* Exiting Gnus:: Stop reading news and get some work done.
* Group Topics:: A folding group mode divided into topics.
* Non-ASCII Group Names:: Accessing groups of non-English names.
+* Searching:: Mail search engines.
* Misc Group Stuff:: Other stuff that you can to do.
@end menu
question to @code{gnus-newsgroup-variables}. @xref{Various Summary
Stuff}. So if you want to set @code{message-from-style} via the group
parameters, then you may need the following statement elsewhere in your
-@file{~/.gnus} file:
+@file{~/.gnus.el} file:
@lisp
(add-to-list 'gnus-newsgroup-variables 'message-from-style)
header will be displayed incorrectly in the article buffer.
+@node Searching
+@section Searching
+
+@menu
+* nnir:: Searching on IMAP, with swish, namazu, etc.
+* nnmairix:: Searching maildir, MH or mbox with Mairix.
+@end menu
+
+@cindex Searching
+
+FIXME: This node is a stub.
+
+FIXME: Add a brief overview of Gnus search capabilities. A brief
+comparison of nnir, nnmairix, contrib/gnus-namazu would be nice
+as well.
+
+FIXME: Explain difference to @ref{Searching for Articles}, add reference
+and back-reference.
+
+@node nnir
+@subsection nnir
+
+FIXME: As a first step, convert the commentary of @file{nnir} to texi.
+@cindex nnir
+
+@node nnmairix
+@subsection nnmairix
+
+@cindex mairix
+@cindex nnmairix
+This paragraph describes how to set up mairix and the back end
+@code{nnmairix} for indexing and searching your mail from within
+Gnus. Additionally, you can create permanent ``smart'' groups which are
+bound to mairix searches and are automatically updated.
+
+@menu
+* About mairix:: About the mairix mail search engine
+* nnmairix requirements:: What you will need for using nnmairix
+* What nnmairix does:: What does nnmairix actually do?
+* Setting up mairix:: Set up your mairix installation
+* Configuring nnmairix:: Set up the nnmairix back end
+* nnmairix keyboard shortcuts:: List of available keyboard shortcuts
+* Propagating marks:: How to propagate marks from nnmairix groups
+* nnmairix tips and tricks:: Some tips, tricks and examples
+* nnmairix caveats:: Some more stuff you might want to know
+@end menu
+
+@c FIXME: The markup in this section needs improvement. E.g. add
+@c @sample{...}, maybe remove some @strong{...}, convert ` - ' to `---`,
+@c ...
+
+@node About mairix
+@subsubsection About mairix
+
+Mairix is a tool for indexing and searching words in locally stored
+mail. It was written by Richard Curnow and is licensed under the
+GPL. Mairix comes with most popular GNU/Linux distributions, but it also
+runs under Windows (with cygwin), Mac OS X and Solaris. The homepage can
+be found at
+@uref{http://www.rpcurnow.force9.co.uk/mairix/index.html}
+
+Though mairix might not be as flexible as other search tools like
+swish++ or namazu, which you can use via the @code{nnir} back end, it
+has the prime advantage of being incredibly fast. On current systems, it
+can easily search through headers and message bodies of thousands and
+thousands of mails in well under a second. Building the database
+necessary for searching might take a minute or two, but only has to be
+done once fully. Afterwards, the updates are done incrementally and
+therefore are really fast, too. Additionally, mairix is very easy to set
+up.
+
+For maximum speed though, mairix should be used with mails stored in
+@code{Maildir} or @code{MH} format (this includes the @code{nnml} back
+end), although it also works with mbox. Mairix presents the search
+results by populating a @emph{virtual} maildir/MH folder with symlinks
+which point to the ``real'' message files (if mbox is used, copies are
+made). Since mairix already presents search results in such a virtual
+mail folder, it is very well suited for using it as an external program
+for creating @emph{smart} mail folders, which represent certain mail
+searches. This is similar to a Kiboze group (@pxref{Kibozed Groups}),
+but much faster.
+
+@node nnmairix requirements
+@subsubsection nnmairix requirements
+
+Mairix searches local mail - that means, mairix absolutely must have
+direct access to your mail folders. If your mail resides on another
+server (e.g. an @acronym{IMAP} server) and you happen to have shell
+access, @code{nnmairix} supports running mairix remotely, e.g. via ssh.
+
+Additionally, @code{nnmairix} only supports the following Gnus back
+ends: @code{nnml}, @code{nnmaildir}, and @code{nnimap}. You
+@strong{must} use one of these back ends for using
+@code{nnmairix}. Other back ends, like @code{nnmbox}, @code{nnfolder} or
+@code{nnmh}, won't work.
+
+If you absolutely must use mbox and still want to use @code{nnmairix},
+you can set up a local @acronym{IMAP} server, which you then access via
+@code{nnimap}. This is a rather massive setup for accessing some mbox
+files, so just change to MH or Maildir already...
+
+@node What nnmairix does
+@subsubsection What nnmairix does
+
+The back end @code{nnmairix} enables you to call mairix from within Gnus,
+either to query mairix with a search term or to update the
+database. While visiting a message in the summary buffer, you can use
+several pre-defined shortcuts for calling mairix, e.g. to quickly
+search for all mails from the sender of the current message or to
+display the whole thread associated with the message, even if the
+mails are in different folders.
+
+Additionally, you can create permanent @code{nnmairix} groups which are bound
+to certain mairix searches. This way, you can easily create a group
+containing mails from a certain sender, with a certain subject line or
+even for one specific thread based on the Message-ID. If you check for
+new mail in these folders (e.g. by pressing @kbd{g} or @kbd{M-g}), they
+automatically update themselves by calling mairix.
+
+You might ask why you need @code{nnmairix} at all, since mairix already
+creates the group, populates it with links to the mails so that you can
+then access it with Gnus, right? Well, this @emph{might} work, but often
+does not - at least not without problems. Most probably you will get
+strange article counts, and sometimes you might see mails which Gnus
+claims have already been canceled and are inaccessible. This is due to
+the fact that Gnus isn't really amused when things are happening behind
+its back. Another problem can be the mail back end itself, e.g. if you
+use mairix with an @acronym{IMAP} server (I had Dovecot complaining
+about corrupt index files when mairix changed the contents of the search
+group). Using @code{nnmairix} should circumvent these problems.
+
+@code{nnmairix} is not really a mail back end - it's actually more like a
+wrapper, sitting between a ``real'' mail back end where mairix stores the
+searches and the Gnus front end. You can choose between three different
+mail back ends for the mairix folders: @code{nnml}, @code{nnmaildir} or
+@code{nnimap}. @code{nnmairix} will call the mairix binary so that the
+search results are stored in folders named
+@code{zz_mairix-<NAME>-<NUMBER>} on this mail back end, but it will
+present these folders in the Gnus front end only with @code{<NAME>}. You
+can use an existing mail back end where you already store your mail, but
+if you're uncomfortable with @code{nnmairix} creating new mail groups
+alongside your other mail, you can also create e.g. a new
+@code{nnmaildir} server exclusively for mairix. However, a special case
+exists if you want to use mairix remotely on an IMAP server with
+@code{nnimap} - here the mairix folders and your other mail must be on
+the same @code{nnimap} back end.
+
+@node Setting up mairix
+@subsubsection Setting up mairix
+
+First: create a backup of your mail folders (@pxref{nnmairix caveats}).
+
+Setting up mairix is easy: simply create a @file{.mairixrc} file with
+(at least) the following entries:
+
+@example
+# Your Maildir/MH base folder
+base=~/Maildir
+@end example
+
+This is the base folder for your mails. All the following paths are
+relative to this base folder. If you want to use @code{nnmairix} with
+@code{nnimap}, this base path has to point to the mail path where the
+@acronym{IMAP} server stores the mail folders!
+
+@c FIXME: Add typical examples?
+@example
+maildir= ... your maildir folders which should be indexed ...
+mh= ... your nnml/mh folders which should be indexed ...
+mbox = ... your mbox files which should be indexed ...
+@end example
+
+Specify all your maildir/nnml folders and mbox files (relative to the
+base path!) you want to index with mairix. See the man-page for
+mairixrc for details.
+
+@example
+omit=zz_mairix-*
+@end example
+
+@vindex nnmairix-group-prefix
+This should make sure that you don't accidentally index the mairix
+search results. You can change the prefix of these folders with the
+variable @code{nnmairix-group-prefix}.
+
+@c FIXME: Add typical examples?
+@example
+mformat= ... 'maildir' or 'mh' ...
+database= ... location of database file ...
+@end example
+
+The @code{format} setting specifies the output format for the mairix
+search folder. Set this to @code{mh} if you want to access search results
+with @code{nnml}. Otherwise choose @code{maildir}.
+
+See the man pages for mairix and mairixrc for further options. Now
+simply call @code{mairix} to create the index for the first time.
+
+@node Configuring nnmairix
+@subsubsection Configuring nnmairix
+
+In group mode, type @kbd{G b c}
+(@code{nnmairix-create-server-and-default-group}). This will ask you for all
+necessary information and create a @code{nnmairix} server as a foreign
+server. You will have to specify the following:
+
+@itemize @bullet
+
+@item
+The @strong{name} of the @code{nnmairix} server - choose whatever you
+want.
+
+@item
+The @strong{mail back end} where mairix should stores its
+searches. Currently @code{nnmaildir}, @code{nnimap} and @code{nnml} are
+supported. As explained above, for locally stored mails, this can be an
+existing mail back end where you store your mails. However, you can also
+create e.g. a new @code{nnmaildir} server exclusively for
+@code{nnmairix} in your secondary select methods (@pxref{Finding the
+News}). If you want to use mairix remotely on an @acronym{IMAP} server,
+you have to choose the corresponding @code{nnimap} back end here.
+
+@item
+@vindex nnmairix-mairix-search-options
+The @strong{command} to call the mairix binary. This will usually just
+be @code{mairix}, but you can also choose something like @code{ssh
+SERVER mairix} if you want to call mairix remotely, e.g. on your
+@acronym{IMAP} server. If you want to add some default options to
+mairix, you could do this here, but better use the variable
+@code{nnmairix-mairix-search-options} instead.
+
+@item
+The name of the @strong{default search group}. This will be the group
+where all temporary mairix searches are stored, i.e. all searches which
+are not bound to permanent @code{nnmairix} groups. Choose whatever you
+like.
+
+@item
+If the mail back end is @code{nnimap} or @code{nnmaildir}, you will be
+asked if you work with @strong{Maildir++}, i.e. with hidden maildir
+folders (=beginning with a dot). For example, you have to answer
+@samp{yes} here if you work with the Dovecot @acronym{IMAP}
+server. Otherwise, you should answer @samp{no} here.
+
+@end itemize
+
+@node nnmairix keyboard shortcuts
+@subsubsection nnmairix keyboard shortcuts
+
+In group mode:
+
+@table @kbd
+
+@item G b c
+@kindex G b c (Group)
+@findex nnmairix-create-server-and-default-group
+Creates @code{nnmairix} server and default search group for this server
+(@code{nnmairix-create-server-and-default-group}). You should have done
+this by now (@pxref{Configuring nnmairix}).
+
+@item G b s
+@kindex G b s (Group)
+@findex nnmairix-search
+Prompts for query which is then sent to the mairix binary. Search
+results are put into the default search group which is automatically
+displayed (@code{nnmairix-search}).
+
+@item G b m
+@kindex G b m (Group)
+@findex nnmairix-widget-search
+Allows you to create a mairix search or a permanent group more
+comfortably using graphical widgets, similar to a customization
+group. Just try it to see how it works (@code{nnmairix-widget-search}).
