|
0
|
1 test:
|
|
|
2 - make sure mmap()s work properly with NFS
|
|
|
3 - make sure first_hole_records is updated properly
|
|
|
4 - make sure locking is done properly when opening/switching modifylog
|
|
|
5 - make sure index->header->flags are updated at correct times
|
|
|
6 - make sure SELECT rebuilds index properly when next_uid is near 32bit value
|
|
|
7 - make sure io_buffer_send() handles huge inputs properly
|
|
|
8 - make sure rfc822_parse_date() works properly
|
|
|
9 - make sure imap_match functions work properly
|
|
|
10 - try imap_message_send() / io_buffer_send_file()
|
|
|
11 - make sure connection limits work
|
|
|
12
|
|
|
13 index:
|
|
|
14 - optimization:
|
|
|
15 - optimize so that when all mail is deleted, the indexes will be
|
|
|
16 truncated
|
|
|
17 - could hash function be better..? like uid*uid? what about changing
|
|
|
18 probe strategy from linear to something else?
|
|
|
19 - support shrinking hash file when it becomes 99% empty or so
|
|
|
20 - save part sizes + positions so MessagePart could be regenerated from
|
|
|
21 index..? probably only needed if BODY/BODYSTRUCTURE is saved, as it's
|
|
|
22 only useful with FETCH BODY[mime.sections] and they wouldn't be known
|
|
|
23 without bodystructure..
|
|
|
24 - index->lookup_uid_range(): first_uid could quite often be either the
|
|
|
25 first UID or some UIDs below the first. optimize these by remembering
|
|
|
26 the first UID in index.
|
|
|
27 - mbox:
|
|
|
28 - BUG: adding new mail after indexes are created doesn't work
|
|
|
29 - save MD5 sums for messages?
|
|
|
30 - update Status and X-Status headers when flags are changed
|
|
|
31 - last \n shouldn't be sent for messages. also remember to fix
|
|
|
32 the From-checks to check for [\r]\nFrom instead then..
|
|
|
33 - EXPUNGE doesn't delete the mail from the mbox file
|
|
|
34 - fsck should probably (or optionally?) really scan the message body
|
|
|
35 for "\nFrom " text instead of just jumping over the message body.
|
|
|
36 Quite useless actually, but this would make it fully reliable with md5
|
|
|
37 anyway..
|
|
|
38 - there's some race condition issues when opening mailboxes..
|
|
|
39 - when opening index files, check the flags and do what's needed. fsck and
|
|
|
40 rebuild is supported currently. compression and hash rebuilding is still
|
|
|
41 needed. and the cache_fields .. not sure when that'd be done, preferably
|
|
|
42 in the separate compress-process..
|
|
|
43 - set_lock() is ugly and horrible and should really be done something.
|
|
|
44 does the syncing really need to be there? maybe put it into separate
|
|
|
45 function which can be called after set_lock() by functions which actually
|
|
|
46 care about the sync state (fetch, search, store, etc).
|
|
|
47 - read-only support so we could use an index where we don't have
|
|
|
48 write-access? we should use MAP_PRIVATE everywhere with these boxes to
|
|
|
49 make sure we don't get exploited .. and anyway recheck everything to make
|
|
|
50 sure there wouldn't be a way to exploit them.
|
|
|
51 - if index was just rebuilt, modify log complains about indexid mismatch
|
|
|
52 at first open
|
|
|
53 - does append work?
|
|
|
54
|
|
|
55 lib-storage:
|
|
|
56 - support multiple mailbox formats and locations for one user. that would
|
|
|
57 require support for multiple MailStorages, and since we're chroot()ed,
|
|
|
58 usually the only way to communicate with others would be to create
|
|
|
59 RemoteMailStorage which would use TCP/UNIX sockets to connect to another
|
|
|
60 imap session.
|
|
|
61 - DELETE/RENAME: when someone else had the mailbox open, we should
|
|
|
62 disconnect it (when stat() fails with ENOENT while syncing)
|
|
|
63 - optimize SEARCH [UN]SEEN, [UN]DELETED and [UN]RECENT. They're able to
|
|
|
64 skip lots of messages based on the index header data.
|
|
|
65 - use a trie index for fast text searching, like cyrus squat?
|
|
|
66 - hardlink-COPY doesn't copy flags
|
|
|
67 - maildir: atomic COPY could be done by setting a "temporary" flag into the
|
|
|
68 file's name. once copying is done, set an ignore-temporary field into
|
|
|
69 index's header. at next sync the temporary flag will be removed.
|
|
|
70 - mbox: internal_date isn't saved
|
|
|
71 - select "" shouldn't work.
