mairix.1

.TH MAIRIX 1 "January 2006"
.de Sx
.PP
.ne \\$1
.nf
.na
.RS 7
..
.de Ex
.RE 
.fi
.ad
.PP
..
.de Sy
.PP
.ne \\$1
.nf
.na
.RS 12
..
.de Ey
.RE 
.fi
.ad
.IP "" 7
..
.SH NAME
mairix \- index and search mail folders
.SH SYNOPSIS
.SS Indexing
.B mairix
[
.BR \-v | \-\-verbose
] [
.BR \-p | \-\-purge
] [
.BR \-f | \-\-rcfile
.I mairixrc
] [
.BR \-F | \-\-fast-index
]

.SS Searching
.B mairix
[
.BR \-v | \-\-verbose
] [
.BR \-f | \-\-rcfile
.I mairixrc
] [
.BR \-r | \-\-raw-output
] [
.BR \-x | \-\-excerpt-output
] [
.BR \-H | \-\-force-hardlinks
] [
.BR \-o | \-\-mfolder
.I mfolder
] [
.BR \-a | \-\-augment
] [
.BR \-t | \-\-threads
]
.I search-patterns

.SS Other
.B mairix
[
.BR \-h | \-\-help
]

.B mairix
[
.BR \-V | \-\-version
]

.B mairix
[
.BR \-d | \-\-dump
]

.SH DESCRIPTION
.I mairix
indexes and searches a collection of email messages.  The folders containing
the messages for indexing are defined in the configuration file.  The indexing
stage produces a database file.  The database file provides rapid access to
details of the indexed messages during searching operations.  A search normally
produces a folder (so-called
.BR mfolder )
containing the matched messages.  However, a raw mode
.RB ( \-r )
exists which just lists the matched messages instead.
.PP
It can operate with the following folder types
.IP *
maildir
.IP *
MH (compatible with the MH folder formats used by xmh, sylpheed, claws-mail, nnml (Gnus) and evolution)
.IP *
mbox (including mboxes that have been compressed with gzip or bzip2)
.PP
If maildir or MH source folders are used, and a search outputs its matches to
an mfolder in maildir or MH format, symbolic links are used to reference the
original messages inside the mfolder.  However, if mbox folders are involved,
copies of messages are made instead.

.SH OPTIONS

.B mairix
decides whether indexing or searching is required by looking for the presence of any
.I search-patterns
on the command line.

.SS Special modes
.TP
.B -h, --help
.br
Show usage summary and exit

.TP
.B -V, --version
Show program version and exit

.TP
.B -d
.br
Dump the database's contents in human-readable form to stdout.

.SS General options
.TP
.BI "-f " mairixrc
.br
.ns
.TP
.BI "--rcfile " mairixrc
.br
Specify an alternative configuration file to use.  The default configuration file is
.IR ~/.mairixrc .

.TP
.B -v, --verbose
.br
Make the output more verbose

.TP
.B -Q, --no-integrity-checks
.br
Normally
.I mairix
will do some internal integrity tests on the database.  The
.B -Q
option removes these checks, making
.I mairix
run faster, but it will be less likely to detect internal problems if any bugs creep in.

The
.I nochecks
directive in the rc file has the same effect.

.TP
.B \-\-unlock
.br
.I mairix
locks its database file during any indexing or searching operation to prevent
multiple indexing runs interfering with each other, or an indexing run
interfering with search runs.  The
.B --unlock
option removes the lockfile before doing the requested indexing or searching
operation.  This is a convenient way of cleaning up a stale lockfile if an
earlier run crashed for some reason or was aborted.

.SS Indexing options

.TP
.B -p, --purge
.br
Cause stale (dead) messages to be purged from the database during an indexing
run.  (Normally, stale messages are left in the database because of the
additional cost of compacting away the storage that they take up.)

.TP
.B -F, --fast-index
.br
When processing maildir and MH folders,
.I mairix
normally compares the mtime and size of each message against the values stored
in the database.  If they have changed, the message will be rescanned.  This
check requires each message file to be stat'ed.  For large numbers of messages
in these folder types, this can be a sizeable overhead.

This option tells
.I mairix
to assume that when a message currently on-disc has a name matching one already
in the database, it should assume the message is unchanged.

