annotate src/plugins/fts-solr/fts-backend-solr.c @ 9604:cf0da2cd31fb HEAD

fts-solr: Replace characters not valid for XML with replacement char.
author Timo Sirainen <tss@iki.fi>
date Fri, 20 Aug 2010 20:38:26 +0100
parents 00cd9aacd03c
children ac0855b1bbea
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
9532
00cd9aacd03c Updated copyright notices to include year 2010.
Timo Sirainen <tss@iki.fi>
parents: 9292
diff changeset
1 /* Copyright (c) 2006-2010 Dovecot authors, see the included COPYING file */
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
2
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
3 #include "lib.h"
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
4 #include "array.h"
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
5 #include "str.h"
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
6 #include "strescape.h"
9604
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
7 #include "unichar.h"
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
8 #include "mail-storage-private.h"
8109
e7929190cd32 fts-solr: Fixed compiling with recent struct mail_user changes.
Timo Sirainen <tss@iki.fi>
parents: 8006
diff changeset
9 #include "mail-namespace.h"
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
10 #include "solr-connection.h"
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
11 #include "fts-solr-plugin.h"
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
12
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
13 #include <ctype.h>
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
14
8001
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
15 #define SOLR_CMDBUF_SIZE (1024*64)
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
16 #define SOLR_MAX_ROWS 100000
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
17 #define FTS_SOLR_MAX_BOX_INC_PATTERNS 5
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
18 #define FTS_SOLR_MAX_BOX_EXC_PATTERNS 5
8001
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
19
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
20 struct solr_fts_backend {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
21 struct fts_backend backend;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
22 char *id_username, *id_namespace, *id_box_name;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
23 struct mail_namespace *default_ns;
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
24 };
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
25
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
26 struct solr_fts_backend_build_context {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
27 struct fts_backend_build_context ctx;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
28
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
29 struct solr_connection_post *post;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
30 uint32_t prev_uid, uid_validity;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
31 string_t *cmd;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
32 bool headers;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
33 };
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
34
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
35 struct solr_virtual_uid_map_context {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
36 struct fts_backend *backend;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
37 struct mailbox *box;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
38 string_t *vname;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
39 };
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
40
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
41 struct fts_backend_solr_get_last_uids_context {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
42 struct fts_backend *backend;
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
43 pool_t pool;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
44 ARRAY_TYPE(fts_backend_uid_map) *last_uids;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
45
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
46 struct mailbox *box;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
47 string_t *vname;
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
48 };
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
49
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
50 static struct solr_connection *solr_conn = NULL;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
51
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
52 static void fts_box_name_get_root(struct mail_namespace **ns, const char **name)
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
53 {
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
54 struct mail_namespace *orig_ns = *ns;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
55
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
56 while ((*ns)->alias_for != NULL)
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
57 *ns = (*ns)->alias_for;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
58
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
59 if (**name == '\0' && *ns != orig_ns &&
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
60 ((*ns)->flags & NAMESPACE_FLAG_INBOX) != 0) {
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
61 /* ugly workaround to allow selecting INBOX from a Maildir/
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
62 when it's not in the inbox=yes namespace. */
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
63 *name = "INBOX";
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
64 }
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
65 }
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
66
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
67 static const char *
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
68 fts_box_get_root(struct mailbox *box, struct mail_namespace **ns_r)
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
69 {
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
70 struct mail_namespace *ns = box->storage->ns;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
71 const char *name = box->name;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
72
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
73 fts_box_name_get_root(&ns, &name);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
74 *ns_r = ns;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
75 return name;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
76 }
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
77
9604
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
78 static bool is_valid_xml_char(unichar_t chr)
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
79 {
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
80 /* Valid characters in XML:
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
81
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
82 #x9 | #xA | #xD | [#x20-#xD7FF] | [#xE000-#xFFFD] |
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
83 [#x10000-#x10FFFF]
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
84
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
85 This function gets called only for #x80 and higher */
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
86 if (chr > 0xd7ff && chr < 0xe000)
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
87 return FALSE;
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
88 if (chr > 0xfffd && chr < 0x10000)
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
89 return FALSE;
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
90 return chr < 0x10ffff;
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
91 }
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
92
8000
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
93 static void
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
94 xml_encode_data(string_t *dest, const unsigned char *data, unsigned int len)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
95 {
9604
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
96 unichar_t chr;
8000
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
97 unsigned int i;
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
98
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
99 for (i = 0; i < len; i++) {
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
100 switch (data[i]) {
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
101 case '&':
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
102 str_append(dest, "&amp;");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
103 break;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
104 case '<':
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
105 str_append(dest, "&lt;");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
106 break;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
107 case '>':
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
108 str_append(dest, "&gt;");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
109 break;
8848
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
110 case '\t':
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
111 case '\n':
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
112 case '\r':
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
113 /* exceptions to the following control char check */
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
114 str_append_c(dest, data[i]);
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
115 break;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
116 default:
8848
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
117 if (data[i] < 32) {
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
118 /* SOLR doesn't like control characters.