+
+@item G b i
+@kindex G b i (Group)
+@findex nnmairix-search-interactive
+Another command for creating a mairix query more comfortably, but uses
+only the minibuffer (@code{nnmairix-search-interactive}).
+
+@item G b g
+@kindex G b g (Group)
+@findex nnmairix-create-search-group
+Creates a permanent group which is associated with a search query
+(@code{nnmairix-create-search-group}). The @code{nnmairix} back end
+automatically calls mairix when you update this group with @kbd{g} or
+@kbd{M-g}.
+
+@item G b q
+@kindex G b q (Group)
+@findex nnmairix-group-change-query-this-group
+Changes the search query for the @code{nnmairix} group under cursor
+(@code{nnmairix-group-change-query-this-group}).
+
+@item G b t
+@kindex G b t (Group)
+@findex nnmairix-group-toggle-threads-this-group
+Toggles the 'threads' parameter for the @code{nnmairix} group under cursor,
+i.e. if you want see the whole threads of the found messages
+(@code{nnmairix-group-toggle-threads-this-group}).
+
+@item G b u
+@kindex G b u (Group)
+@findex nnmairix-update-database
+@vindex nnmairix-mairix-update-options
+Calls mairix binary for updating the database
+(@code{nnmairix-update-database}). The default parameters are @code{-F}
+and @code{-Q} for making this as fast as possible (see variable
+@code{nnmairix-mairix-update-options} for defining these default
+options).
+
+@item G b r
+@kindex G b r (Group)
+@findex nnmairix-group-toggle-readmarks-this-group
+Keep articles in this @code{nnmairix} group always read or unread, or leave the
+marks unchanged (@code{nnmairix-group-toggle-readmarks-this-group}).
+
+@item G b d
+@kindex G b d (Group)
+@findex nnmairix-group-delete-recreate-this-group
+Recreate @code{nnmairix} group on the ``real'' mail back end
+(@code{nnmairix-group-delete-recreate-this-group}). You can do this if
+you always get wrong article counts with a @code{nnmairix} group.
+
+@item G b a
+@kindex G b a (Group)
+@findex nnmairix-group-toggle-allowfast-this-group
+Toggles the @code{allow-fast} parameters for group under cursor
+(@code{nnmairix-group-toggle-allowfast-this-group}). The default
+behavior of @code{nnmairix} is to do a mairix search every time you
+update or enter the group. With the @code{allow-fast} parameter set,
+mairix will only be called when you explicitly update the group, but not
+upon entering. This makes entering the group faster, but it may also
+lead to dangling symlinks if something changed between updating and
+entering the group which is not yet in the mairix database.
+
+@item G b p
+@kindex G b p (Group)
+@findex nnmairix-group-toggle-propmarks-this-group
+Toggle marks propagation for this group
+(@code{nnmairix-group-toggle-propmarks-this-group}). (@pxref{Propagating
+marks}).
+
+@item G b o
+@kindex G b o (Group)
+@findex nnmairix-propagate-marks
+Manually propagate marks (@code{nnmairix-propagate-marks}); needed only when
+@code{nnmairix-propagate-marks-upon-close} is set to @code{nil}.
+
+@end table
+
+In summary mode:
+
+@table @kbd
+
+@item $ m
+@kindex $ m (Summary)
+@findex nnmairix-widget-search-from-this-article
+Allows you to create a mairix query or group based on the current
+message using graphical widgets (same as @code{nnmairix-widget-search})
+(@code{nnmairix-widget-search-from-this-article}).
+
+@item $ g
+@kindex $ g (Summary)
+@findex nnmairix-create-search-group-from-message
+Interactively creates a new search group with query based on the current
+message, but uses the minibuffer instead of graphical widgets
+(@code{nnmairix-create-search-group-from-message}).
+
+@item $ t
+@kindex $ t (Summary)
+@findex nnmairix-search-thread-this-article
+Searches thread for the current article
+(@code{nnmairix-search-thread-this-article}). This is effectively a
+shortcut for calling @code{nnmairix-search} with @samp{m:msgid} of the
+current article and enabled threads.
+
+@item $ f
+@kindex $ f (Summary)
+@findex nnmairix-search-from-this-article
+Searches all messages from sender of the current article
+(@code{nnmairix-search-from-this-article}). This is a shortcut for
+calling @code{nnmairix-search} with @samp{f:From}.
+
+@item $ o
+@kindex $ o (Summary)
+@findex nnmairix-goto-original-article
+(Only in @code{nnmairix} groups!) Tries determine the group this article
+originally came from and displays the article in this group, so that
+e.g. replying to this article the correct posting styles/group
+parameters are applied (@code{nnmairix-goto-original-article}). This
+function will use the registry if available, but can also parse the
+article file path as a fallback method.
+
+@item $ u
+@kindex $ u (Summary)
+@findex nnmairix-remove-tick-mark-original-article
+Remove possibly existing tick mark from original article
+(@code{nnmairix-remove-tick-mark-original-article}). (@pxref{nnmairix
+tips and tricks}).
+
+@end table
+
+@node Propagating marks
+@subsubsection Propagating marks
+
+First of: you really need a patched mairix binary for using the marks
+propagation feature efficiently. Otherwise, you would have to update
+the mairix database all the time. You can get the patch at
+
+@uref{http://m61s02.vlinux.de/mairix-maildir-patch.tar}
+
+You need the mairix v0.21 source code for this patch; everything else
+is explained in the accompanied readme file. If you don't want to use
+marks propagation, you don't have to apply these patches, but they also
+fix some annoyances regarding changing maildir flags, so it might still
+be useful to you.
+
+With the patched mairix binary, you can use @code{nnmairix} as an
+alternative to mail splitting (@pxref{Fancy Mail Splitting}). For
+example, instead of splitting all mails from @samp{david@@foobar.com}
+into a group, you can simply create a search group with the query
+@samp{f:david@@foobar.com}. This is actually what ``smart folders'' are
+all about: simply put everything in one mail folder and dynamically
+create searches instead of splitting. This is more flexible, since you
+can dynamically change your folders any time you want to. This also
+implies that you will usually read your mails in the @code{nnmairix}
+groups instead of your ``real'' mail groups.
+
+There is one problem, though: say you got a new mail from
+@samp{david@@foobar.com} - it will now show up in two groups, the
+``real'' group (your INBOX, for example) and in the @code{nnmairix}
+search group (provided you have updated the mairix database). Now you
+enter the @code{nnmairix} group and read the mail. The mail will be
+marked as read, but only in the @code{nnmairix} group - in the ``real''
+mail group it will be still shown as unread.
+
+You could now catch up the mail group (@pxref{Group Data}), but this is
+tedious and error prone, since you may overlook mails you don't have
+created @code{nnmairix} groups for. Of course, you could first use
+@code{nnmairix-goto-original-article} (@pxref{nnmairix keyboard
+shortcuts}) and then read the mail in the original group, but that's
+even more cumbersome.
+
+Clearly, the easiest way would be if marks could somehow be
+automatically set for the original article. This is exactly what
+@emph{marks propagation} is about.
+
+Marks propagation is deactivated by default. You can activate it for a
+certain @code{nnmairix} group with
+@code{nnmairix-group-toggle-propmarks-this-group} (bound to @kbd{G b
+p}). This function will warn you if you try to use it with your default
+search group; the reason is that the default search group is used for
+temporary searches, and it's easy to accidentally propagate marks from
+this group. However, you can ignore this warning if you really want to.
+
+With marks propagation enabled, all the marks you set in a @code{nnmairix}
+group should now be propagated to the original article. For example,
+you can now tick an article (by default with @kbd{!}) and this mark should
+magically be set for the original article, too.
+
+A few more remarks which you may or may not want to know:
+
+@vindex nnmairix-propagate-marks-upon-close
+Marks will not be set immediately, but only upon closing a group. This
+not only makes marks propagation faster, it also avoids problems with
+dangling symlinks when dealing with maildir files (since changing flags
+will change the file name). You can also control when to propagate marks
+via @code{nnmairix-propagate-marks-upon-close} (see the doc-string for
+details).
+
+Obviously, @code{nnmairix} will have to look up the original group for every
+article you want to set marks for. If available, @code{nnmairix} will first use
+the registry for determining the original group. The registry is very
+fast, hence you should really, really enable the registry when using
+marks propagation. If you don't have to worry about RAM and disc space,
+set @code{gnus-registry-max-entries} to a large enough value; to be on
+the safe side, choose roughly the amount of mails you index with mairix.
+
+@vindex nnmairix-only-use-registry
+If you don't want to use the registry or the registry hasn't seen the
+original article yet, @code{nnmairix} will use an additional mairix search
+for determining the file path of the article. This, of course, is way
+slower than the registry - if you set hundreds or even thousands of
+marks this way, it might take some time. You can avoid this situation
+by setting @code{nnmairix-only-use-registry} to t.
+
+Maybe you also want to propagate marks the other way round, i.e. if you
+tick an article in a "real" mail group, you'd like to have the same
+article in a @code{nnmairix} group ticked, too. For several good
+reasons, this can only be done efficiently if you use maildir. To
+immediately contradict myself, let me mention that it WON'T work with
+@code{nnmaildir}, since @code{nnmaildir} stores the marks externally and
+not in the file name. Therefore, propagating marks to @code{nnmairix}
+groups will usually only work if you use an IMAP server which uses
+maildir as its file format.
+
+@vindex nnmairix-propagate-marks-to-nnmairix-groups
+If you work with this setup, just set
+@code{nnmairix-propagate-marks-to-nnmairix-groups} to @code{t} and see what
+happens. If you don't like what you see, just set it to @code{nil} again. One
+problem might be that you get a wrong number of unread articles; this
+usually happens when you delete or expire articles in the original
+groups. When this happens, you can recreate the @code{nnmairix} group on the
+back end using @kbd{G b d}.
+
+@node nnmairix tips and tricks
+@subsubsection nnmairix tips and tricks
+
+@itemize
+@item
+Checking Mail
+
+@findex nnmairix-update-groups
+I put all my important mail groups at group level 1. The mairix groups
+have group level 5, so they do not get checked at start up (@pxref{Group
+Levels}).
+
+I use the following to check for mails:
+
+@lisp
+(defun my-check-mail-mairix-update (level)
+ (interactive "P")
+ ;; if no prefix given, set level=1
+ (gnus-group-get-new-news (or level 1))
+ (nnmairix-update-groups "mairixsearch" t t)
+ (gnus-group-list-groups))
+
+(define-key gnus-group-mode-map "g" 'my-check-mail-mairix-update)
+@end lisp
+
+Instead of @samp{"mairixsearch"} use the name of your @code{nnmairix}
+server. See the doc string for @code{nnmairix-update-groups} for
+details.
+
+@item
+Example: search group for ticked articles
+
+For example, you can create a group for all ticked articles, where the
+articles always stay unread:
+
+Hit @kbd{G b g}, enter group name (e.g. @samp{important}), use
+@samp{F:f} as query and do not include threads.
+
+Now activate marks propagation for this group by using @kbd{G b p}. Then
+activate the always-unread feature by using @kbd{G b r} twice.
+
+So far so good - but how do you remove the tick marks in the @code{nnmairix}
+group? There are two options: You may simply use
+@code{nnmairix-remove-tick-mark-original-article} (bound to @kbd{$ u}) to remove
+tick marks from the original article. The other possibility is to set
+@code{nnmairix-propagate-marks-to-nnmairix-groups} to @code{t}, but see the above
+comments about this option. If it works for you, the tick marks should
+also exist in the @code{nnmairix} group and you can remove them as usual,
+e.g. by marking an article as read.
