content_type;
269 if (g_mime_content_type_is_type (content_type, "text", "html")) {
270 filter->states = html_states;
271 filter->first_skipping_state = first_html_skipping_state;
273 filter->states = uuencode_states;
274 filter->first_skipping_state = first_uuencode_skipping_state;
277 return (GMimeFilter *) filter;
280 /* We're finally down to a single (NAME + address) email "mailbox". */
282 _index_address_mailbox (notmuch_message_t *message,
283 const char *prefix_name,
284 InternetAddress *address)
286 InternetAddressMailbox *mailbox = INTERNET_ADDRESS_MAILBOX (address);
287 const char *name, *addr, *combined;
288 void *local = talloc_new (message);
290 name = internet_address_get_name (address);
291 addr = internet_address_mailbox_get_addr (mailbox);
293 /* Combine the name and address and index them as a phrase. */
295 combined = talloc_asprintf (local, "%s %s", name, addr);
302 _notmuch_message_gen_terms (message, prefix_name, combined);
308 _index_address_list (notmuch_message_t *message,
309 const char *prefix_name,
310 InternetAddressList *addresses);
312 /* The outer loop over the InternetAddressList wasn't quite enough.
313 * There can actually be a tree here where a single member of the list
314 * is a "group" containing another list. Recurse please.
317 _index_address_group (notmuch_message_t *message,
318 const char *prefix_name,
319 InternetAddress *address)
321 InternetAddressGroup *group;
322 InternetAddressList *list;
324 group = INTERNET_ADDRESS_GROUP (address);
325 list = internet_address_group_get_members (group);
330 _index_address_list (message, prefix_name, list);
334 _index_address_list (notmuch_message_t *message,
335 const char *prefix_name,
336 InternetAddressList *addresses)
339 InternetAddress *address;
341 if (addresses == NULL)
344 for (i = 0; i < internet_address_list_length (addresses); i++) {
345 address = internet_address_list_get_address (addresses, i);
346 if (INTERNET_ADDRESS_IS_MAILBOX (address)) {
347 _index_address_mailbox (message, prefix_name, address);
348 } else if (INTERNET_ADDRESS_IS_GROUP (address)) {
349 _index_address_group (message, prefix_name, address);
351 INTERNAL_ERROR ("GMime InternetAddress is neither a mailbox nor a group.\n");
357 _index_content_type (notmuch_message_t *message, GMimeObject *part)
359 GMimeContentType *content_type = g_mime_object_get_content_type (part);
362 char *mime_string = g_mime_content_type_get_mime_type (content_type);
364 _notmuch_message_gen_terms (message, "mimetype", mime_string);
365 g_free (mime_string);
371 _index_encrypted_mime_part (notmuch_message_t *message, notmuch_indexopts_t *indexopts,
372 GMimeMultipartEncrypted *part,
373 _notmuch_message_crypto_t *msg_crypto);
375 /* Callback to generate terms for each mime part of a message. */
377 _index_mime_part (notmuch_message_t *message,
378 notmuch_indexopts_t *indexopts,
380 _notmuch_message_crypto_t *msg_crypto)
382 GMimeStream *stream, *filter;
383 GMimeFilter *discard_non_term_filter;
384 GMimeDataWrapper *wrapper;
385 GByteArray *byte_array;
386 GMimeContentDisposition *disposition;
387 GMimeContentType *content_type;
390 GMimeObject *repaired_part = NULL;
393 _notmuch_database_log (notmuch_message_get_database (message),
394 "Warning: Not indexing empty mime part.\n");
398 repaired_part = _notmuch_repair_mixed_up_mangled (part);
400 /* This was likely "Mixed Up" in transit! We will instead use
401 * the more likely-to-be-correct variant. */
402 notmuch_message_add_property (message, "index.repaired", "mixedup");