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
119 replace them with spaces. */
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
120 str_append_c(dest, ' ');
9604
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
121 } else if (data[i] >= 0x80) {
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
122 /* make sure the character is valid for XML
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
123 so we don't get XML parser errors */
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
124 unsigned int char_len =
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
125 uni_utf8_char_bytes(data[0]);
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
126 if (i + char_len <= len &&
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
127 uni_utf8_get_char_n(data, len, &chr) == 0 &&
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
128 is_valid_xml_char(chr))
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
129 str_append_n(dest, data + i, char_len);
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
130 else {
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
131 str_append_n(dest, utf8_replacement_char,
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
132 UTF8_REPLACEMENT_CHAR_LEN);
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
133 }
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
134 i += char_len - 1;
8848
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
135 } else {
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
136 str_append_c(dest, data[i]);
dd441977c578 fts-solr: Solr breaks when it sees control characters, so replace them with spaces.
Timo Sirainen <tss@iki.fi>
parents: 8590
diff changeset
137 }
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
138 break;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
139 }
9604
cf0da2cd31fb fts-solr: Replace characters not valid for XML with replacement char.
Timo Sirainen <tss@iki.fi>
parents: 9532
diff changeset
140 i += uni_utf8_char_bytes(data[0]);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
141 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
142 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
143
8000
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
144 static void xml_encode(string_t *dest, const char *str)
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
145 {
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
146 xml_encode_data(dest, (const unsigned char *)str, strlen(str));
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
147 }
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
148
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
149 static const char *solr_escape_id_str(const char *str)
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
150 {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
151 string_t *tmp;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
152 const char *p;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
153
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
154 for (p = str; *p != '\0'; p++) {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
155 if (*p == '/' || *p == '!')
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
156 break;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
157 }
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
158 if (*p == '\0')
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
159 return str;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
160
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
161 tmp = t_str_new(64);
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
162 for (p = str; *p != '\0'; p++) {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
163 switch (*p) {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
164 case '/':
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
165 str_append(tmp, "!\\");
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
166 break;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
167 case '!':
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
168 str_append(tmp, "!!");
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
169 break;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
170 default:
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
171 str_append_c(tmp, *p);
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
172 break;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
173 }
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
174 }
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
175 return str_c(tmp);
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
176 }
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
177
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
178 static void solr_quote(string_t *dest, const char *str)
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
179 {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
180 str_append_c(dest, '"');
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
181 str_append(dest, str_escape(str));
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
182 str_append_c(dest, '"');
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
183 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
184
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
185 static void solr_quote_http(string_t *dest, const char *str)
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
186 {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
187 str_append(dest, "%22");
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
188 solr_connection_http_escape(solr_conn, dest, str);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
189 str_append(dest, "%22");
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
190 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
191
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
192 static struct fts_backend *
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
193 fts_backend_solr_init(struct mailbox *box)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
194 {
8002
e814614ffb65 fts-solr: Added break-imap-search setting to use Solr with SEARCH TEXT/BODY
Timo Sirainen <tss@iki.fi>
parents: 8001
diff changeset
195 const struct fts_solr_settings *set = &fts_solr_settings;
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
196 struct solr_fts_backend *backend;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
197 struct mail_namespace *ns;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
198 const char *str, *box_name;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
199
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
200
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
201 box_name = fts_box_get_root(box, &ns);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
202 i_assert(*box_name != '\0');
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
203
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
204 if (solr_conn == NULL)
8002
e814614ffb65 fts-solr: Added break-imap-search setting to use Solr with SEARCH TEXT/BODY
Timo Sirainen <tss@iki.fi>
parents: 8001
diff changeset
205 solr_conn = solr_connection_init(set->url, set->debug);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
206
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
207 backend = i_new(struct solr_fts_backend, 1);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
208 str = fts_solr_settings.default_ns_prefix;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
209 if (str != NULL) {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
210 backend->default_ns =
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
211 mail_namespace_find_prefix(ns->user->namespaces, str);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
212 if (backend->default_ns == NULL) {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
213 i_fatal("fts_solr: default_ns setting points to "
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
214 "nonexisting namespace");
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
215 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
216 } else {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
217 backend->default_ns =
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
218 mail_namespace_find_inbox(ns->user->namespaces);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
219 }
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
220 while (backend->default_ns->alias_for != NULL)
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
221 backend->default_ns = backend->default_ns->alias_for;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
222
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
223 str = solr_escape_id_str(ns->user->username);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
224 backend->id_username = i_strdup(str);
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
225 if (ns != backend->default_ns) {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
226 str = solr_escape_id_str(ns->prefix);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
227 backend->id_namespace = i_strdup(str);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
228 }
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
229 backend->id_box_name = i_strdup(box_name);
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
230 backend->backend = fts_backend_solr;
8002
e814614ffb65 fts-solr: Added break-imap-search setting to use Solr with SEARCH TEXT/BODY
Timo Sirainen <tss@iki.fi>
parents: 8001
diff changeset
231
e814614ffb65 fts-solr: Added break-imap-search setting to use Solr with SEARCH TEXT/BODY
Timo Sirainen <tss@iki.fi>
parents: 8001
diff changeset
232 if (set->substring_search)
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
233 backend->backend.flags |= FTS_BACKEND_FLAG_SUBSTRING_LOOKUPS;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
234 return &backend->backend;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