+
+When you have removed a tick mark from the original article, this
+article should vanish from the @code{nnmairix} group after you have updated the
+mairix database and updated the group. Fortunately, there is a function
+for doing exactly that: @code{nnmairix-update-groups}. See the previous code
+snippet and the doc string for details.
+
+@item
+Dealing with auto-subscription of mail groups
+
+As described before, all @code{nnmairix} groups are in fact stored on
+the mail back end in the form @samp{zz_mairix-<NAME>-<NUMBER>}. You can
+see them when you enter the back end server in the server buffer. You
+should not subscribe these groups! Unfortunately, these groups will
+usually get @strong{auto-subscribed} when you use @code{nnmaildir} or
+@code{nnml}, i.e. you will suddenly see groups of the form
+@samp{zz_mairix*} pop up in your group buffer. If this happens to you,
+simply kill these groups with C-k. For avoiding this, turn off
+auto-subscription completely by setting the variable
+@code{gnus-auto-subscribed-groups} to @code{nil} (@pxref{Filtering New
+Groups}), or if you like to keep this feature use the following kludge
+for turning it off for all groups beginning with @samp{zz_}:
+
+@lisp
+(setq gnus-auto-subscribed-groups
+ "^\\(nnml\\|nnfolder\\|nnmbox\\|nnmh\\|nnbabyl\\|nnmaildir\\).*:\\([^z]\\|z$\\|\\z[^z]\\|zz$\\|zz[^_]\\|zz_$\\).*")
+@end lisp
+
+@end itemize
+
+@node nnmairix caveats
+@subsubsection nnmairix caveats
+
+@itemize
+@item
+If you use the Gnus agent (@pxref{Gnus Unplugged}): don't agentize
+@code{nnmairix} groups (though I have no idea what happens if you do).
+
+@item
+If you use the Gnus registry: don't use the registry with
+@code{nnmairix} groups (put them in
+@code{gnus-registry-unfollowed-groups}). Be @strong{extra careful} if
+you use @code{gnus-registry-split-fancy-with-parent} - mails which are
+split into @code{nnmairix} groups are usually gone for good as soon as
+you check the group for new mail (yes, it has happened to me...).
+
+@item
+Therefore: @strong{Never ever} put ``real'' mails into @code{nnmairix}
+groups (you shouldn't be able to, anyway).
+
+@item
+mairix does only support us-ascii characters.
+
+@item
+@code{nnmairix} uses a rather brute force method to force Gnus to
+completely reread the group on the mail back end after mairix was
+called - it simply deletes and re-creates the group on the mail
+back end. So far, this has worked for me without any problems, and I
+don't see how @code{nnmairix} could delete other mail groups than its
+own, but anyway: you really should have a backup of your mail
+folders.
+
+@item
+All necessary information is stored in the group parameters
+(@pxref{Group Parameters}). This has the advantage that no active file
+is needed, but also implies that when you kill a @code{nnmairix} group,
+it is gone for good.
+
+@item
+@findex nnmairix-purge-old-groups
+If you create and kill a lot of @code{nnmairix} groups, the
+``zz_mairix-*'' groups will accumulate on the mail back end server. To
+delete old groups which are no longer needed, call
+@code{nnmairix-purge-old-groups}. Note that this assumes that you don't
+save any ``real'' mail in folders of the form
+@code{zz_mairix-<NAME>-<NUMBER>}. You can change the prefix of
+@code{nnmairix} groups by changing the variable
+@code{nnmairix-group-prefix}.
+
+@item
+The following only applies if you @strong{don't} use the mentioned patch
+for mairix (@pxref{Propagating marks}):
+
+A problem can occur when using @code{nnmairix} with maildir folders and
+comes with the fact that maildir stores mail flags like @samp{Seen} or
+@samp{Replied} by appending chars @samp{S} and @samp{R} to the message
+file name, respectively. This implies that currently you would have to
+update the mairix database not only when new mail arrives, but also when
+mail flags are changing. The same applies to new mails which are indexed
+while they are still in the @samp{new} folder but then get moved to
+@samp{cur} when Gnus has seen the mail. If you don't update the database
+after this has happened, a mairix query can lead to symlinks pointing to
+non-existing files. In Gnus, these messages will usually appear with
+``(none)'' entries in the header and can't be accessed. If this happens
+to you, using @kbd{G b u} and updating the group will usually fix this.
+
+@end itemize
+
@node Misc Group Stuff
@section Misc Group Stuff
--- /dev/null
+;;; nnir.el --- search mail with various search engines -*- coding: iso-8859-1 -*-
+
+;; Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004,
+;; 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
+
+;; Author: Kai Großjohann <grossjohann@ls6.cs.uni-dortmund.de>
+;; Swish-e and Swish++ backends by:
+;; Christoph Conrad <christoph.conrad@gmx.de>.
+;; IMAP backend by: Simon Josefsson <jas@pdc.kth.se>.
+;; IMAP search by: Torsten Hilbrich <torsten.hilbrich <at> gmx.net>
+;; IMAP search improved by Daniel Pittman <daniel@rimspace.net>.
+;; nnmaildir support for Swish++ and Namazu backends by:
+;; Justus Piater <Justus <at> Piater.name>
+
+;; TODO: Documentation in the Gnus manual
+
+;; From: Reiner Steib
+;; Subject: Re: Including nnir.el
+;; Newsgroups: gmane.emacs.gnus.general
+;; Message-ID: <v9d5dnp6aq.fsf@marauder.physik.uni-ulm.de>
+;; Date: 2006-06-05 22:49:01 GMT
+;;
+;; On Sun, Jun 04 2006, Sascha Wilde wrote:
+;;
+;; > The one thing most hackers like to forget: Documentation. By now the
+;; > documentation is only in the comments at the head of the source, I
+;; > would use it as basis to cook up some minimal texinfo docs.
+;; >
+;; > Where in the existing gnus manual would this fit best?
+
+;; Maybe (info "(gnus)Combined Groups") for a general description.
+;; `gnus-group-make-nnir-group' might be described in (info
+;; "(gnus)Foreign Groups") as well.
+
+;; Keywords: news mail searching ir
+
+;; This file is part of GNU Emacs.
+
+;; This is free software; you can redistribute it and/or modify
+;; it under the terms of the GNU General Public License as published by
+;; the Free Software Foundation; either version 3, or (at your option)
+;; any later version.
+
+;; GNU Emacs is distributed in the hope that it will be useful,
+;; but WITHOUT ANY WARRANTY; without even the implied warranty of
+;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+;; GNU General Public License for more details.
+
+;; You should have received a copy of the GNU General Public License
+;; along with GNU Emacs; see the file COPYING. If not, write to the
+;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+;; Boston, MA 02110-1301, USA.
+
+;;; Commentary:
+
+;; The most recent version of this can always be fetched from the Gnus
+;; CVS repository. See http://www.gnus.org/ for more information.
+
+;; This code is still in the development stage but I'd like other
+;; people to have a look at it. Please do not hesitate to contact me
+;; with your ideas.
+
+;; What does it do? Well, it allows you to index your mail using some
+;; search engine (freeWAIS-sf, swish-e and others -- see later),
+;; then type `G G' in the Group buffer and issue a query to the search
+;; engine. You will then get a buffer which shows all articles
+;; matching the query, sorted by Retrieval Status Value (score).
+
+;; When looking at the retrieval result (in the Summary buffer) you
+;; can type `G T' (aka M-x gnus-summary-nnir-goto-thread RET) on an
+;; article. You will be teleported into the group this article came
+;; from, showing the thread this article is part of. (See below for
+;; restrictions.)
+
+;; The Lisp installation is simple: just put this file on your
+;; load-path, byte-compile it, and load it from ~/.gnus or something.
+;; This will install a new command `G G' in your Group buffer for
+;; searching your mail. Note that you also need to configure a number
+;; of variables, as described below.
+
+;; Restrictions:
+;;
+;; * If you don't use HyREX as your search engine, this expects that
+;; you use nnml or another one-file-per-message backend, because the
+;; others doesn't support nnfolder.
+;; * It can only search the mail backend's which are supported by one
+;; search engine, because of different query languages.
+;; * There are restrictions to the Wais setup.
+;; * There are restrictions to the imap setup.
+;; * gnus-summary-nnir-goto-thread: Fetches whole group first, before
+;; limiting to the right articles. This is much too slow, of
+;; course. May issue a query for number of articles to fetch; you
+;; must accept the default of all articles at this point or things
+;; may break.
+
+;; The Lisp setup involves setting a few variables and setting up the
+;; search engine. You can define the variables in the server definition
+;; like this :
+;; (setq gnus-secondary-select-methods '(
+;; (nnimap "" (nnimap-address "localhost")
+;; (nnir-search-engine hyrex)
+;; (nnir-hyrex-additional-switches ("-d" "ddl-nnimap.xml"))
+;; )))
+;; Or you can define the global ones. The variables set in the mailer-
+;; definition will be used first.
+;; The variable to set is `nnir-search-engine'. Choose one of the engines
+;; listed in `nnir-engines'. (Actually `nnir-engines' is an alist,
+;; type `C-h v nnir-engines RET' for more information; this includes
+;; examples for setting `nnir-search-engine', too.)
+;;
+;; The variable nnir-mail-backend isn't used anymore.
+;;
+
+;; You must also set up a search engine. I'll tell you about the two
+;; search engines currently supported:
+
+;; 1. freeWAIS-sf
+;;
+;; As always with freeWAIS-sf, you need a so-called `format file'. I
+;; use the following file:
+;;
+;; ,-----
+;; | # Kai's format file for freeWAIS-sf for indexing mails.
+;; | # Each mail is in a file, much like the MH format.
+;; |
+;; | # Document separator should never match -- each file is a document.
+;; | record-sep: /^@this regex should never match@$/
+;; |
+;; | # Searchable fields specification.
+;; |
+;; | region: /^[sS]ubject:/ /^[sS]ubject: */
+;; | subject "Subject header" stemming TEXT BOTH
+;; | end: /^[^ \t]/
+;; |
+;; | region: /^([tT][oO]|[cC][cC]):/ /^([tT][oO]|[cC][cC]): */
+;; | to "To and Cc headers" SOUNDEX BOTH
+;; | end: /^[^ \t]/
+;; |
+;; | region: /^[fF][rR][oO][mM]:/ /^[fF][rR][oO][mM]: */
+;; | from "From header" SOUNDEX BOTH
+;; | end: /^[^ \t]/
+;; |
+;; | region: /^$/
+;; | stemming TEXT GLOBAL
+;; | end: /^@this regex should never match@$/
+;; `-----
+;;
+;; 1998-07-22: waisindex would dump core on me for large articles with
+;; the above settings. I used /^$/ as the end regex for the global
+;; field. That seemed to work okay.
+
+;; There is a Perl module called `WAIS.pm' which is available from
+;; CPAN as well as ls6-ftp.cs.uni-dortmund.de:/pub/wais/Perl. This
+;; module comes with a nifty tool called `makedb', which I use for
+;; indexing. Here's my `makedb.conf':
+;;
+;; ,-----
+;; | # Config file for makedb
+;; |
+;; | # Global options
+;; | waisindex = /usr/local/bin/waisindex
+;; | wais_opt = -stem -t fields
+;; | # `-stem' option necessary when `stemming' is specified for the
+;; | # global field in the *.fmt file
+;; |
+;; | # Own variables
+;; | homedir = /home/kai
+;; |
+;; | # The mail database.