A later indexing run without using this option will fix up any rescans that
were missed due to its use.

.SS Search options
.TP
.B -a, --augment
.br
Append newly matches messages to the current mfolder instead of creating the
mfolder from scratch.

.TP
.B -t, --threads
.br
As well as returning the matched messages, also return every message in the
same thread as one of the real matches.

.TP
.B -r, --raw-output
.br
Instead of creating an mfolder containing the matched messages, just show their
paths on stdout.

.TP
.B -x, --excerpt-output
.br
Instead of creating an mfolder containing the matched messages, display an
excerpt from their headers on stdout.  The excerpt shows To, Cc, From, Subject
and Date.

.TP
.B -H, --force-hardlinks
.br
Instead of creating symbolic links, force the use of hardlinks.  This helps
mailers such as alpine to realize that there are new mails in the search
folder.

.TP
.BI "-o " mfolder
.br
.ns
.TP
.BI "--mfolder " mfolder
.br
Specify a temporary alternative path for the mfolder to use, overriding the
.I mfolder
directive in the rc file.

.B mairix
will refuse to output search results into any folder that appears to be amongst
those that are indexed.  This is to prevent accidental deletion of emails.

.SS Search patterns
.TP
.BI t: word
.br
Match
.I word
in the To: header.

.TP
.BI c: word
.br
Match
.I word
in the Cc: header.

.TP
.BI f: word
.br
Match
.I word
in the From: header.

.TP
.BI s: word
.br
Match
.I word
in the Subject: header.

.TP
.BI m: word
.br
Match
.I word
in the Message-ID: header.

.TP
.BI b: word
.br
Match
.I word
in the message body.

.B Message body
is taken to mean any body part of type text/plain or text/html.  For text/html,
text within meta tags is ignored.  In particular, the URLs inside <A
HREF="..."> tags are not currently indexed.  Non-text attachments are ignored.
If there's an attachment of type message/rfc822, this is parsed and the match
is performed on this sub-message too.  If a hit occurs, the enclosing message
is treated as having a hit.

.TP
.BI d: "[start-datespec]" - "[end-datespec]"
.br
Match messages with Date: headers lying in the specific range.

.TP
.BI z: "[low-size]" - "[high-size]"
.br
Match messages whose size lies in the specified range.  If the
.I low-size
argument is omitted it defaults to zero.  If the
.I high-size
argument is omitted it defaults to infinite size.

For example, to match messages between 10kilobytes and 20kilobytes in size, the
following search term can be used:
.Sy 1
mairix z:10k-20k
.Ey

The suffix 'k' on a number means multiply by 1024, and the suffix 'M' on a
number means multiply by 1024*1024.

.TP
.BI n: word
.br
Match
.I word
occurring as the name of an attachment in the message.  Since attachment names
are usually long, this option would usually be used in the substring form.  So
.Sy 1
mairix n:mairix=
.Ey

would match all messages which have attachments whose names contain the
substring
.IR mairix .

The attachment name is determined from the name=xxx or filename=xxx qualifiers
on the Content-Type: and Content-Disposition: headers respectively.

.TP
.BI F: flags
.br
Match messages with particular flag settings.  The available flags are 's'
meaning seen, 'r' meaning replied, and 'f' meaning flagged.  The flags are
case-insensitive.  A flag letter may be prefixed by a '-' to negate its sense.  Thus

.Sy 1
mairix F:-s d:1w-
.Ey

would match any unread message less than a week old, and

.Sy 1
mairix F:f-r d:-1m
.Ey

would match any flagged message older than a month which you haven't replied to yet.

Note that the flag characters and their meanings agree with those used as the
suffix letters on message filenames in maildir folders.

.SS Searching for a match amongst more than one part of a message
.PP
Multiple body parts may be grouped together, if a match in any of them is
sought.  Common examples follow.

.TP
.BI tc: word
.br
Match
.I word
in either the To: or Cc: headers (or both).

.TP
.BI bs: word
.br
Match
.I word
in either the Subject: header or the message body (or both).

.PP
The
.B a:
search pattern is an abbreviation for
.BR tcf: ;
i.e. match the word in the To:, Cc: or From: headers.  ("a" stands for
"address" in this case.)