235 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
236
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
237 static void fts_backend_solr_deinit(struct fts_backend *_backend)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
238 {
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
239 struct solr_fts_backend *backend = (struct solr_fts_backend *)_backend;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
240
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
241 i_free(backend->id_box_name);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
242 i_free(backend->id_namespace);
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
243 i_free(backend->id_username);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
244 i_free(backend);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
245 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
246
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
247 static void
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
248 solr_add_ns_query(string_t *str, struct fts_backend *_backend,
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
249 struct mail_namespace *ns, bool neg)
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
250 {
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
251 struct solr_fts_backend *backend = (struct solr_fts_backend *)_backend;
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
252
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
253 while (ns->alias_for != NULL)
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
254 ns = ns->alias_for;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
255
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
256 if (ns == backend->default_ns || *ns->prefix == '\0') {
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
257 if (!neg)
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
258 str_append(str, " -ns:[* TO *]");
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
259 else
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
260 str_append(str, " +ns:[* TO *]");
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
261 } else {
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
262 if (!neg)
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
263 str_append(str, " +ns:");
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
264 else
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
265 str_append(str, " -ns:");
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
266 solr_quote(str, ns->prefix);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
267 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
268 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
269
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
270 static void
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
271 solr_add_ns_query_http(string_t *str, struct fts_backend *backend,
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
272 struct mail_namespace *ns)
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
273 {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
274 string_t *tmp;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
275
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
276 tmp = t_str_new(64);
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
277 solr_add_ns_query(tmp, backend, ns, FALSE);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
278 solr_connection_http_escape(solr_conn, str, str_c(tmp));
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
279 }
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
280
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
281 static int fts_backend_solr_get_last_uid_fallback(struct fts_backend *backend,
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
282 uint32_t *last_uid_r)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
283 {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
284 struct mailbox *box = backend->box;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
285 struct mail_namespace *ns;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
286 struct mailbox_status status;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
287 ARRAY_TYPE(seq_range) uids;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
288 const struct seq_range *uidvals;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
289 const char *box_name;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
290 unsigned int count;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
291 string_t *str;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
292
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
293 str = t_str_new(256);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
294 str_append(str, "fl=uid&rows=1&sort=uid+desc&q=");
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
295
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
296 box_name = fts_box_get_root(box, &ns);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
297
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
298 mailbox_get_status(box, STATUS_UIDVALIDITY, &status);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
299 str_printfa(str, "uidv:%u+box:", status.uidvalidity);
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
300 solr_quote_http(str, box_name);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
301 solr_add_ns_query_http(str, backend, ns);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
302 str_append(str, "+user:");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
303 solr_quote_http(str, ns->user->username);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
304
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
305 t_array_init(&uids, 1);
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
306 if (solr_connection_select(solr_conn, str_c(str),
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
307 NULL, NULL, &uids, NULL) < 0)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
308 return -1;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
309
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
310 uidvals = array_get(&uids, &count);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
311 if (count == 0) {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
312 /* nothing indexed yet for this mailbox */
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
313 *last_uid_r = 0;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
314 } else if (count == 1 && uidvals[0].seq1 == uidvals[0].seq2) {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
315 *last_uid_r = uidvals[0].seq1;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
316 } else {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
317 i_error("fts_solr: Last UID lookup returned multiple rows");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
318 return -1;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
319 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
320 return 0;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
321 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
322
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
323 static int fts_backend_solr_get_last_uid(struct fts_backend *backend,
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
324 uint32_t *last_uid_r)
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
325 {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
326 struct mailbox *box = backend->box;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
327 struct mail_namespace *ns;
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
328 struct mailbox_status status;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
329 ARRAY_TYPE(seq_range) uids;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
330 const struct seq_range *uidvals;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
331 const char *box_name;
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
332 unsigned int count;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
333 string_t *str;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
334
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
335 str = t_str_new(256);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
336 str_append(str, "fl=uid&rows=1&q=last_uid:TRUE+");
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
337
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
338 box_name = fts_box_get_root(box, &ns);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
339
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
340 mailbox_get_status(box, STATUS_UIDVALIDITY, &status);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
341 str_printfa(str, "uidv:%u+box:", status.uidvalidity);
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
342 solr_quote_http(str, box_name);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
343 solr_add_ns_query_http(str, backend, ns);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
344 str_append(str, "+user:");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
345 solr_quote_http(str, ns->user->username);
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
346
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
347 t_array_init(&uids, 1);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
348 if (solr_connection_select(solr_conn, str_c(str),
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
349 NULL, NULL, &uids, NULL) < 0)
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
350 return -1;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
351
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
352 uidvals = array_get(&uids, &count);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
353 if (count == 0) {
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
354 /* either nothing is indexed or we're converting from an
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
355 older database format without the last_uid fields */
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