+;; | database = mail
+;; | files = `find $homedir/Mail -name \*[0-9] -print`
+;; | dbdir = $homedir/.wais
+;; | limit = 100
+;; `-----
+;;
+;; The Lisp setup involves the `nnir-wais-*' variables. The most
+;; difficult to understand variable is probably
+;; `nnir-wais-remove-prefix'. Here's what it does: the output of
+;; `waissearch' basically contains the file name and the (full)
+;; directory name. As Gnus works with group names rather than
+;; directory names, the directory name is transformed into a group
+;; name as follows: first, a prefix is removed from the (full)
+;; directory name, then all `/' are replaced with `.'. The variable
+;; `nnir-wais-remove-prefix' should contain a regex matching exactly
+;; this prefix. It defaults to `$HOME/Mail/' (note the trailing
+;; slash).
+
+;; 2. Namazu
+;;
+;; The Namazu backend requires you to have one directory containing all
+;; index files, this is controlled by the `nnir-namazu-index-directory'
+;; variable. To function the `nnir-namazu-remove-prefix' variable must
+;; also be correct, see the documentation for `nnir-wais-remove-prefix'
+;; above.
+;;
+;; It is particularly important not to pass any any switches to namazu
+;; that will change the output format. Good switches to use include
+;; `--sort', `--ascending', `--early' and `--late'. Refer to the Namazu
+;; documentation for further information on valid switches.
+;;
+;; To index my mail with the `mknmz' program I use the following
+;; configuration file:
+;;
+;; ,----
+;; | package conf; # Don't remove this line!
+;; |
+;; | # Paths which will not be indexed. Don't use `^' or `$' anchors.
+;; | $EXCLUDE_PATH = "spam|sent";
+;; |
+;; | # Header fields which should be searchable. case-insensitive
+;; | $REMAIN_HEADER = "from|date|message-id|subject";
+;; |
+;; | # Searchable fields. case-insensitive
+;; | $SEARCH_FIELD = "from|date|message-id|subject";
+;; |
+;; | # The max length of a word.
+;; | $WORD_LENG_MAX = 128;
+;; |
+;; | # The max length of a field.
+;; | $MAX_FIELD_LENGTH = 256;
+;; `----
+;;
+;; My mail is stored in the directories ~/Mail/mail/, ~/Mail/lists/ and
+;; ~/Mail/archive/, so to index them I go to the directory set in
+;; `nnir-namazu-index-directory' and issue the following command.
+;;
+;; mknmz --mailnews ~/Mail/archive/ ~/Mail/mail/ ~/Mail/lists/
+;;
+;; For maximum searching efficiency I have a cron job set to run this
+;; command every four hours.
+
+;; 3. HyREX
+;;
+;; The HyREX backend requires you to have one directory from where all
+;; your relative paths are to, if you use them. This directory must be
+;; set in the `nnir-hyrex-index-directory' variable, which defaults to
+;; your home directory. You must also pass the base, class and
+;; directory options or simply your dll to the `nnir-hyrex-programm' by
+;; setting the `nnir-hyrex-additional-switches' variable accordently.
+;; To function the `nnir-hyrex-remove-prefix' variable must also be
+;; correct, see the documentation for `nnir-wais-remove-prefix' above.
+
+;; 4. find-grep
+;;
+;; The find-grep engine simply runs find(1) to locate eligible
+;; articles and searches them with grep(1). This, of course, is much
+;; slower than using a proper search engine but OTOH doesn't require
+;; maintenance of an index and is still faster than using any built-in
+;; means for searching. The method specification of the server to
+;; search must include a directory for this engine to work (E.g.,
+;; `nnml-directory'). The tools must be POSIX compliant. GNU Find
+;; prior to version 4.2.12 (4.2.26 on Linux due to incorrect ARG_MAX
+;; handling) does not work.
+;; ,----
+;; | ;; find-grep configuration for searching the Gnus Cache
+;; |
+;; | (nnml "cache"
+;; | (nnml-get-new-mail nil)
+;; | (nnir-search-engine find-grep)
+;; | (nnml-directory "~/News/cache/")
+;; | (nnml-active-file "~/News/cache/active"))
+;; `----
+
+;; Developer information:
+
+;; I have tried to make the code expandable. Basically, it is divided
+;; into two layers. The upper layer is somewhat like the `nnvirtual'
+;; or `nnkiboze' backends: given a specification of what articles to
+;; show from another backend, it creates a group containing exactly
+;; those articles. The lower layer issues a query to a search engine
+;; and produces such a specification of what articles to show from the
+;; other backend.
+
+;; The interface between the two layers consists of the single
+;; function `nnir-run-query', which just selects the appropriate
+;; function for the search engine one is using. The input to
+;; `nnir-run-query' is a string, representing the query as input by
+;; the user. The output of `nnir-run-query' is supposed to be a
+;; vector, each element of which should in turn be a three-element
+;; vector. The first element should be full group name of the article,
+;; the second element should be the article number, and the third
+;; element should be the Retrieval Status Value (RSV) as returned from
+;; the search engine. An RSV is the score assigned to the document by
+;; the search engine. For Boolean search engines, the
+;; RSV is always 1000 (or 1 or 100, or whatever you like).
+
+;; The sorting order of the articles in the summary buffer created by
+;; nnir is based on the order of the articles in the above mentioned
+;; vector, so that's where you can do the sorting you'd like. Maybe
+;; it would be nice to have a way of displaying the search result
+;; sorted differently?
+
+;; So what do you need to do when you want to add another search
+;; engine? You write a function that executes the query. Temporary
+;; data from the search engine can be put in `nnir-tmp-buffer'. This
+;; function should return the list of articles as a vector, as
+;; described above. Then, you need to register this backend in
+;; `nnir-engines'. Then, users can choose the backend by setting
+;; `nnir-search-engine'.
+
+;; Todo, or future ideas:
+
+;; * It should be possible to restrict search to certain groups.
+;;
+;; * There is currently no error checking.
+;;
+;; * The summary buffer display is currently really ugly, with all the
+;; added information in the subjects. How could I make this
+;; prettier?
+;;
+;; * A function which can be called from an nnir summary buffer which
+;; teleports you into the group the current article came from and
+;; shows you the whole thread this article is part of.
+;; Implementation suggestions?
+;; (1998-07-24: There is now a preliminary implementation, but
+;; it is much too slow and quite fragile.)
+;;
+;; * Support other mail backends. In particular, probably quite a few
+;; people use nnfolder. How would one go about searching nnfolders
+;; and producing the right data needed? The group name and the RSV
+;; are simple, but what about the article number?
+;; - The article number is encoded in the `X-Gnus-Article-Number'
+;; header of each mail.
+;; - The HyREX engine supports nnfolder.
+;;
+;; * Support compressed mail files. Probably, just stripping off the
+;; `.gz' or `.Z' file name extension is sufficient.
+;;
+;; * At least for imap, the query is performed twice.
+;;
+
+;; Have you got other ideas?
+
+;;; Setup Code:
+
+(require 'nnoo)
+(require 'gnus-group)
+(require 'gnus-sum)
+(require 'message)
+(require 'gnus-util)
+(eval-and-compile
+ (require 'cl))
+
+(nnoo-declare nnir)
+(nnoo-define-basics nnir)
+
+(gnus-declare-backend "nnir" 'mail)
+
+(defvar nnir-imap-search-field "TEXT"
+ "The IMAP search item when doing an nnir search")
+
+(defvar nnir-imap-search-arguments
+ '(("Whole message" . "TEXT")
+ ("Subject" . "SUBJECT")
+ ("To" . "TO")
+ ("From" . "FROM")
+ (nil . "HEADER \"%s\""))
+ "Mapping from user readable strings to IMAP search items for use in nnir")
+
+(defvar nnir-imap-search-argument-history ()
+ "The history for querying search options in nnir")
+
+;;; Developer Extension Variable:
+
+(defvar nnir-engines
+ `((wais nnir-run-waissearch
+ ())
+ (imap nnir-run-imap
+ ((criteria
+ "Search in: " ; Prompt
+ ,nnir-imap-search-arguments ; alist for completing
+ nil ; no filtering
+ nil ; allow any user input
+ nil ; initial value
+ nnir-imap-search-argument-history ; the history to use
+ ,nnir-imap-search-field ; default
+ )))
+ (swish++ nnir-run-swish++
+ ((group . "Group spec: ")))
+ (swish-e nnir-run-swish-e
+ ((group . "Group spec: ")))
+ (namazu nnir-run-namazu
+ ())
+ (hyrex nnir-run-hyrex
+ ((group . "Group spec: ")))
+ (find-grep nnir-run-find-grep
+ ((grep-options . "Grep options: "))))
+ "Alist of supported search engines.
+Each element in the alist is a three-element list (ENGINE FUNCTION ARGS).
+ENGINE is a symbol designating the searching engine. FUNCTION is also
+a symbol, giving the function that does the search. The third element
+ARGS is a list of cons pairs (PARAM . PROMPT). When issuing a query,
+the FUNCTION will issue a query for each of the PARAMs, using PROMPT.
+
+The value of `nnir-search-engine' must be one of the ENGINE symbols.
+For example, use the following line for searching using freeWAIS-sf:
+ (setq nnir-search-engine 'wais)
+Use the following line if you read your mail via IMAP and your IMAP
+server supports searching:
+ (setq nnir-search-engine 'imap)
+Note that you have to set additional variables for most backends. For
+example, the `wais' backend needs the variables `nnir-wais-program',
+`nnir-wais-database' and `nnir-wais-remove-prefix'.
+
+Add an entry here when adding a new search engine.")
+
+;;; User Customizable Variables:
+
+(defgroup nnir nil
+ "Search nnmh and nnml groups in Gnus with swish-e, freeWAIS-sf, or EWS."
+ :group 'gnus)
+
+;; Mail backend.
+
+;; TODO:
+;; If `nil', use server parameters to find out which server to search. CCC
+;;
+(defcustom nnir-mail-backend '(nnml "")
+ "*Specifies which backend should be searched.
+More precisely, this is used to determine from which backend to fetch the
+messages found.
+
+This must be equal to an existing server, so maybe it is best to use
+something like the following:
+ (setq nnir-mail-backend (nth 0 gnus-secondary-select-methods))
+The above line works fine if the mail backend you want to search is
+the first element of gnus-secondary-select-methods (`nth' starts counting
+at zero)."
+ :type '(sexp)
+ :group 'nnir)
+
+;; Search engine to use.
+
+(defcustom nnir-search-engine 'wais
+ "*The search engine to use. Must be a symbol.
+See `nnir-engines' for a list of supported engines, and for example
+settings of `nnir-search-engine'."
+ :type '(sexp)
+ :group 'nnir)
+
+;; freeWAIS-sf.
+
+(defcustom nnir-wais-program "waissearch"
+ "*Name of waissearch executable."
+ :type '(string)
+ :group 'nnir)
+
+(defcustom nnir-wais-database (expand-file-name "~/.wais/mail")
+ "*Name of Wais database containing the mail.
+
+Note that this should be a file name without extension. For example,
+if you have a file /home/john/.wais/mail.fmt, use this:
+ (setq nnir-wais-database \"/home/john/.wais/mail\")
+The string given here is passed to `waissearch -d' as-is."
+ :type '(file)
+ :group 'nnir)
+
+(defcustom nnir-wais-remove-prefix (concat (getenv "HOME") "/Mail/")
+ "*The prefix to remove from each directory name returned by waissearch
+in order to get a group name (albeit with / instead of .). This is a
+regular expression.