403 part = repaired_part;
406 _index_content_type (message, part);
408 if (GMIME_IS_MULTIPART (part)) {
409 GMimeMultipart *multipart = GMIME_MULTIPART (part);
412 if (GMIME_IS_MULTIPART_SIGNED (multipart))
413 _notmuch_message_add_term (message, "tag", "signed");
415 if (GMIME_IS_MULTIPART_ENCRYPTED (multipart))
416 _notmuch_message_add_term (message, "tag", "encrypted");
418 for (i = 0; i < g_mime_multipart_get_count (multipart); i++) {
420 if (GMIME_IS_MULTIPART_SIGNED (multipart)) {
421 /* Don't index the signature, but index its content type. */
422 if (i == GMIME_MULTIPART_SIGNED_SIGNATURE) {
423 _index_content_type (message,
424 g_mime_multipart_get_part (multipart, i));
426 } else if (i != GMIME_MULTIPART_SIGNED_CONTENT) {
427 _notmuch_database_log (notmuch_message_get_database (message),
428 "Warning: Unexpected extra parts of multipart/signed. Indexing anyway.\n");
431 if (GMIME_IS_MULTIPART_ENCRYPTED (multipart)) {
432 _index_content_type (message,
433 g_mime_multipart_get_part (multipart, i));
434 if (i == GMIME_MULTIPART_ENCRYPTED_CONTENT) {
435 _index_encrypted_mime_part (message, indexopts,
436 GMIME_MULTIPART_ENCRYPTED (part),
439 if (i != GMIME_MULTIPART_ENCRYPTED_VERSION) {
440 _notmuch_database_log (notmuch_message_get_database (message),
441 "Warning: Unexpected extra parts of multipart/encrypted.\n");
446 child = g_mime_multipart_get_part (multipart, i);
447 GMimeObject *toindex = child;
448 if (_notmuch_message_crypto_potential_payload (msg_crypto, child, part, i) &&
449 msg_crypto->decryption_status == NOTMUCH_MESSAGE_DECRYPTED_FULL) {
450 toindex = _notmuch_repair_crypto_payload_skip_legacy_display (child);
451 if (toindex != child)
452 notmuch_message_add_property (message, "index.repaired", "skip-protected-headers-legacy-display");
454 _index_mime_part (message, indexopts, toindex, msg_crypto);
459 if (GMIME_IS_MESSAGE_PART (part)) {
460 GMimeMessage *mime_message;
462 mime_message = g_mime_message_part_get_message (GMIME_MESSAGE_PART (part));
464 _index_mime_part (message, indexopts, g_mime_message_get_mime_part (mime_message), msg_crypto);
469 if (! (GMIME_IS_PART (part))) {
470 _notmuch_database_log (notmuch_message_get_database (message),
471 "Warning: Not indexing unknown mime part: %s.\n",
472 g_type_name (G_OBJECT_TYPE (part)));
476 disposition = g_mime_object_get_content_disposition (part);
478 strcasecmp (g_mime_content_disposition_get_disposition (disposition),
479 GMIME_DISPOSITION_ATTACHMENT) == 0) {
480 const char *filename = g_mime_part_get_filename (GMIME_PART (part));
482 _notmuch_message_add_term (message, "tag", "attachment");
483 _notmuch_message_gen_terms (message, "attachment", filename);
485 /* XXX: Would be nice to call out to something here to parse
486 * the attachment into text and then index that. */
490 byte_array = g_byte_array_new ();
492 stream = g_mime_stream_mem_new_with_byte_array (byte_array);
493 g_mime_stream_mem_set_owner (GMIME_STREAM_MEM (stream), false);
495 filter = g_mime_stream_filter_new (stream);
497 content_type = g_mime_object_get_content_type (part);
498 discard_non_term_filter = notmuch_filter_discard_non_term_new (content_type);
500 g_mime_stream_filter_add (GMIME_STREAM_FILTER (filter),
501 discard_non_term_filter);
503 charset = g_mime_object_get_content_type_parameter (part, "charset");
505 GMimeFilter *charset_filter;
506 charset_filter = g_mime_filter_charset_new (charset, "UTF-8");
507 /* This result can be NULL for things like "unknown-8bit".