356 return fts_backend_solr_get_last_uid_fallback(backend,
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
357 last_uid_r);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
358 } else if (count == 1 && uidvals[0].seq1 == uidvals[0].seq2) {
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
359 *last_uid_r = uidvals[0].seq1;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
360 } else {
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
361 i_error("fts_solr: Last UID lookup returned multiple rows");
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
362 return -1;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
363 }
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
364 return 0;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
365 }
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
366
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
367 static struct mail_namespace *
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
368 solr_get_namespaces(struct fts_backend *_backend,
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
369 struct mailbox *box, const char *ns_prefix)
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
370 {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
371 struct solr_fts_backend *backend = (struct solr_fts_backend *)_backend;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
372 struct mail_namespace *namespaces = box->storage->ns->user->namespaces;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
373
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
374 if (ns_prefix == NULL)
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
375 return backend->default_ns;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
376 else
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
377 return mail_namespace_find_prefix(namespaces, ns_prefix);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
378 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
379
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
380 static bool
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
381 solr_virtual_get_last_uids(const char *ns_prefix, const char *mailbox,
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
382 uint32_t uidvalidity, uint32_t *uid, void *context)
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
383 {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
384 struct fts_backend_solr_get_last_uids_context *ctx = context;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
385 struct fts_backend_uid_map *map;
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
386 struct mail_namespace *ns;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
387 const char *vname;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
388
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
389 ns = solr_get_namespaces(ctx->backend, ctx->box, ns_prefix);
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
390 for (; ns != NULL; ns = ns->alias_chain_next) {
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
391 vname = mail_namespace_get_vname(ns, ctx->vname, mailbox);
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
392 map = array_append_space(ctx->last_uids);
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
393 map->mailbox = p_strdup(ctx->pool, vname);
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
394 map->uidvalidity = uidvalidity;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
395 map->uid = *uid;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
396 }
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
397 return FALSE;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
398 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
399
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
400 static void
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
401 solr_add_pattern(string_t *str, const struct mailbox_virtual_pattern *pattern)
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
402 {
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
403 struct mail_namespace *ns = pattern->ns;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
404 const char *name, *p;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
405
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
406 name = pattern->pattern;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
407 if (!mail_namespace_update_name(pattern->ns, &name))
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
408 name = mail_namespace_fix_sep(pattern->ns, name);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
409
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
410 fts_box_name_get_root(&ns, &name);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
411
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
412 if (strcmp(name, "*") == 0) {
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
413 str_append(str, "[* TO *]");
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
414 return;
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
415 }
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
416
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
417 /* first check if there are any wildcards in the pattern */
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
418 for (p = name; *p != '\0'; p++) {
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
419 if (*p == '%' || *p == '*')
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
420 break;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
421 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
422 if (*p == '\0') {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
423 /* full mailbox name */
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
424 solr_quote(str, name);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
425 return;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
426 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
427
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
428 /* there are at least some wildcards. */
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
429 for (p = name; *p != '\0'; p++) {
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
430 if (*p == '%' || *p == '*') {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
431 if (p == name || (p[-1] != '%' && p[-1] != '*'))
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
432 str_append_c(str, '*');
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
433 } else {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
434 if (!i_isalnum(*p))
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
435 str_append_c(str, '\\');
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
436 str_append_c(str, *p);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
437 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
438 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
439 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
440
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
441 static void
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
442 fts_backend_solr_filter_mailboxes(struct fts_backend *_backend,
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
443 string_t *str, struct mailbox *box)
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
444 {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
445 ARRAY_TYPE(mailbox_virtual_patterns) includes_arr, excludes_arr;
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
446 struct mail_namespace *ns;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
447 const struct mailbox_virtual_pattern *includes, *excludes;
8519
f456043e10d3 fts-solr: Minor bugfix
Timo Sirainen <tss@iki.fi>
parents: 8505
diff changeset
448 unsigned int i, inc_count, exc_count;
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
449 string_t *fq;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
450
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
451 t_array_init(&includes_arr, 16);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
452 t_array_init(&excludes_arr, 16);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
453 mailbox_get_virtual_box_patterns(box, &includes_arr, &excludes_arr);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
454 includes = array_get(&includes_arr, &inc_count);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
455 excludes = array_get(&excludes_arr, &exc_count);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
456 i_assert(inc_count > 0);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
457
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
458 /* First see if there are any patterns that begin with a wildcard.
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
459 Solr doesn't allow them, so in that case we'll need to return
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
460 all mailboxes. */
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
461 for (i = 0; i < inc_count; i++) {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
462 if (*includes[i].pattern == '*' ||
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
463 *includes[i].pattern == '%')
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
464 break;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
465 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
466
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
467 fq = t_str_new(128);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
468 if (i == inc_count && inc_count <= FTS_SOLR_MAX_BOX_INC_PATTERNS) {
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
469 /* we can filter what mailboxes we want returned */
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
470 str_append_c(fq, '(');
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
471 for (i = 0; i < inc_count; i++) {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
472 if (i != 0)
8879
e81dfb69bfc5 fts-solr: Don't double-escape "+" if we're filtering what mailboxes we want.