+
+For example, suppose that Wais returns file names such as
+\"/home/john/Mail/mail/misc/42\". For this example, use the following
+setting: (setq nnir-wais-remove-prefix \"/home/john/Mail/\")
+Note the trailing slash. Removing this prefix gives \"mail/misc/42\".
+`nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
+arrive at the correct group name, \"mail.misc\"."
+ :type '(regexp)
+ :group 'nnir)
+
+(defcustom nnir-swish++-configuration-file
+ (expand-file-name "~/Mail/swish++.conf")
+ "*Configuration file for swish++."
+ :type '(file)
+ :group 'nnir)
+
+(defcustom nnir-swish++-program "search"
+ "*Name of swish++ search executable."
+ :type '(string)
+ :group 'nnir)
+
+(defcustom nnir-swish++-additional-switches '()
+ "*A list of strings, to be given as additional arguments to swish++.
+
+Note that this should be a list. Ie, do NOT use the following:
+ (setq nnir-swish++-additional-switches \"-i -w\") ; wrong
+Instead, use this:
+ (setq nnir-swish++-additional-switches '(\"-i\" \"-w\"))"
+ :type '(repeat (string))
+ :group 'nnir)
+
+(defcustom nnir-swish++-remove-prefix (concat (getenv "HOME") "/Mail/")
+ "*The prefix to remove from each file name returned by swish++
+in order to get a group name (albeit with / instead of .). This is a
+regular expression.
+
+This variable is very similar to `nnir-wais-remove-prefix', except
+that it is for swish++, not Wais."
+ :type '(regexp)
+ :group 'nnir)
+
+;; Swish-E.
+;; URL: http://sunsite.berkeley.edu/SWISH-E/
+;; New version: http://www.boe.es/swish-e
+;; Variables `nnir-swish-e-index-file', `nnir-swish-e-program' and
+;; `nnir-swish-e-additional-switches'
+
+(make-obsolete-variable 'nnir-swish-e-index-file
+ 'nnir-swish-e-index-files)
+(defcustom nnir-swish-e-index-file
+ (expand-file-name "~/Mail/index.swish-e")
+ "*Index file for swish-e.
+This could be a server parameter.
+It is never consulted once `nnir-swish-e-index-files', which should be
+used instead, has been customized."
+ :type '(file)
+ :group 'nnir)
+
+(defcustom nnir-swish-e-index-files
+ (list nnir-swish-e-index-file)
+ "*List of index files for swish-e.
+This could be a server parameter."
+ :type '(repeat (file))
+ :group 'nnir)
+
+(defcustom nnir-swish-e-program "swish-e"
+ "*Name of swish-e search executable.
+This cannot be a server parameter."
+ :type '(string)
+ :group 'nnir)
+
+(defcustom nnir-swish-e-additional-switches '()
+ "*A list of strings, to be given as additional arguments to swish-e.
+
+Note that this should be a list. Ie, do NOT use the following:
+ (setq nnir-swish-e-additional-switches \"-i -w\") ; wrong
+Instead, use this:
+ (setq nnir-swish-e-additional-switches '(\"-i\" \"-w\"))
+
+This could be a server parameter."
+ :type '(repeat (string))
+ :group 'nnir)
+
+(defcustom nnir-swish-e-remove-prefix (concat (getenv "HOME") "/Mail/")
+ "*The prefix to remove from each file name returned by swish-e
+in order to get a group name (albeit with / instead of .). This is a
+regular expression.
+
+This variable is very similar to `nnir-wais-remove-prefix', except
+that it is for swish-e, not Wais.
+
+This could be a server parameter."
+ :type '(regexp)
+ :group 'nnir)
+
+;; HyREX engine, see <URL:http://ls6-www.cs.uni-dortmund.de/>
+
+(defcustom nnir-hyrex-program "nnir-search"
+ "*Name of the nnir-search executable."
+ :type '(string)
+ :group 'nnir)
+
+(defcustom nnir-hyrex-additional-switches '()
+ "*A list of strings, to be given as additional arguments for nnir-search.
+Note that this should be a list. Ie, do NOT use the following:
+ (setq nnir-hyrex-additional-switches \"-ddl ddl.xml -c nnir\") ; wrong !
+Instead, use this:
+ (setq nnir-hyrex-additional-switches '(\"-ddl\" \"ddl.xml\" \"-c\" \"nnir\"))"
+ :type '(repeat (string))
+ :group 'nnir)
+
+(defcustom nnir-hyrex-index-directory (getenv "HOME")
+ "*Index directory for HyREX."
+ :type '(directory)
+ :group 'nnir)
+
+(defcustom nnir-hyrex-remove-prefix (concat (getenv "HOME") "/Mail/")
+ "*The prefix to remove from each file name returned by HyREX
+in order to get a group name (albeit with / instead of .).
+
+For example, suppose that HyREX returns file names such as
+\"/home/john/Mail/mail/misc/42\". For this example, use the following
+setting: (setq nnir-hyrex-remove-prefix \"/home/john/Mail/\")
+Note the trailing slash. Removing this prefix gives \"mail/misc/42\".
+`nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
+arrive at the correct group name, \"mail.misc\"."
+ :type '(directory)
+ :group 'nnir)
+
+;; Namazu engine, see <URL:http://ww.namazu.org/>
+
+(defcustom nnir-namazu-program "namazu"
+ "*Name of Namazu search executable."
+ :type '(string)
+ :group 'nnir)
+
+(defcustom nnir-namazu-index-directory (expand-file-name "~/Mail/namazu/")
+ "*Index directory for Namazu."
+ :type '(directory)
+ :group 'nnir)
+
+(defcustom nnir-namazu-additional-switches '()
+ "*A list of strings, to be given as additional arguments to namazu.
+The switches `-q', `-a', and `-s' are always used, very few other switches
+make any sense in this context.
+
+Note that this should be a list. Ie, do NOT use the following:
+ (setq nnir-namazu-additional-switches \"-i -w\") ; wrong
+Instead, use this:
+ (setq nnir-namazu-additional-switches '(\"-i\" \"-w\"))"
+ :type '(repeat (string))
+ :group 'nnir)
+
+(defcustom nnir-namazu-remove-prefix (concat (getenv "HOME") "/Mail/")
+ "*The prefix to remove from each file name returned by Namazu
+in order to get a group name (albeit with / instead of .).
+
+This variable is very similar to `nnir-wais-remove-prefix', except
+that it is for Namazu, not Wais."
+ :type '(directory)
+ :group 'nnir)
+
+;;; Internal Variables:
+
+(defvar nnir-current-query nil
+ "Internal: stores current query (= group name).")
+
+(defvar nnir-current-server nil
+ "Internal: stores current server (does it ever change?).")
+
+(defvar nnir-current-group-marked nil
+ "Internal: stores current list of process-marked groups.")
+
+(defvar nnir-artlist nil
+ "Internal: stores search result.")
+
+(defvar nnir-tmp-buffer " *nnir*"
+ "Internal: temporary buffer.")
+
+;;; Code:
+
+;; Gnus glue.
+
+(defun gnus-group-make-nnir-group (extra-parms query)
+ "Create an nnir group. Asks for query."
+ (interactive "P\nsQuery: ")
+ (setq nnir-current-query nil
+ nnir-current-server nil
+ nnir-current-group-marked nil
+ nnir-artlist nil)
+ (let ((parms nil))
+ (if extra-parms
+ (setq parms (nnir-read-parms query))
+ (setq parms (list (cons 'query query))))
+ (add-to-list 'parms (cons 'unique-id (message-unique-id)) t)
+ (gnus-group-read-ephemeral-group
+ (concat "nnir:" (prin1-to-string parms)) '(nnir "") t
+ (cons (current-buffer)
+ gnus-current-window-configuration)
+ nil)))
+
+(defun nnir-group-mode-hook ()
+ (define-key gnus-group-mode-map (kbd "G G")
+ 'gnus-group-make-nnir-group))
+(add-hook 'gnus-group-mode-hook 'nnir-group-mode-hook)
+
+;; Why is this needed? Is this for compatibility with old/new gnusae? Using
+;; gnus-group-server instead works for me. -- Justus Piater
+(defmacro nnir-group-server (group)
+ "Return the server for a newsgroup GROUP.
+The returned format is as `gnus-server-to-method' needs it. See
+`gnus-group-real-prefix' and `gnus-group-real-name'."
+ `(let ((gname ,group))
+ (if (string-match "^\\([^:]+\\):" gname)
+ (progn
+ (setq gname (match-string 1 gname))
+ (if (string-match "^\\([^+]+\\)\\+\\(.+\\)$" gname)
+ (format "%s:%s" (match-string 1 gname) (match-string 2 gname))
+ (concat gname ":")))
+ (format "%s:%s" (car gnus-select-method) (cadr gnus-select-method)))))
+
+;; Summary mode commands.
+
+(defun gnus-summary-nnir-goto-thread ()
+ "Only applies to nnir groups. Go to group this article came from
+and show thread that contains this article."
+ (interactive)
+ (unless (eq 'nnir (car (gnus-find-method-for-group gnus-newsgroup-name)))
+ (error "Can't execute this command unless in nnir group."))
+ (let* ((cur (gnus-summary-article-number))
+ (group (nnir-artlist-artitem-group nnir-artlist cur))
+ (backend-number (nnir-artlist-artitem-number nnir-artlist cur))
+ server backend-group)
+ (setq server (nnir-group-server group))
+ (setq backend-group (gnus-group-real-name group))
+ (gnus-group-read-ephemeral-group
+ backend-group
+ (gnus-server-to-method server)
+ t ; activate
+ (cons (current-buffer)
+ 'summary) ; window config
+ nil
+ (list backend-number))
+ (gnus-summary-limit (list backend-number))
+ (gnus-summary-refer-thread)))
+
+(if (fboundp 'eval-after-load)
+ (eval-after-load "gnus-sum"
+ '(define-key gnus-summary-goto-map
+ "T" 'gnus-summary-nnir-goto-thread))
+ (add-hook 'gnus-summary-mode-hook
+ (function (lambda ()
+ (define-key gnus-summary-goto-map
+ "T" 'gnus-summary-nnir-goto-thread)))))
+
+
+
+;; Gnus backend interface functions.
+
+(deffoo nnir-open-server (server &optional definitions)
+ ;; Just set the server variables appropriately.
+ (nnoo-change-server 'nnir server definitions))
+
+(deffoo nnir-request-group (group &optional server fast)
+ "GROUP is the query string."
+ (nnir-possibly-change-server server)
+ ;; Check for cache and return that if appropriate.
+ (if (and (equal group nnir-current-query)
+ (equal gnus-group-marked nnir-current-group-marked)
+ (or (null server)
+ (equal server nnir-current-server)))
+ nnir-artlist
+ ;; Cache miss.
+ (setq nnir-artlist (nnir-run-query group)))
+ (save-excursion
+ (set-buffer nntp-server-buffer)
+ (if (zerop (length nnir-artlist))
+ (progn
+ (setq nnir-current-query nil
+ nnir-current-server nil
+ nnir-current-group-marked nil
+ nnir-artlist nil)
+ (nnheader-report 'nnir "Search produced empty results."))
+ ;; Remember data for cache.