|
|
|
72
|
|
|
73 general:
|
|
|
74 - capabilities:
|
|
|
75 - acl (rfc2086)
|
|
|
76 - quota (rfc2087)
|
|
|
77 - namespace (rfc2342), id (rfc2971), mailbox-referrals (rfc2193),
|
|
|
78 literal+ (rfc2088), idle (rfc2177), uidplus (rfc2359)
|
|
|
79 - drafts: listext, children, unselect, multiappend, annotatemore
|
|
|
80 - sort, thread: are these really useful for clients? do any actually
|
|
|
81 use them? i'd think most clients want to know all the messages
|
|
|
82 anyway and can do the sorting/threading themselves.
|
|
|
83 - http://www.imc.org/ids.html
|
|
|
84 - check if t_push()/t_pop() should be added somewhere
|
|
|
85 - rfc-2231 continuation support
|
|
|
86 - "UID FETCH|SEARCH|STORE *" doesn't work if latest message was deleted.
|
|
|
87 should we bother to fix this? I doubt there's a client that would use this.
|
|
|
88 - RENAME INBOX isn't atomic with Maildir. And in general, RENAME can't
|
|
|
89 be moved to another storage. Maybe support doing also using COPY + delete
|
|
|
90 once COPY is atomic?
|
|
|
91
|
|
|
92 - go through .temp files and delete them
|
|
|
93 - grep for FIXME
|
|
|
94 - cache keeps the last message mmap()ed .. is there some case when it's not
|
|
|
95 a good idea? like the file changes in the background? cache should be
|
|
|
96 updated then. yes, especially with mbox support. the mmap should be
|
|
|
97 removed after unlocking. also, it shouldn't depend on mmap() anyway as
|
|
|
98 it's not possible to use it with eg. SQL storage.. except if we make
|
|
|
99 mmap()ing it optional, just give it some function which in some way
|
|
|
100 generates const char *msg + size_t.
|
|
|
101 - if auth process died and login couldn't immediately reconnect to it, it's
|
|
|
102 left until next user connects. however the connection needs to read the
|
|
|
103 init data before it can be used, so the user gets "NO Unknown
|
|
|
104 authentication method" error the first time
|
|
|
105 - ulimit / setrlimit() should be set somewhere
|
|
|
106 - create indexer binary
|
|
|
107 - SEARCH CHARSET support, iconv()?
|
|
|
108 - Fix the blocking SSL handshake
|
|
|
109 - SRP authentication support?
|
|
|
110 - Digest-MD5: support integrity protection, and maybe crypting. Do it
|
|
|
111 through imap-login like SSL is done?
|
|
|
112 - imap-auth should limit how fast authentication requests are allowed from
|
|
|
113 login processes. especially if there's one login/connection the speed
|
|
|
114 should be something like once/sec.
|
|
|
115 - support executing each login in it's own process, so if an exploit is ever
|
|
|
116 found from it, the attacker can't see other users' passwords
|
|
|
117 - the error messages given in command replies can sometimes be quite
|
|
|
118 specific, eg. rename(/full/path, /full/new/path) failed: xxx. These
|
|
|
119 probably shouldn't be shown to user, instead just print some "internal
|
|
|
120 error" with a timestamp and the real error would be written into syslog.
|
|
|
121 all errors from lib-index should be done this, and maybe some/all
|
|
|
122 lib-storage errors as well (there's separate error vs. critical)
|
|
|
123 - mmap()ing large messages isn't very good idea. we need to support doing
|
|
|
124 everything in pieces .. add mmap() support for iobuffers, and use them
|
|
|
125 everywhere?
|
|
|
126 - Make sure messages of size INT_MAX..UINT_MAX work correctly
|
|
|
127 - allocating readwrite pools now just uses system_pool .. so pool_unref()
|
|
|
128 can't free memory used by it .. what to do about it? at least count the
|
|
|
129 malloc/free calls and make complain if at the exit they don't match
|
|
|
130 - put IMAP_LOGFILE into config file. and the timestamp format.
|
|
|
131 - SIGHUPing master should reload the configuration
|
|
|
132 - Something's wrong with expunging mails from maildir ..
|
|
|
133
|
|
|
134 optional optimizations:
|
|
|
135 - provide some helper binary to save new mail into mailboxes with CR+LF
|
|
|
136 line breaks?
|
|
|
137 - disk I/O is the biggest problem, so split the mail into multiple computers
|
|
|
138 based on user and have a proxy in the front redirecting the connection.
|
|
|
139 cyrus had something like this except a lot more complicated - it tried
|
|
|
140 to fix the problem of having shared mailboxes. we have the same problem
|
|
|
141 with local shared mailboxes as we chroot(), so locally we could communicate
|
|
|
142 with UNIX sockets, remotely that could be done with TCP sockets.
|