Timo Sirainen <tss@iki.fi>
parents: 8848
diff changeset
473 str_append(fq, " OR +");
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
474 str_append_c(fq, '(');
8879
e81dfb69bfc5 fts-solr: Don't double-escape "+" if we're filtering what mailboxes we want.
Timo Sirainen <tss@iki.fi>
parents: 8848
diff changeset
475 str_append(fq, "+box:");
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
476 solr_add_pattern(fq, &includes[i]);
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
477 solr_add_ns_query(fq, _backend, includes[i].ns, FALSE);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
478 str_append_c(fq, ')');
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
479 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
480 str_append_c(fq, ')');
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
481 }
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
482 exc_count = I_MIN(FTS_SOLR_MAX_BOX_EXC_PATTERNS, exc_count);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
483 for (i = 0; i < exc_count; i++) {
8519
f456043e10d3 fts-solr: Minor bugfix
Timo Sirainen <tss@iki.fi>
parents: 8505
diff changeset
484 if (str_len(fq) > 0)
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
485 str_append_c(fq, ' ');
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
486 str_append(fq, "NOT (");
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
487 str_append(fq, "box:");
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
488 solr_add_pattern(fq, &excludes[i]);
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
489
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
490 for (ns = excludes[i].ns; ns->alias_for != NULL; )
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
491 ns = ns->alias_for;
8945
43b68bd6023a fts-solr: Fixes with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8879
diff changeset
492 solr_add_ns_query(fq, _backend, ns, FALSE);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
493 str_append_c(fq, ')');
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
494 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
495 if (str_len(fq) > 0) {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
496 str_append(str, "&fq=");
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
497 solr_connection_http_escape(solr_conn, str, str_c(fq));
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
498 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
499 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
500
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
501 static int
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
502 fts_backend_solr_get_all_last_uids(struct fts_backend *backend, pool_t pool,
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
503 ARRAY_TYPE(fts_backend_uid_map) *last_uids)
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
504 {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
505 struct fts_backend_solr_get_last_uids_context ctx;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
506 string_t *str;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
507
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
508 memset(&ctx, 0, sizeof(ctx));
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
509 ctx.backend = backend;
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
510 ctx.pool = pool;
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
511 ctx.last_uids = last_uids;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
512 ctx.box = backend->box;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
513 ctx.vname = t_str_new(256);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
514
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
515 str = t_str_new(256);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
516 str_printfa(str, "fl=uid,box,uidv,ns&rows=%u&q=last_uid:TRUE+user:",
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
517 SOLR_MAX_ROWS);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
518 solr_quote_http(str, backend->box->storage->ns->user->username);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
519 fts_backend_solr_filter_mailboxes(backend, str, backend->box);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
520
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
521 return solr_connection_select(solr_conn, str_c(str),
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
522 solr_virtual_get_last_uids, &ctx,
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
523 NULL, NULL);
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
524 }
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
525
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
526 static int
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
527 fts_backend_solr_build_init(struct fts_backend *backend, uint32_t *last_uid_r,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
528 struct fts_backend_build_context **ctx_r)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
529 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
530 struct solr_fts_backend_build_context *ctx;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
531 struct mailbox_status status;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
532
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
533 *last_uid_r = (uint32_t)-1;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
534
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
535 ctx = i_new(struct solr_fts_backend_build_context, 1);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
536 ctx->ctx.backend = backend;
8001
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
537 ctx->cmd = str_new(default_pool, SOLR_CMDBUF_SIZE);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
538
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
539 mailbox_get_status(backend->box, STATUS_UIDVALIDITY, &status);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
540 ctx->uid_validity = status.uidvalidity;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
541
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
542 *ctx_r = &ctx->ctx;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
543 return 0;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
544 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
545
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
546 static void
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
547 fts_backend_solr_add_doc_prefix(struct solr_fts_backend_build_context *ctx,
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
548 uint32_t uid)
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
549 {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
550 struct solr_fts_backend *backend =
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
551 (struct solr_fts_backend *)ctx->ctx.backend;
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
552 struct mailbox *box = ctx->ctx.backend->box;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
553 struct mail_namespace *ns;
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
554 const char *box_name;
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
555
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
556 str_printfa(ctx->cmd, "<doc>"
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
557 "<field name=\"uid\">%u</field>"
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
558 "<field name=\"uidv\">%u</field>",
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
559 uid, ctx->uid_validity);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
560
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
561 box_name = fts_box_get_root(box, &ns);
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
562
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
563 if (ns != backend->default_ns) {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
564 str_append(ctx->cmd, "<field name=\"ns\">");
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
565 xml_encode(ctx->cmd, ns->prefix);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
566 str_append(ctx->cmd, "</field>");
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
567 }
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
568 str_append(ctx->cmd, "<field name=\"box\">");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
569 xml_encode(ctx->cmd, box_name);
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
570 str_append(ctx->cmd, "</field><field name=\"user\">");
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
571 xml_encode(ctx->cmd, ns->user->username);
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
572 str_append(ctx->cmd, "</field>");
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
573 }
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
574
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
575 static void xml_encode_id(string_t *str, struct fts_backend *_backend,