+ (setq nnir-current-query group)
+ (when server (setq nnir-current-server server))
+ (setq nnir-current-group-marked gnus-group-marked)
+ (nnheader-insert "211 %d %d %d %s\n"
+ (nnir-artlist-length nnir-artlist) ; total #
+ 1 ; first #
+ (nnir-artlist-length nnir-artlist) ; last #
+ group)))) ; group name
+
+(deffoo nnir-retrieve-headers (articles &optional group server fetch-old)
+ (save-excursion
+ (let ((artlist (copy-sequence articles))
+ art artitem artgroup artno artrsv artfullgroup
+ novitem novdata foo server)
+ (while (not (null artlist))
+ (setq art (car artlist))
+ (or (numberp art)
+ (nnheader-report
+ 'nnir
+ "nnir-retrieve-headers doesn't grok message ids: %s"
+ art))
+ (setq artitem (nnir-artlist-article nnir-artlist art))
+ (setq artrsv (nnir-artitem-rsv artitem))
+ (setq artfullgroup (nnir-artitem-group artitem))
+ (setq artno (nnir-artitem-number artitem))
+ (setq artgroup (gnus-group-real-name artfullgroup))
+ (setq server (nnir-group-server artfullgroup))
+ ;; retrieve NOV or HEAD data for this article, transform into
+ ;; NOV data and prepend to `novdata'
+ (set-buffer nntp-server-buffer)
+ (nnir-possibly-change-server server)
+ (let ((gnus-override-method
+ (gnus-server-to-method server)))
+ (case (setq foo (gnus-retrieve-headers (list artno) artfullgroup nil))
+ (nov
+ (goto-char (point-min))
+ (setq novitem (nnheader-parse-nov))
+ (unless novitem
+ (pop-to-buffer nntp-server-buffer)
+ (error
+ "nnheader-parse-nov returned nil for article %s in group %s"
+ artno artfullgroup)))
+ (headers
+ (goto-char (point-min))
+ (setq novitem (nnheader-parse-head))
+ (unless novitem
+ (pop-to-buffer nntp-server-buffer)
+ (error
+ "nnheader-parse-head returned nil for article %s in group %s"
+ artno artfullgroup)))
+ (t (error "Unknown header type %s while requesting article %s of group %s"
+ foo artno artfullgroup))))
+ ;; replace article number in original group with article number
+ ;; in nnir group
+ (mail-header-set-number novitem art)
+ (mail-header-set-from novitem
+ (mail-header-from novitem))
+ (mail-header-set-subject
+ novitem
+ (format "[%d: %s/%d] %s"
+ artrsv artgroup artno
+ (mail-header-subject novitem)))
+ ;;-(mail-header-set-extra novitem nil)
+ (push novitem novdata)
+ (setq artlist (cdr artlist)))
+ (setq novdata (nreverse novdata))
+ (set-buffer nntp-server-buffer) (erase-buffer)
+ (mapc 'nnheader-insert-nov novdata)
+ 'nov)))
+
+(deffoo nnir-request-article (article
+ &optional group server to-buffer)
+ (if (stringp article)
+ (nnheader-report
+ 'nnir
+ "nnir-retrieve-headers doesn't grok message ids: %s"
+ article)
+ (save-excursion
+ (let* ((artitem (nnir-artlist-article nnir-artlist
+ article))
+ (artfullgroup (nnir-artitem-group artitem))
+ (artno (nnir-artitem-number artitem))
+ ;; Bug?
+ ;; Why must we bind nntp-server-buffer here? It won't
+ ;; work if `buf' is used, say. (Of course, the set-buffer
+ ;; line below must then be updated, too.)
+ (nntp-server-buffer (or to-buffer nntp-server-buffer)))
+ (set-buffer nntp-server-buffer)
+ (erase-buffer)
+ (message "Requesting article %d from group %s"
+ artno artfullgroup)
+ (gnus-request-article artno artfullgroup nntp-server-buffer)
+ (cons artfullgroup artno)))))
+
+
+(nnoo-define-skeleton nnir)
+
+
+(defmacro nnir-add-result (dirnam artno score prefix server artlist)
+ "Ask `nnir-compose-result' to construct a result vector,
+and if it is non-nil, add it to artlist."
+ `(let ((result (nnir-compose-result ,dirnam ,artno ,score ,prefix ,server)))
+ (when (not (null result))
+ (push result ,artlist))))
+
+(autoload 'nnmaildir-base-name-to-article-number "nnmaildir")
+
+;; Helper function currently used by the Swish++ and Namazu backends;
+;; perhaps useful for other backends as well
+(defun nnir-compose-result (dirnam article score prefix server)
+ "Extract the group from dirnam, and create a result vector
+ready to be added to the list of search results."
+
+ ;; remove nnir-*-remove-prefix from beginning of dirnam filename
+ (when (string-match (concat "^" prefix) dirnam)
+ (setq dirnam (replace-match "" t t dirnam)))
+
+ (when (file-readable-p (concat prefix dirnam article))
+ ;; remove trailing slash and, for nnmaildir, cur/new/tmp
+ (setq dirnam
+ (substring dirnam 0 (if (string= server "nnmaildir:") -5 -1)))
+
+ ;; Set group to dirnam without any leading dots or slashes,
+ ;; and with all subsequent slashes replaced by dots
+ (let ((group (gnus-replace-in-string
+ (gnus-replace-in-string dirnam "^[./\\]" "" t)
+ "[/\\]" "." t)))
+
+ (vector (nnir-group-full-name group server)
+ (if (string= server "nnmaildir:")
+ (nnmaildir-base-name-to-article-number
+ (substring article 0 (string-match ":" article))
+ group nil)
+ (string-to-number article))
+ (string-to-number score)))))
+
+;;; Search Engine Interfaces:
+
+;; freeWAIS-sf interface.
+(defun nnir-run-waissearch (query server &optional group)
+ "Run given query agains waissearch. Returns vector of (group name, file name)
+pairs (also vectors, actually)."
+ (when group
+ (error "The freeWAIS-sf backend cannot search specific groups."))
+ (save-excursion
+ (let ((qstring (cdr (assq 'query query)))
+ (prefix (nnir-read-server-parm 'nnir-wais-remove-prefix server))
+ artlist score artno dirnam)
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+ (message "Doing WAIS query %s..." query)
+ (call-process nnir-wais-program
+ nil ; input from /dev/null
+ t ; output to current buffer
+ nil ; don't redisplay
+ "-d" (nnir-read-server-parm 'nnir-wais-database server) ; database to search
+ qstring)
+ (message "Massaging waissearch output...")
+ ;; remove superfluous lines
+ (keep-lines "Score:")
+ ;; extract data from result lines
+ (goto-char (point-min))
+ (while (re-search-forward
+ "Score: +\\([0-9]+\\).*'\\([0-9]+\\) +\\([^']+\\)/'" nil t)
+ (setq score (match-string 1)
+ artno (match-string 2)
+ dirnam (match-string 3))
+ (unless (string-match prefix dirnam)
+ (nnheader-report 'nnir "Dir name %s doesn't contain prefix %s"
+ dirnam prefix))
+ (setq group (substitute ?. ?/ (replace-match "" t t dirnam)))
+ (push (vector (nnir-group-full-name group server)
+ (string-to-number artno)
+ (string-to-number score))
+ artlist))
+ (message "Massaging waissearch output...done")
+ (apply 'vector
+ (sort* artlist
+ (function (lambda (x y)
+ (> (nnir-artitem-rsv x)
+ (nnir-artitem-rsv y)))))))))
+
+;; IMAP interface.
+;; todo:
+;; nnir invokes this two (2) times???!
+;; we should not use nnimap at all but open our own server connection
+;; we should not LIST * but use nnimap-list-pattern from defs
+;; send queries as literals
+;; handle errors
+
+(autoload 'nnimap-open-server "nnimap")
+(defvar nnimap-server-buffer) ;; nnimap.el
+(autoload 'imap-mailbox-select "imap")
+(autoload 'imap-search "imap")
+(autoload 'imap-quote-specials "imap")
+
+(defun nnir-run-imap (query srv &optional group-option)
+ "Run a search against an IMAP back-end server.
+This uses a custom query language parser; see `nnir-imap-make-query' for
+details on the language and supported extensions"
+ (save-excursion
+ (let ((qstring (cdr (assq 'query query)))
+ (server (cadr (gnus-server-to-method srv)))
+ (group (or group-option (gnus-group-group-name)))
+ (defs (caddr (gnus-server-to-method srv)))
+ (criteria (or (cdr (assq 'criteria query))
+ nnir-imap-search-field))
+ artlist buf)
+ (message "Opening server %s" server)
+ (condition-case ()
+ (when (nnimap-open-server server defs) ;; xxx
+ (setq buf nnimap-server-buffer) ;; xxx
+ (message "Searching %s..." group)
+ (let ((arts 0)
+ (mbx (gnus-group-real-name group)))
+ (when (imap-mailbox-select mbx nil buf)
+ (mapc
+ (lambda (artnum)
+ (push (vector group artnum 1) artlist)
+ (setq arts (1+ arts)))
+ (imap-search (nnir-imap-make-query criteria qstring) buf))
+ (message "Searching %s... %d matches" mbx arts)))
+ (message "Searching %s...done" group))
+ (quit nil))
+ (reverse artlist))))
+
+(defun nnir-imap-make-query (criteria qstring)
+ "Parse the query string and criteria into an appropriate IMAP search
+expression, returning the string query to make.
+
+This implements a little language designed to return the expected results
+to an arbitrary query string to the end user.
+
+The search is always case-insensitive, as defined by RFC2060, and supports
+the following features (inspired by the Google search input language):
+
+Automatic \"and\" queries
+ If you specify multiple words then they will be treated as an \"and\"
+ expression intended to match all components.
+
+Phrase searches
+ If you wrap your query in double-quotes then it will be treated as a
+ literal string.
+
+Negative terms
+ If you precede a term with \"-\" then it will negate that.
+
+\"OR\" queries
+ If you include an upper-case \"OR\" in your search it will cause the
+ term before it and the term after it to be treated as alternatives.
+
+In future the following will be added to the language:
+ * support for date matches
+ * support for location of text matching within the query
+ * from/to/etc headers
+ * additional search terms
+ * flag based searching
+ * anything else that the RFC supports, basically."
+ ;; Walk through the query and turn it into an IMAP query string.
+ (nnir-imap-query-to-imap criteria (nnir-imap-parse-query qstring)))
+
+
+(defun nnir-imap-query-to-imap (criteria query)
+ "Turn a s-expression format query into IMAP."
+ (mapconcat
+ ;; Turn the expressions into IMAP text
+ (lambda (item)
+ (nnir-imap-expr-to-imap criteria item))
+ ;; The query, already in s-expr format.
+ query
+ ;; Append a space between each expression
+ " "))
+
+
+(defun nnir-imap-expr-to-imap (criteria expr)
+ "Convert EXPR into an IMAP search expression on CRITERIA"
+ ;; What sort of expression is this, eh?
+ (cond
+ ;; Simple string term
+ ((stringp expr)
+ (format "%s \"%s\"" criteria (imap-quote-specials expr)))
+ ;; Trivial term: and
+ ((eq expr 'and) nil)
+ ;; Composite term: or expression
+ ((eq (car-safe expr) 'or)
+ (format "OR %s %s"
+ (nnir-imap-expr-to-imap criteria (second expr))
+ (nnir-imap-expr-to-imap criteria (third expr))))
+ ;; Composite term: just the fax, mam
+ ((eq (car-safe expr) 'not)
+ (format "NOT (%s)" (nnir-imap-query-to-imap criteria (rest expr))))
+ ;; Composite term: just expand it all.
+ ((and (not (null expr)) (listp expr))
+ (format "(%s)" (nnir-imap-query-to-imap criteria expr)))
+ ;; Complex value, give up for now.
+ (t (error "Unhandled input: %S" expr))))
+
+
+(defun nnir-imap-parse-query (string)
+ "Turn STRING into an s-expression based query based on the IMAP
+query language as defined in `nnir-imap-make-query'.
