1 /* message.c - Utility functions for parsing an email message for notmuch.
3 * Copyright © 2009 Carl Worth
5 * This program is free software: you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation, either version 3 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License
16 * along with this program. If not, see https://www.gnu.org/licenses/ .
18 * Author: Carl Worth <cworth@cworth.org>
23 #include "notmuch-private.h"
25 #include <gmime/gmime.h>
27 #include <glib.h> /* GHashTable */
29 struct _notmuch_message_file {
30 /* open stream to (possibly gzipped) file */
34 /* Cache for decoded headers */
37 GMimeMessage *message;
41 _notmuch_message_file_destructor (notmuch_message_file_t *message)
44 g_hash_table_destroy (message->headers);
47 g_object_unref (message->message);
52 /* Create a new notmuch_message_file_t for 'filename' with 'ctx' as
53 * the talloc owner. */
54 notmuch_message_file_t *
55 _notmuch_message_file_open_ctx (notmuch_database_t *notmuch,
56 void *ctx, const char *filename)
58 notmuch_message_file_t *message;
60 message = talloc_zero (ctx, notmuch_message_file_t);
61 if (unlikely (message == NULL))
64 message->filename = talloc_strdup (message, filename);
65 if (message->filename == NULL)
68 talloc_set_destructor (message, _notmuch_message_file_destructor);
70 message->stream = g_mime_stream_gzfile_open (filename);
71 if (message->stream == NULL)
77 _notmuch_database_log (notmuch, "Error opening %s: %s\n",
78 filename, strerror (errno));
79 _notmuch_message_file_close (message);
84 notmuch_message_file_t *
85 _notmuch_message_file_open (notmuch_database_t *notmuch,
88 return _notmuch_message_file_open_ctx (notmuch, NULL, filename);
92 _notmuch_message_file_get_filename (notmuch_message_file_t *message_file)
94 return message_file->filename;
98 _notmuch_message_file_close (notmuch_message_file_t *message)
100 talloc_free (message);
104 _is_mbox (GMimeStream *stream)
110 if (g_mime_stream_read (stream, from_buf, sizeof (from_buf)) == sizeof(from_buf) &&
111 strncmp (from_buf, "From ", 5) == 0)
114 g_mime_stream_reset (stream);
120 _notmuch_message_file_parse (notmuch_message_file_t *message)
123 notmuch_status_t status = NOTMUCH_STATUS_SUCCESS;
124 static int initialized = 0;
127 if (message->message)
128 return NOTMUCH_STATUS_SUCCESS;
130 is_mbox = _is_mbox (message->stream);
137 message->headers = g_hash_table_new_full (strcase_hash, strcase_equal,
139 if (! message->headers)
140 return NOTMUCH_STATUS_OUT_OF_MEMORY;
142 parser = g_mime_parser_new_with_stream (message->stream);
143 g_mime_parser_set_scan_from (parser, is_mbox);
145 message->message = g_mime_parser_construct_message (parser, NULL);
146 if (! message->message) {
147 status = NOTMUCH_STATUS_FILE_NOT_EMAIL;
151 if (is_mbox && ! g_mime_parser_eos (parser)) {
153 * This is a multi-message mbox. (For historical reasons, we
154 * do support single-message mboxes.)
156 status = NOTMUCH_STATUS_FILE_NOT_EMAIL;
160 g_mime_stream_reset (message->stream);
161 g_object_unref (parser);
164 g_hash_table_destroy (message->headers);
165 message->headers = NULL;
167 if (message->message) {
168 g_object_unref (message->message);
169 message->message = NULL;
178 _notmuch_message_file_get_mime_message (notmuch_message_file_t *message,
179 GMimeMessage **mime_message)
181 notmuch_status_t status;
183 status = _notmuch_message_file_parse (message);
187 *mime_message = message->message;
189 return NOTMUCH_STATUS_SUCCESS;
193 * Get all instances of a header decoded and concatenated.
195 * The result must be freed using g_free().