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
576 uint32_t uid, uint32_t uid_validity)
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
577 {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
578 struct solr_fts_backend *backend = (struct solr_fts_backend *)_backend;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
579
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
580 if (uid != 0)
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
581 str_printfa(str, "%u/", uid);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
582 else
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
583 str_append(str, "L/");
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
584 if (backend->id_namespace != NULL) {
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
585 xml_encode(str, backend->id_namespace);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
586 str_append_c(str, '/');
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
587 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
588 str_printfa(str, "%u/", uid_validity);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
589 xml_encode(str, backend->id_username);
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
590 str_append_c(str, '/');
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
591 xml_encode(str, backend->id_box_name);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
592 }
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
593
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
594 static int
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
595 fts_backend_solr_build_more(struct fts_backend_build_context *_ctx,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
596 uint32_t uid, const unsigned char *data,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
597 size_t size, bool headers)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
598 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
599 struct solr_fts_backend_build_context *ctx =
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
600 (struct solr_fts_backend_build_context *)_ctx;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
601 string_t *cmd = ctx->cmd;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
602
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
603 /* body comes first, then headers */
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
604 if (ctx->prev_uid != uid) {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
605 /* uid changed */
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
606 if (ctx->post == NULL) {
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
607 ctx->post = solr_connection_post_begin(solr_conn);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
608 str_append(cmd, "<add>");
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
609 } else {
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
610 str_append(cmd, "</field></doc>");
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
611 }
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
612 ctx->prev_uid = uid;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
613
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
614 fts_backend_solr_add_doc_prefix(ctx, uid);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
615 str_printfa(cmd, "<field name=\"id\">");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
616 xml_encode_id(cmd, _ctx->backend, uid, ctx->uid_validity);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
617 str_append(cmd, "</field>");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
618
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
619 ctx->headers = headers;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
620 if (headers) {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
621 str_append(cmd, "<field name=\"hdr\">");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
622 } else {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
623 str_append(cmd, "<field name=\"body\">");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
624 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
625 } else if (headers && !ctx->headers) {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
626 str_append(cmd, "</field><field name=\"hdr\">");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
627 } else {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
628 i_assert(!(!headers && ctx->headers));
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
629 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
630
8000
61be6a4722c5 fts-solr: Encode entities when indexing data.
Timo Sirainen <tss@iki.fi>
parents: 7999
diff changeset
631 xml_encode_data(cmd, data, size);
8001
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
632 if (str_len(cmd) > SOLR_CMDBUF_SIZE-128) {
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
633 solr_connection_post_more(ctx->post, str_data(cmd),
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
634 str_len(cmd));
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
635 str_truncate(cmd, 0);
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
636 }
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
637 return 0;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
638 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
639
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
640 static int
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
641 fts_backed_solr_build_commit(struct solr_fts_backend_build_context *ctx)
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
642 {
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
643 int ret;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
644
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
645 if (ctx->post == NULL)
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
646 return 0;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
647
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
648 str_append(ctx->cmd, "</field></doc>");
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
649
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
650 /* Update the mailbox's last_uid field, replacing the existing
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
651 document. Note that since there is no locking, it's possible that
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
652 if another session is indexing at the same time, the last_uid value
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
653 may shrink. This doesn't really matter, we'll simply do more work
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
654 in future by reindexing some messages. */
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
655 fts_backend_solr_add_doc_prefix(ctx, ctx->prev_uid);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
656 str_printfa(ctx->cmd, "<field name=\"last_uid\">TRUE</field>"
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
657 "<field name=\"id\">");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
658 xml_encode_id(ctx->cmd, ctx->ctx.backend, 0, ctx->uid_validity);
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
659 str_append(ctx->cmd, "</field></doc></add>");
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
660
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
661 solr_connection_post_more(ctx->post, str_data(ctx->cmd),
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
662 str_len(ctx->cmd));
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
663 ret = solr_connection_post_end(ctx->post);
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
664 /* commit and wait until the documents we just indexed are
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
665 visible to the following search */
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
666 if (solr_connection_post(solr_conn, "<commit waitFlush=\"false\" "
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
667 "waitSearcher=\"true\"/>") < 0)
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
668 ret = -1;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
669 return ret;
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
670 }
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
671
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
672 static int
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
673 fts_backend_solr_build_deinit(struct fts_backend_build_context *_ctx)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
674 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
675 struct solr_fts_backend_build_context *ctx =
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
676 (struct solr_fts_backend_build_context *)_ctx;
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
677 int ret;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
678
8492
ee1ecdda8a6b fts-solr: Keep last_uid field updated containing the mailbox's last added UID.