.SS Match words
The
.I word
argument to the search strings can take various forms.

.TP
.I ~word
.br
Match messages
.B not
containing the word.

.TP
.I word1,word2
.br
This matches if both the words are matched in the specified message part.

.TP
.I word1/word2
.br
This matches if either of the words are matched in the specified message part.

.TP
.I substring=
.br
Match any word containing
.I substring
as a substring

.TP
.I substring=N
.br
Match any word containing
.IR substring ,
allowing up to
.I N
errors in the match.  For example, if
.I N
is 1, a single error is allowed, where an error can be
.IP *
a missing letter
.IP *
an extra letter
.IP *
a different letter.

.TP
.I ^substring=
.br
Match any word containing
.I substring
as a substring, with the requirement that
.I substring
occurs at the beginning of the matched word.

.SS Precedence matters

The binding order of the constructions is:

.IP "1." 
Individual command line arguments define separate conditions which are AND-ed
together

.IP "2."
Within a single argument, the letters before the colon define which message
parts the expression applies to.  If there is no colon, the expression applies
to all the headers listed earlier and the body.

.IP "3."
After the colon, commas delineate separate disjuncts, which are
OR-ed together.

.IP "4."
Each disjunct may contain separate conjuncts, which are separated
by plus signs.  These conditions are AND-ed together.

.IP "5."
Each conjunct may start with a tilde to negate it, and may be
followed by a slash to indicate a substring match, optionally
followed by an integer to define the maximum number of errors
allowed.

.SS Date specification
.PP
This section describes the syntax used for specifying dates when
searching using the `d:' option.

Dates are specified as a range.  The start and end of the range can both be
specified.  Alternatively, if the start is omitted, it is treated as being the
beginning of time.  If the end is omitted, it is treated as the current time.

There are 4 basic formats:
.TP
.BI d: start-end
.br
Specify both start and end explicitly
.TP
.BI d: start-
Specify start, end is the current time
.TP
.BI d: -end
Specify end, start is 'a long time ago' (i.e. early enough to include any
message).
.TP
.BI d: period
Specify start and end implicitly, as the start and end of the
period given.

.PP
The start and end can be specified either absolute or relative.  A relative
endpoint is given as a number followed by a single letter defining the scaling:

.TS
box tab(&);
lb | lb | lb | lb.
letter & short for & example & meaning
=
.T&
l | l | l | l.
d & days   & 3d & 3 days
w & weeks  & 2w & 2 weeks (14 days)
m & months & 5m & 5 months (150 days)
y & years  & 4y & 4 years (4*365 days)
.TE

.PP
Months are always treated as 30 days, and years as 365 days, for
this purpose.

Absolute times can be specified in many forms.  Some forms have different
meanings when they define a start date from that when they define an end date.
Where a single expression specifies both the start and end (i.e. where the
argument to d: doesn't contain a `-'), it will usually have different
interpretations in the two cases.

In the examples below, suppose the current date is Sunday May 18th,
2003 (when I started to write this material.)

.TS
box tab(&);
l | l | l | l.
Example & Start date & End date & Notes
=
d:20030301\-20030425 & March 1st, 2003 & 25th April, 2003
d:030301\-030425 & March 1st, 2003 & April 25th, 2003 & century assumed
d:mar1\-apr25    & March 1st, 2003 & April 25th, 2003
d:Mar1\-Apr25    & March 1st, 2003 & April 25th, 2003 & case insensitive
d:MAR1\-APR25    & March 1st, 2003 & April 25th, 2003 & case insensitive
d:1mar\-25apr    & March 1st, 2003 & April 25th, 2003 & date and month in either order
d:2002          & January 1st, 2002 & December 31st, 2002 & whole year
d:mar           & March 1st, 2003 & March 31st, 2003 & most recent March
d:oct           & October 1st, 2002 & October 31st, 2002 & most recent October
d:21oct\-mar     & October 21st, 2002 & March 31st, 2003 & start before end
d:21apr\-mar     & April 21st, 2002 & March 31st, 2003 & start before end
d:21apr\-        & April 21st, 2003 & May 18th, 2003 & end omitted
d:\-21apr        & January 1st, 1900 & April 21st, 2003 & start omitted
d:6w\-2w         & April 6th, 2003 & May 4th, 2003 & both dates relative
d:21apr\-1w      & April 21st, 2003 & May 11th, 2003 & one date relative
d:21apr\-2y      & April 21st, 2001 & May 11th, 2001 & start before end
d:99\-11         & January 1st, 1999 & May 11th, 2003 &T{
2 digits are a day of the month if possible, otherwise a year
T}
d:99oct\-1oct    & October 1st, 1999 & October 1st, 2002 &T{
end before now, single digit is a day of the month
T}
d:99oct\-01oct   & October 1st, 1999 & October 31st, 2001 &T{
2 digits starting with zero treated as a year
T}
d:oct99\-oct1    & October 1st, 1999 & October 1st, 2002 &T{
day and month in either order
T}
d:oct99\-oct01   & October 1st, 1999 & October 31st, 2001 &T{
year and month in either order
T}
.TE