197 * Return NULL on errors, empty string for non-existing headers.
201 _extend_header (char *combined, const char *value) {
204 decoded = g_mime_utils_header_decode_text (NULL, value);
214 char *tmp = g_strdup_printf ("%s %s", combined, decoded);
231 _notmuch_message_file_get_combined_header (notmuch_message_file_t *message,
234 char *combined = NULL;
235 GMimeHeaderList *headers;
237 headers = g_mime_object_get_header_list (GMIME_OBJECT (message->message));
242 for (int i=0; i < g_mime_header_list_get_count (headers); i++) {
244 GMimeHeader *g_header = g_mime_header_list_get_header_at (headers, i);
246 if (strcasecmp (g_mime_header_get_name (g_header), header) != 0)
249 /* GMime retains ownership of value, we hope */
250 value = g_mime_header_get_value (g_header);
252 combined = _extend_header (combined, value);
255 /* Return empty string for non-existing headers. */
257 combined = g_strdup ("");
263 _notmuch_message_file_get_header (notmuch_message_file_t *message,
269 if (_notmuch_message_file_parse (message))
272 /* If we have a cached decoded value, use it. */
273 value = g_hash_table_lookup (message->headers, header);
277 if (strcasecmp (header, "received") == 0) {
279 * The Received: header is special. We concatenate all
280 * instances of the header as we use this when analyzing the
281 * path the mail has taken from sender to recipient.
283 decoded = _notmuch_message_file_get_combined_header (message, header);
285 value = g_mime_object_get_header (GMIME_OBJECT (message->message),
288 decoded = g_mime_utils_header_decode_text (NULL, value);
290 decoded = g_strdup ("");
296 /* Cache the decoded value. We also own the strings. */
297 g_hash_table_insert (message->headers, xstrdup (header), decoded);
303 _notmuch_message_file_get_headers (notmuch_message_file_t *message_file,
304 const char **from_out,
305 const char **subject_out,
307 const char **date_out,
308 char **message_id_out)
310 notmuch_status_t ret;
312 const char *from, *to, *subject, *date;
313 char *message_id = NULL;
315 /* Parse message up front to get better error status. */
316 ret = _notmuch_message_file_parse (message_file);
320 /* Before we do any real work, (especially before doing a
321 * potential SHA-1 computation on the entire file's contents),
322 * let's make sure that what we're looking at looks like an
323 * actual email message.
325 from = _notmuch_message_file_get_header (message_file, "from");
326 subject = _notmuch_message_file_get_header (message_file, "subject");
327 to = _notmuch_message_file_get_header (message_file, "to");
328 date = _notmuch_message_file_get_header (message_file, "date");
330 if ((from == NULL || *from == '\0') &&
331 (subject == NULL || *subject == '\0') &&
332 (to == NULL || *to == '\0')) {
333 ret = NOTMUCH_STATUS_FILE_NOT_EMAIL;
337 /* Now that we're sure it's mail, the first order of business
338 * is to find a message ID (or else create one ourselves).
340 header = _notmuch_message_file_get_header (message_file, "message-id");
341 if (header && *header != '\0') {
342 message_id = _notmuch_message_id_parse (message_file, header, NULL);
344 /* So the header value isn't RFC-compliant, but it's
345 * better than no message-id at all.
347 if (message_id == NULL)
348 message_id = talloc_strdup (message_file, header);
351 if (message_id == NULL ) {
352 /* No message-id at all, let's generate one by taking a
353 * hash over the file's contents.
355 char *sha1 = _notmuch_sha1_of_file (_notmuch_message_file_get_filename (message_file));
357 /* If that failed too, something is really wrong. Give up. */
359 ret = NOTMUCH_STATUS_FILE_ERROR;
363 message_id = talloc_asprintf (message_file, "notmuch-sha1-%s", sha1);
367 if (ret == NOTMUCH_STATUS_SUCCESS) {
371 *subject_out = subject;
377 *message_id_out = message_id;