Timo Sirainen <tss@iki.fi>
parents: 8484
diff changeset
679 ret = fts_backed_solr_build_commit(ctx);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
680 str_free(&ctx->cmd);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
681 i_free(ctx);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
682 return ret;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
683 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
684
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
685 static void
8493
149c6ccc3df4 fts-solr: Use '!' as escape character in usernames. Escape '/' and '!'.
Timo Sirainen <tss@iki.fi>
parents: 8492
diff changeset
686 fts_backend_solr_expunge(struct fts_backend *backend, struct mail *mail)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
687 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
688 struct mailbox_status status;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
689
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
690 mailbox_get_status(mail->box, STATUS_UIDVALIDITY, &status);
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
691
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
692 T_BEGIN {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
693 string_t *cmd;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
694
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
695 cmd = t_str_new(256);
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
696 str_append(cmd, "<delete><id>");
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
697 xml_encode_id(cmd, backend, mail->uid, status.uidvalidity);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
698 str_append(cmd, "</id></delete>");
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
699
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
700 (void)solr_connection_post(solr_conn, str_c(cmd));
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
701 } T_END;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
702 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
703
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
704 static void
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
705 fts_backend_solr_expunge_finish(struct fts_backend *backend ATTR_UNUSED,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
706 struct mailbox *box ATTR_UNUSED,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
707 bool committed ATTR_UNUSED)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
708 {
8001
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
709 solr_connection_post(solr_conn,
3fbfdda3e5d3 fts-solr: Send data to Solr in larger packets. Commit after expunges.
Timo Sirainen <tss@iki.fi>
parents: 8000
diff changeset
710 "<commit waitFlush=\"false\" waitSearcher=\"false\"/>");
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
711 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
712
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
713 static int fts_backend_solr_lock(struct fts_backend *backend ATTR_UNUSED)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
714 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
715 return 1;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
716 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
717
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
718 static void fts_backend_solr_unlock(struct fts_backend *backend ATTR_UNUSED)
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
719 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
720 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
721
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
722 static bool solr_virtual_uid_map(const char *ns_prefix, const char *mailbox,
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
723 uint32_t uidvalidity, uint32_t *uid,
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
724 void *context)
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
725 {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
726 struct solr_virtual_uid_map_context *ctx = context;
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
727 struct mail_namespace *ns;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
728 const char *vname;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
729 bool convert_inbox;
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
730
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
731 ns = solr_get_namespaces(ctx->backend, ctx->box, ns_prefix);
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
732 convert_inbox = (ns->flags & NAMESPACE_FLAG_INBOX) != 0 &&
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
733 strcmp(mailbox, "INBOX") == 0;
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
734 for (; ns != NULL; ns = ns->alias_chain_next) {
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
735 vname = convert_inbox ? ns->prefix :
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
736 mail_namespace_get_vname(ns, ctx->vname, mailbox);
8500
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
737 if (mailbox_get_virtual_uid(ctx->box, vname, uidvalidity,
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
738 *uid, uid))
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
739 return TRUE;
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
740 }
3efcdc45d111 Added alias_for setting for namespaces. Fixes namespace issues with fts.
Timo Sirainen <tss@iki.fi>
parents: 8499
diff changeset
741 return FALSE;
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
742 }
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
743
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
744 static int fts_backend_solr_lookup(struct fts_backend_lookup_context *ctx,
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
745 ARRAY_TYPE(seq_range) *definite_uids,
8006
0a3b5ddb0869 Added support for sorting by X-SCORE. Currently it's only set by fts-solr.
Timo Sirainen <tss@iki.fi>
parents: 8004
diff changeset
746 ARRAY_TYPE(seq_range) *maybe_uids,
0a3b5ddb0869 Added support for sorting by X-SCORE. Currently it's only set by fts-solr.
Timo Sirainen <tss@iki.fi>
parents: 8004
diff changeset
747 ARRAY_TYPE(fts_score_map) *scores)
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
748 {
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
749 struct mailbox *box = ctx->backend->box;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
750 struct mail_namespace *ns;
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
751 struct solr_virtual_uid_map_context uid_map_ctx;
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
752 const struct fts_backend_lookup_field *fields;
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
753 const char *box_name;
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
754 unsigned int i, count;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
755 struct mailbox_status status;
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
756 string_t *str;
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
757 bool virtual;
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
758
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
759 virtual = strcmp(box->storage->name, "virtual") == 0;
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
760 mailbox_get_status(box, STATUS_UIDVALIDITY, &status);
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
761
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
762 str = t_str_new(256);
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
763 if (!virtual) {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
764 str_printfa(str, "fl=uid,score&rows=%u&sort=uid+asc&q=",
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
765 status.uidnext);
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
766 } else {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
767 str_printfa(str, "fl=uid,score,box,uidv,ns&rows=%u"
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
768 "&sort=box+asc,uid+asc&q=",
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
769 SOLR_MAX_ROWS);
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
770 }
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
771
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
772 /* build a lucene search query from the fields */
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
773 fields = array_get(&ctx->fields, &count);
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
774 for (i = 0; i < count; i++) {
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
775 if (i > 0)
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
776 str_append_c(str, '+');
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
777
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
778 if ((fields[i].flags & FTS_LOOKUP_FLAG_INVERT) != 0)
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
779 str_append_c(str, '-');
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
780
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
781 if ((fields[i].flags & FTS_LOOKUP_FLAG_HEADER) == 0) {
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
782 /* body only */
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
783 i_assert((fields[i].flags & FTS_LOOKUP_FLAG_BODY) != 0);
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
784 str_append(str, "body:");
9141
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
785 solr_quote_http(str, fields[i].key);
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
786 } else if ((fields[i].flags & FTS_LOOKUP_FLAG_BODY) == 0) {
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
787 /* header only */
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
788 str_append(str, "hdr:");
9141
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
789 solr_quote_http(str, fields[i].key);
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
790 } else {
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
791 /* both */
9141
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
792 str_append(str, "(body:");
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
793 solr_quote_http(str, fields[i].key);
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
794 str_append(str, "+OR+hdr:");
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
795 solr_quote_http(str, fields[i].key);
d21bd1a1710a solr: Don't use "any" copyfield, it doubles the index size.