508 * Don't set a NULL filter as that makes GMime print
509 * annoying assertion-failure messages on stderr. */
510 if (charset_filter) {
511 g_mime_stream_filter_add (GMIME_STREAM_FILTER (filter),
513 g_object_unref (charset_filter);
517 wrapper = g_mime_part_get_content (GMIME_PART (part));
519 g_mime_data_wrapper_write_to_stream (wrapper, filter);
521 g_object_unref (stream);
522 g_object_unref (filter);
523 g_object_unref (discard_non_term_filter);
525 g_byte_array_append (byte_array, (guint8 *) "\0", 1);
526 body = (char *) g_byte_array_free (byte_array, false);
529 _notmuch_message_gen_terms (message, NULL, body);
535 g_object_unref (repaired_part);
538 /* descend (if desired) into the cleartext part of an encrypted MIME
539 * part while indexing. */
541 _index_encrypted_mime_part (notmuch_message_t *message,
542 notmuch_indexopts_t *indexopts,
543 GMimeMultipartEncrypted *encrypted_data,
544 _notmuch_message_crypto_t *msg_crypto)
546 notmuch_status_t status;
548 notmuch_database_t *notmuch = NULL;
549 GMimeObject *clear = NULL;
551 if (! indexopts || (notmuch_indexopts_get_decrypt_policy (indexopts) == NOTMUCH_DECRYPT_FALSE))
554 notmuch = notmuch_message_get_database (message);
556 bool attempted = false;
557 GMimeDecryptResult *decrypt_result = NULL;
558 bool get_sk = (notmuch_indexopts_get_decrypt_policy (indexopts) == NOTMUCH_DECRYPT_TRUE);
559 clear = _notmuch_crypto_decrypt (&attempted, notmuch_indexopts_get_decrypt_policy (indexopts),
560 message, encrypted_data, get_sk ? &decrypt_result : NULL, &err);
563 if (err || ! clear) {
565 g_object_unref (decrypt_result);
567 _notmuch_database_log (notmuch, "Failed to decrypt during indexing. (%d:%d) [%s]\n",
568 err->domain, err->code, err->message);
571 _notmuch_database_log (notmuch, "Failed to decrypt during indexing. (unknown error)\n");
573 /* Indicate that we failed to decrypt during indexing */
574 status = notmuch_message_add_property (message, "index.decryption", "failure");
576 _notmuch_database_log_append (notmuch, "failed to add index.decryption "
577 "property (%d)\n", status);
580 if (decrypt_result) {
581 status = _notmuch_message_crypto_successful_decryption (msg_crypto);
583 _notmuch_database_log_append (notmuch, "failed to mark the message as decrypted (%s)\n",
584 notmuch_status_to_string (status));
586 status = notmuch_message_add_property (message, "session-key",
587 g_mime_decrypt_result_get_session_key (decrypt_result));
589 _notmuch_database_log (notmuch, "failed to add session-key "
590 "property (%d)\n", status);
592 g_object_unref (decrypt_result);
594 GMimeObject *toindex = clear;
595 if (_notmuch_message_crypto_potential_payload (msg_crypto, clear, GMIME_OBJECT (encrypted_data), GMIME_MULTIPART_ENCRYPTED_CONTENT) &&
596 msg_crypto->decryption_status == NOTMUCH_MESSAGE_DECRYPTED_FULL) {
597 toindex = _notmuch_repair_crypto_payload_skip_legacy_display (clear);
598 if (toindex != clear)
599 notmuch_message_add_property (message, "index.repaired", "skip-protected-headers-legacy-display");
601 _index_mime_part (message, indexopts, toindex, msg_crypto);
602 g_object_unref (clear);
604 status = notmuch_message_add_property (message, "index.decryption", "success");
606 _notmuch_database_log (notmuch, "failed to add index.decryption "
607 "property (%d)\n", status);
611 static notmuch_status_t
612 _notmuch_message_index_user_headers (notmuch_message_t *message, GMimeMessage *mime_message)
615 notmuch_database_t *notmuch = notmuch_message_get_database (message);
616 notmuch_string_map_iterator_t *iter = _notmuch_database_user_headers (notmuch);
618 for (; _notmuch_string_map_iterator_valid (iter);
619 _notmuch_string_map_iterator_move_to_next (iter)) {
621 const char *prefix_name = _notmuch_string_map_iterator_key (iter);
623 const char *header_name = _notmuch_string_map_iterator_value (iter);
625 const char *header = g_mime_object_get_header (GMIME_OBJECT (mime_message), header_name);
627 _notmuch_message_gen_terms (message, prefix_name, header);
631 _notmuch_string_map_iterator_destroy (iter);
632 return NOTMUCH_STATUS_SUCCESS;
637 _notmuch_message_index_file (notmuch_message_t *message,
638 notmuch_indexopts_t *indexopts,
639 notmuch_message_file_t *message_file)
641 GMimeMessage *mime_message;
642 InternetAddressList *addresses;
644 notmuch_status_t status;
645 _notmuch_message_crypto_t *msg_crypto;
647 status = _notmuch_message_file_get_mime_message (message_file,
652 addresses = g_mime_message_get_from (mime_message);
654 _index_address_list (message, "from", addresses);
657 addresses = g_mime_message_get_all_recipients (mime_message);
659 _index_address_list (message, "to", addresses);
660 g_object_unref (addresses);
663 subject = g_mime_message_get_subject (mime_message);
664 _notmuch_message_gen_terms (message, "subject", subject);
666 status = _notmuch_message_index_user_headers (message, mime_message);
668 msg_crypto = _notmuch_message_crypto_new (NULL);
669 _index_mime_part (message, indexopts, g_mime_message_get_mime_part (mime_message), msg_crypto);
670 if (msg_crypto && msg_crypto->payload_subject) {
671 _notmuch_message_gen_terms (message, "subject", msg_crypto->payload_subject);
672 _notmuch_message_update_subject (message, msg_crypto->payload_subject);
675 talloc_free (msg_crypto);
677 return NOTMUCH_STATUS_SUCCESS;