.PP
The principles in the table work as follows.
.IP \(bu
When the expression defines a period of more than a day (i.e. if a month or
year is specified), the earliest day in the period is taken when the start date
is defined, and the last day in the period if the end of the range is being
defined.
.IP \(bu
The end date is always taken to be on or before the current date.
.IP \(bu
The start date is always taken to be on or before the end date.

.SH "SETTING UP THE MATCH FOLDER"

If the match folder does not exist when running in search mode, it is
automatically created.  For 'mformat=maildir' (the default), this
should be all you need to do.  If you use 'mformat=mh', you may have to
run some commands before your mailer will recognize the folder.  e.g.
for mutt, you could do
.Sx 2
mkdir -p /home/richard/Mail/mfolder
touch /home/richard/Mail/mfolder/.mh_sequences
.Ex
which seems to work.  Alternatively, within mutt, you could set MBOX_TYPE to
'mh' and save a message to '+mfolder' to have mutt set up the structure for you
in advance.

If you use Sylpheed, the best way seems to be to create the new folder from
within Sylpheed before letting mairix write into it.

.SH EXAMPLES
.PP
Suppose my email address is <richard@doesnt.exist>.

Either of the following will match all messages newer than 3 months from me
with the word 'chrony' in the subject line:
.Sx 2
mairix d:3m- f:richard+doesnt+exist s:chrony
mairix d:3m- f:richard@doesnt.exist s:chrony
.Ex
Suppose I don't mind a few spurious matches on the address, I want a wider date
range, and I suspect that some messages I replied to might have had the subject
keyword spelt wrongly (let's allow up to 2 errors):
.Sx 1
mairix d:6m- f:richard s:chrony=2
.Ex

.SH NOTES
.PP
.B mairix
works exclusively in terms of
.IR words .
The index that's built
in indexing mode contains a table of which words occur in which
messages.  Hence, the search capability is based on finding messages
that contain particular words.
.B mairix
defines a word as any string of alphanumeric characters + underscore.  Any
whitespace, punctuation, hyphens etc are treated as word boundaries.

.B mairix
has special handling for the To:, Cc: and From: headers.
Besides the normal word scan, these headers are scanned a second time,
where the characters '@', '-' and '.' are also treated as word
characters.  This allows most (if not all) email addresses to appear in
the database as single words.  So if you have a mail from
wibble@foobar.zzz, it will match on both these searches

.Sx 2
mairix f:foobar
mairix f:wibble@foobar.zzz
.Ex
It should be clear by now that the searching cannot be used to find messages
matching general regular expressions.  This has never been much of a
limitation.  Most searches are for particular keywords that were in the
messages, or details of the recipients, or the approximate date.

It's also worth pointing out that there is no 'locality' information
stored, so you can't search for messages that have one words 'close' to
some other word.  For every message and every word, there is a simple
yes/no condition stored - whether the message contains the word in a
particular header or in the body.  So far this has proved to be
adequate.
.B mairix
has a similar feel to using an Internet search engine.

.SH FILES
.I ~/.mairixrc

.SH AUTHOR
Copyright (C) 2002-2006 Richard P. Curnow <rc@rc0.org.uk>
.SH "SEE ALSO"
mairixrc(5)
.SH BUGS
.PP
We need a plugin scheme to allow more types of attachment to be scanned and indexed.