+
+This involves turning individual tokens into higher level terms
+that the search language can then understand and use."
+ (with-temp-buffer
+ ;; Set up the parsing environment.
+ (insert string)
+ (goto-char (point-min))
+ ;; Now, collect the output terms and return them.
+ (let (out)
+ (while (not (nnir-imap-end-of-input))
+ (push (nnir-imap-next-expr) out))
+ (reverse out))))
+
+
+(defun nnir-imap-next-expr (&optional count)
+ "Return the next expression from the current buffer."
+ (let ((term (nnir-imap-next-term count))
+ (next (nnir-imap-peek-symbol)))
+ ;; Are we looking at an 'or' expression?
+ (cond
+ ;; Handle 'expr or expr'
+ ((eq next 'or)
+ (list 'or term (nnir-imap-next-expr 2)))
+ ;; Anything else
+ (t term))))
+
+
+(defun nnir-imap-next-term (&optional count)
+ "Return the next TERM from the current buffer."
+ (let ((term (nnir-imap-next-symbol count)))
+ ;; What sort of term is this?
+ (cond
+ ;; and -- just ignore it
+ ((eq term 'and) 'and)
+ ;; negated term
+ ((eq term 'not) (list 'not (nnir-imap-next-expr)))
+ ;; generic term
+ (t term))))
+
+
+(defun nnir-imap-peek-symbol ()
+ "Return the next symbol from the current buffer, but don't consume it."
+ (save-excursion
+ (nnir-imap-next-symbol)))
+
+(defun nnir-imap-next-symbol (&optional count)
+ "Return the next symbol from the current buffer, or nil if we are
+at the end of the buffer. If supplied COUNT skips some symbols before
+returning the one at the supplied position."
+ (when (and (numberp count) (> count 1))
+ (nnir-imap-next-symbol (1- count)))
+ (let ((case-fold-search t))
+ ;; end of input stream?
+ (unless (nnir-imap-end-of-input)
+ ;; No, return the next symbol from the stream.
+ (cond
+ ;; negated expression -- return it and advance one char.
+ ((looking-at "-") (forward-char 1) 'not)
+ ;; quoted string
+ ((looking-at "\"") (nnir-imap-delimited-string "\""))
+ ;; list expression -- we parse the content and return this as a list.
+ ((looking-at "(")
+ (nnir-imap-parse-query (nnir-imap-delimited-string ")")))
+ ;; keyword input -- return a symbol version
+ ((looking-at "\\band\\b") (forward-char 3) 'and)
+ ((looking-at "\\bor\\b") (forward-char 2) 'or)
+ ((looking-at "\\bnot\\b") (forward-char 3) 'not)
+ ;; Simple, boring keyword
+ (t (let ((start (point))
+ (end (if (search-forward-regexp "[[:blank:]]" nil t)
+ (prog1
+ (match-beginning 0)
+ ;; unskip if we hit a non-blank terminal character.
+ (when (string-match "[^[:blank:]]" (match-string 0))
+ (backward-char 1)))
+ (goto-char (point-max)))))
+ (buffer-substring start end)))))))
+
+(defun nnir-imap-delimited-string (delimiter)
+ "Return a delimited string from the current buffer."
+ (let ((start (point)) end)
+ (forward-char 1) ; skip the first delimiter.
+ (while (not end)
+ (unless (search-forward delimiter nil t)
+ (error "Unmatched delimited input with %s in query" delimiter))
+ (let ((here (point)))
+ (unless (equal (buffer-substring (- here 2) (- here 1)) "\\")
+ (setq end (point)))))
+ (buffer-substring (1+ start) (1- end))))
+
+(defun nnir-imap-end-of-input ()
+ "Are we at the end of input?"
+ (skip-chars-forward "[[:blank:]]")
+ (looking-at "$"))
+
+
+;; Swish++ interface.
+;; -cc- Todo
+;; Search by
+;; - group
+;; Sort by
+;; - rank (default)
+;; - article number
+;; - file size
+;; - group
+(defun nnir-run-swish++ (query server &optional group)
+ "Run QUERY against swish++.
+Returns a vector of (group name, file name) pairs (also vectors,
+actually).
+
+Tested with swish++ 4.7 on GNU/Linux and with swish++ 5.0b2 on
+Windows NT 4.0."
+
+ (when group
+ (error "The swish++ backend cannot search specific groups."))
+
+ (save-excursion
+ (let ( (qstring (cdr (assq 'query query)))
+ (groupspec (cdr (assq 'group query)))
+ (prefix (nnir-read-server-parm 'nnir-swish++-remove-prefix server))
+ artlist
+ ;; nnml-use-compressed-files might be any string, but probably this
+ ;; is sufficient. Note that we can't only use the value of
+ ;; nnml-use-compressed-files because old articles might have been
+ ;; saved with a different value.
+ (article-pattern (if (string= server "nnmaildir:")
+ ":[0-9]+"
+ "^[0-9]+\\(\\.[a-z0-9]+\\)?$"))
+ score artno dirnam filenam)
+
+ (when (equal "" qstring)
+ (error "swish++: You didn't enter anything."))
+
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+
+ (if groupspec
+ (message "Doing swish++ query %s on %s..." qstring groupspec)
+ (message "Doing swish++ query %s..." qstring))
+
+ (let* ((cp-list `( ,nnir-swish++-program
+ nil ; input from /dev/null
+ t ; output
+ nil ; don't redisplay
+ "--config-file" ,(nnir-read-server-parm 'nnir-swish++-configuration-file server)
+ ,@(nnir-read-server-parm 'nnir-swish++-additional-switches server)
+ ,qstring ; the query, in swish++ format
+ ))
+ (exitstatus
+ (progn
+ (message "%s args: %s" nnir-swish++-program
+ (mapconcat 'identity (cddddr cp-list) " ")) ;; ???
+ (apply 'call-process cp-list))))
+ (unless (or (null exitstatus)
+ (zerop exitstatus))
+ (nnheader-report 'nnir "Couldn't run swish++: %s" exitstatus)
+ ;; swish++ failure reason is in this buffer, show it if
+ ;; the user wants it.
+ (when (> gnus-verbose 6)
+ (display-buffer nnir-tmp-buffer))))
+
+ ;; The results are output in the format of:
+ ;; V 4.7 Linux
+ ;; rank relative-path-name file-size file-title
+ ;; V 5.0b2:
+ ;; rank relative-path-name file-size topic??
+ ;; where rank is an integer from 1 to 100.
+ (goto-char (point-min))
+ (while (re-search-forward
+ "\\(^[0-9]+\\) \\([^ ]+\\) [0-9]+ \\(.*\\)$" nil t)
+ (setq score (match-string 1)
+ filenam (match-string 2)
+ artno (file-name-nondirectory filenam)
+ dirnam (file-name-directory filenam))
+
+ ;; don't match directories
+ (when (string-match article-pattern artno)
+ (when (not (null dirnam))
+
+ ;; maybe limit results to matching groups.
+ (when (or (not groupspec)
+ (string-match groupspec dirnam))
+ (nnir-add-result dirnam artno score prefix server artlist)))))
+
+ (message "Massaging swish++ output...done")
+
+ ;; Sort by score
+ (apply 'vector
+ (sort* artlist
+ (function (lambda (x y)
+ (> (nnir-artitem-rsv x)
+ (nnir-artitem-rsv y)))))))))
+
+;; Swish-E interface.
+(defun nnir-run-swish-e (query server &optional group)
+ "Run given query against swish-e.
+Returns a vector of (group name, file name) pairs (also vectors,
+actually).
+
+Tested with swish-e-2.0.1 on Windows NT 4.0."
+
+ ;; swish-e crashes with empty parameter to "-w" on commandline...
+ (when group
+ (error "The swish-e backend cannot search specific groups."))
+
+ (save-excursion
+ (let ((qstring (cdr (assq 'query query)))
+ (prefix
+ (or (nnir-read-server-parm 'nnir-swish-e-remove-prefix server)
+ (error "Missing parameter `nnir-swish-e-remove-prefix'")))
+ artlist score artno dirnam group )
+
+ (when (equal "" qstring)
+ (error "swish-e: You didn't enter anything."))
+
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+
+ (message "Doing swish-e query %s..." query)
+ (let* ((index-files
+ (or (nnir-read-server-parm
+ 'nnir-swish-e-index-files server)
+ (error "Missing parameter `nnir-swish-e-index-files'")))
+ (additional-switches
+ (nnir-read-server-parm
+ 'nnir-swish-e-additional-switches server))
+ (cp-list `(,nnir-swish-e-program
+ nil ; input from /dev/null
+ t ; output
+ nil ; don't redisplay
+ "-f" ,@index-files
+ ,@additional-switches
+ "-w"
+ ,qstring ; the query, in swish-e format
+ ))
+ (exitstatus
+ (progn
+ (message "%s args: %s" nnir-swish-e-program
+ (mapconcat 'identity (cddddr cp-list) " "))
+ (apply 'call-process cp-list))))
+ (unless (or (null exitstatus)
+ (zerop exitstatus))
+ (nnheader-report 'nnir "Couldn't run swish-e: %s" exitstatus)
+ ;; swish-e failure reason is in this buffer, show it if
+ ;; the user wants it.
+ (when (> gnus-verbose 6)
+ (display-buffer nnir-tmp-buffer))))
+
+ ;; The results are output in the format of:
+ ;; rank path-name file-title file-size
+ (goto-char (point-min))
+ (while (re-search-forward
+ "\\(^[0-9]+\\) \\([^ ]+\\) \"\\([^\"]+\\)\" [0-9]+$" nil t)
+ (setq score (match-string 1)
+ artno (match-string 3)
+ dirnam (file-name-directory (match-string 2)))
+
+ ;; don't match directories
+ (when (string-match "^[0-9]+$" artno)
+ (when (not (null dirnam))
+
+ ;; remove nnir-swish-e-remove-prefix from beginning of dirname
+ (when (string-match (concat "^" prefix) dirnam)
+ (setq dirnam (replace-match "" t t dirnam)))
+
+ (setq dirnam (substring dirnam 0 -1))
+ ;; eliminate all ".", "/", "\" from beginning. Always matches.
+ (string-match "^[./\\]*\\(.*\\)$" dirnam)
+ ;; "/" -> "."
+ (setq group (substitute ?. ?/ (match-string 1 dirnam)))
+ ;; Windows "\\" -> "."
+ (setq group (substitute ?. ?\\ group))
+
+ (push (vector (nnir-group-full-name group server)
+ (string-to-number artno)
+ (string-to-number score))
+ artlist))))
+
+ (message "Massaging swish-e output...done")
+
+ ;; Sort by score
+ (apply 'vector
+ (sort* artlist
+ (function (lambda (x y)
+ (> (nnir-artitem-rsv x)
+ (nnir-artitem-rsv y)))))))))
+
+;; HyREX interface
+(defun nnir-run-hyrex (query server &optional group)
+ (save-excursion
+ (let ((artlist nil)
+ (groupspec (cdr (assq 'group query)))
+ (qstring (cdr (assq 'query query)))
+ (prefix (nnir-read-server-parm 'nnir-hyrex-remove-prefix server))
+ score artno dirnam)
+ (when (and group groupspec)
+ (error (concat "It does not make sense to use a group spec"
+ " with process-marked groups.")))
+ (when group
+ (setq groupspec (gnus-group-real-name group)))
+ (when (and group (not (equal group (nnir-group-full-name groupspec server))))
+ (message "%s vs. %s" group (nnir-group-full-name groupspec server))
+ (error "Server with groupspec doesn't match group !"))