Timo Sirainen <tss@iki.fi>
parents: 8950
diff changeset
796 str_append_c(str, ')');
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
797 }
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
798 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
799
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
800 /* use a separate filter query for selecting the mailbox. it shouldn't
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
801 affect the score and there could be some caching benefits too. */
8505
5cc1c4f3d38b fts-solr: In filter queries use "+" prefixes to require all the parameters to match.
Timo Sirainen <tss@iki.fi>
parents: 8500
diff changeset
802 str_append(str, "&fq=%2Buser:");
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
803 solr_quote_http(str, box->storage->ns->user->username);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
804 if (virtual)
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
805 fts_backend_solr_filter_mailboxes(ctx->backend, str, box);
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
806 else {
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
807 box_name = fts_box_get_root(box, &ns);
8505
5cc1c4f3d38b fts-solr: In filter queries use "+" prefixes to require all the parameters to match.
Timo Sirainen <tss@iki.fi>
parents: 8500
diff changeset
808 str_printfa(str, "+%%2Buidv:%u+%%2Bbox:", status.uidvalidity);
8950
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
809 solr_quote_http(str, box_name);
ec1ae90af21a fts: Fixes to how virtual mailboxes are searched.
Timo Sirainen <tss@iki.fi>
parents: 8946
diff changeset
810 solr_add_ns_query_http(str, ctx->backend, ns);
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
811 }
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
812
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
813 array_clear(maybe_uids);
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
814 if (!virtual) {
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
815 return solr_connection_select(solr_conn, str_c(str), NULL, NULL,
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
816 definite_uids, scores);
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
817 } else {
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
818 memset(&uid_map_ctx, 0, sizeof(uid_map_ctx));
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
819 uid_map_ctx.backend = ctx->backend;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
820 uid_map_ctx.box = box;
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
821 uid_map_ctx.vname = t_str_new(256);
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
822 return solr_connection_select(solr_conn, str_c(str),
8499
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
823 solr_virtual_uid_map,
252b29ac5f43 fts: Added support for handling multiple namespaces.
Timo Sirainen <tss@iki.fi>
parents: 8495
diff changeset
824 &uid_map_ctx,
8484
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
825 definite_uids, scores);
db78eab2ba5d Initial support for getting full text search working nicely with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8109
diff changeset
826 }
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
827 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
828
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
829 struct fts_backend fts_backend_solr = {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
830 MEMBER(name) "solr",
9292
96b764649cbb fts: Fixed crashing with Squat.
Timo Sirainen <tss@iki.fi>
parents: 9141
diff changeset
831 MEMBER(flags) FTS_BACKEND_FLAG_VIRTUAL_LOOKUPS,
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
832
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
833 {
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
834 fts_backend_solr_init,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
835 fts_backend_solr_deinit,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
836 fts_backend_solr_get_last_uid,
8495
ffb37c392166 Finished making full text search indexes work fast with virtual mailboxes.
Timo Sirainen <tss@iki.fi>
parents: 8493
diff changeset
837 fts_backend_solr_get_all_last_uids,
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
838 fts_backend_solr_build_init,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
839 fts_backend_solr_build_more,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
840 fts_backend_solr_build_deinit,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
841 fts_backend_solr_expunge,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
842 fts_backend_solr_expunge_finish,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
843 fts_backend_solr_lock,
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
844 fts_backend_solr_unlock,
8004
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
845 NULL,
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
846 NULL,
6418d4c0ff16 fts: Added a new lookup API where the backend can look up all the fields
Timo Sirainen <tss@iki.fi>
parents: 8002
diff changeset
847 fts_backend_solr_lookup
7990
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
848 }
662172573fe1 Initial code to support Apache Solr (Lucene indexing server).
Timo Sirainen <tss@iki.fi>
parents:
diff changeset
849 };