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+ (if groupspec
+ (message "Doing hyrex-search query %s on %s..." query groupspec)
+ (message "Doing hyrex-search query %s..." query))
+ (let* ((cp-list
+ `( ,nnir-hyrex-program
+ nil ; input from /dev/null
+ t ; output
+ nil ; don't redisplay
+ "-i",(nnir-read-server-parm 'nnir-hyrex-index-directory server) ; index directory
+ ,@(nnir-read-server-parm 'nnir-hyrex-additional-switches server)
+ ,qstring ; the query, in hyrex-search format
+ ))
+ (exitstatus
+ (progn
+ (message "%s args: %s" nnir-hyrex-program
+ (mapconcat 'identity (cddddr cp-list) " "))
+ (apply 'call-process cp-list))))
+ (unless (or (null exitstatus)
+ (zerop exitstatus))
+ (nnheader-report 'nnir "Couldn't run hyrex-search: %s" exitstatus)
+ ;; nnir-search failure reason is in this buffer, show it if
+ ;; the user wants it.
+ (when (> gnus-verbose 6)
+ (display-buffer nnir-tmp-buffer)))) ;; FIXME: Dont clear buffer !
+ (if groupspec
+ (message "Doing hyrex-search query \"%s\" on %s...done" qstring groupspec)
+ (message "Doing hyrex-search query \"%s\"...done" qstring))
+ (sit-for 0)
+ ;; nnir-search returns:
+ ;; for nnml/nnfolder: "filename mailid weigth"
+ ;; for nnimap: "group mailid weigth"
+ (goto-char (point-min))
+ (delete-non-matching-lines "^\\S + [0-9]+ [0-9]+$")
+ ;; HyREX couldn't search directly in groups -- so filter out here.
+ (when groupspec
+ (keep-lines groupspec))
+ ;; extract data from result lines
+ (goto-char (point-min))
+ (while (re-search-forward
+ "\\(\\S +\\) \\([0-9]+\\) \\([0-9]+\\)" nil t)
+ (setq dirnam (match-string 1)
+ artno (match-string 2)
+ score (match-string 3))
+ (when (string-match prefix dirnam)
+ (setq dirnam (replace-match "" t t dirnam)))
+ (push (vector (nnir-group-full-name (substitute ?. ?/ dirnam) server)
+ (string-to-number artno)
+ (string-to-number score))
+ artlist))
+ (message "Massaging hyrex-search output...done.")
+ (apply 'vector
+ (sort* artlist
+ (function (lambda (x y)
+ (if (string-lessp (nnir-artitem-group x)
+ (nnir-artitem-group y))
+ t
+ (< (nnir-artitem-number x)
+ (nnir-artitem-number y)))))))
+ )))
+
+;; Namazu interface
+(defun nnir-run-namazu (query server &optional group)
+ "Run given query against Namazu. Returns a vector of (group name, file name)
+pairs (also vectors, actually).
+
+Tested with Namazu 2.0.6 on a GNU/Linux system."
+ (when group
+ (error "The Namazu backend cannot search specific groups"))
+ (save-excursion
+ (let ((article-pattern (if (string= server "nnmaildir:")
+ ":[0-9]+"
+ "^[0-9]+$"))
+ artlist
+ (qstring (cdr (assq 'query query)))
+ (prefix (nnir-read-server-parm 'nnir-namazu-remove-prefix server))
+ score group article
+ (process-environment (copy-sequence process-environment)))
+ (setenv "LC_MESSAGES" "C")
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+ (let* ((cp-list
+ `( ,nnir-namazu-program
+ nil ; input from /dev/null
+ t ; output
+ nil ; don't redisplay
+ "-q" ; don't be verbose
+ "-a" ; show all matches
+ "-s" ; use short format
+ ,@(nnir-read-server-parm 'nnir-namazu-additional-switches server)
+ ,qstring ; the query, in namazu format
+ ,(nnir-read-server-parm 'nnir-namazu-index-directory server) ; index directory
+ ))
+ (exitstatus
+ (progn
+ (message "%s args: %s" nnir-namazu-program
+ (mapconcat 'identity (cddddr cp-list) " "))
+ (apply 'call-process cp-list))))
+ (unless (or (null exitstatus)
+ (zerop exitstatus))
+ (nnheader-report 'nnir "Couldn't run namazu: %s" exitstatus)
+ ;; Namazu failure reason is in this buffer, show it if
+ ;; the user wants it.
+ (when (> gnus-verbose 6)
+ (display-buffer nnir-tmp-buffer))))
+
+ ;; Namazu output looks something like this:
+ ;; 2. Re: Gnus agent expire broken (score: 55)
+ ;; /home/henrik/Mail/mail/sent/1310 (4,138 bytes)
+
+ (goto-char (point-min))
+ (while (re-search-forward
+ "^\\([0-9]+\\.\\).*\\((score: \\([0-9]+\\)\\))\n\\([^ ]+\\)"
+ nil t)
+ (setq score (match-string 3)
+ group (file-name-directory (match-string 4))
+ article (file-name-nondirectory (match-string 4)))
+
+ ;; make sure article and group is sane
+ (when (and (string-match article-pattern article)
+ (not (null group)))
+ (nnir-add-result group article score prefix server artlist)))
+
+ ;; sort artlist by score
+ (apply 'vector
+ (sort* artlist
+ (function (lambda (x y)
+ (> (nnir-artitem-rsv x)
+ (nnir-artitem-rsv y)))))))))
+
+(defun nnir-run-find-grep (query server &optional group)
+ "Run find and grep to obtain matching articles."
+ (let* ((method (gnus-server-to-method server))
+ (sym (intern
+ (concat (symbol-name (car method)) "-directory")))
+ (directory (cadr (assoc sym (cddr method))))
+ (regexp (cdr (assoc 'query query)))
+ (grep-options (cdr (assoc 'grep-options query)))
+ artlist)
+ (unless directory
+ (error "No directory found in method specification of server %s"
+ server))
+ (message "Searching %s using find-grep..." (or group server))
+ (save-window-excursion
+ (set-buffer (get-buffer-create nnir-tmp-buffer))
+ (erase-buffer)
+ (if (> gnus-verbose 6)
+ (pop-to-buffer (current-buffer)))
+ (cd directory) ; Using relative paths simplifies postprocessing.
+ (let ((group
+ (if (not group)
+ "."
+ ;; Try accessing the group literally as well as
+ ;; interpreting dots as directory separators so the
+ ;; engine works with plain nnml as well as the Gnus
+ ;; Cache.
+ (find-if 'file-directory-p
+ (let ((group (gnus-group-real-name group)))
+ (list group (gnus-replace-in-string group "\\." "/" t)))))))
+ (unless group
+ (error "Cannot locate directory for group"))
+ (save-excursion
+ (apply
+ 'call-process "find" nil t
+ "find" group "-type" "f" "-name" "[0-9]*" "-exec"
+ "grep"
+ `("-l" ,@(and grep-options (split-string grep-options "\\s-" t))
+ "-e" ,regexp "{}" "+"))))
+
+ ;; Translate relative paths to group names.
+ (while (not (eobp))
+ (let* ((path (split-string
+ (buffer-substring (point) (line-end-position)) "/" t))
+ (art (string-to-number (car (last path)))))
+ (while (string= "." (car path))
+ (setq path (cdr path)))
+ (let ((group (mapconcat 'identity (subseq path 0 -1) ".")))
+ (push (vector (nnir-group-full-name group server) art 0)
+ artlist))
+ (forward-line 1)))
+ (message "Searching %s using find-grep...done" (or group server))
+ artlist)))
+
+;;; Util Code:
+
+(defun nnir-read-parms (query)
+ "Reads additional search parameters according to `nnir-engines'."
+ (let ((parmspec (caddr (assoc nnir-search-engine nnir-engines))))
+ (cons (cons 'query query)
+ (mapcar 'nnir-read-parm parmspec))))
+
+(defun nnir-read-parm (parmspec)
+ "Reads a single search parameter.
+`parmspec' is a cons cell, the car is a symbol, the cdr is a prompt."
+ (let ((sym (car parmspec))
+ (prompt (cdr parmspec)))
+ (if (listp prompt)
+ (let* ((result (apply 'completing-read prompt))
+ (mapping (or (assoc result nnir-imap-search-arguments)
+ (assoc nil nnir-imap-search-arguments))))
+ (cons sym (format (cdr mapping) result)))
+ (cons sym (read-string prompt)))))
+
+(defun nnir-run-query (query)
+ "Invoke appropriate search engine function (see `nnir-engines').
+If some groups were process-marked, run the query for each of the groups
+and concat the results."
+ (let ((q (car (read-from-string query))))
+ (if gnus-group-marked
+ (apply 'vconcat
+ (mapcar (lambda (x)
+ (let ((server (nnir-group-server x))
+ search-func)
+ (setq search-func (cadr
+ (assoc
+ (nnir-read-server-parm 'nnir-search-engine server) nnir-engines)))
+ (if search-func
+ (funcall search-func q server x)
+ nil)))
+ gnus-group-marked)
+ )
+ (apply 'vconcat
+ (mapcar (lambda (x)
+ (if (and (equal (cadr x) 'ok) (not (equal (cadar x) "-ephemeral")))
+ (let ((server (format "%s:%s" (caar x) (cadar x)))
+ search-func)
+ (setq search-func (cadr
+ (assoc
+ (nnir-read-server-parm 'nnir-search-engine server) nnir-engines)))
+ (if search-func
+ (funcall search-func q server nil)
+ nil))
+ nil))
+ gnus-opened-servers)
+ ))
+ ))
+
+(defun nnir-read-server-parm (key server)
+ "Returns the parameter value of for the given server, where server is of
+form 'backend:name'."
+ (let ((method (gnus-server-to-method server)))
+ (cond ((and method (assq key (cddr method)))
+ (nth 1 (assq key (cddr method))))
+ ((and nnir-mail-backend
+ (gnus-server-equal method nnir-mail-backend))
+ (symbol-value key))
+ (t nil))))
+;; (if method
+;; (if (assq key (cddr method))
+;; (nth 1 (assq key (cddr method)))
+;; (symbol-value key))
+;; (symbol-value key))
+;; ))
+
+(defun nnir-group-full-name (shortname server)
+ "For the given group name, return a full Gnus group name.
+The Gnus backend/server information is added."
+ (gnus-group-prefixed-name shortname (gnus-server-to-method server)))
+
+(defun nnir-possibly-change-server (server)
+ (unless (and server (nnir-server-opened server))
+ (nnir-open-server server)))
+
+
+;; Data type article list.
+
+(defun nnir-artlist-length (artlist)
+ "Returns number of articles in artlist."
+ (length artlist))
+
+(defun nnir-artlist-article (artlist n)
+ "Returns from ARTLIST the Nth artitem (counting starting at 1)."
+ (elt artlist (1- n)))
+
+(defun nnir-artitem-group (artitem)
+ "Returns the group from the ARTITEM."
+ (elt artitem 0))
+
+(defun nnir-artlist-artitem-group (artlist n)
+ "Returns from ARTLIST the group of the Nth artitem (counting from 1)."
+ (nnir-artitem-group (nnir-artlist-article artlist n)))
+
+(defun nnir-artitem-number (artitem)
+ "Returns the number from the ARTITEM."
+ (elt artitem 1))
+
+(defun nnir-artlist-artitem-number (artlist n)
+ "Returns from ARTLIST the number of the Nth artitem (counting from 1)."
+ (nnir-artitem-number (nnir-artlist-article artlist n)))
+
+(defun nnir-artitem-rsv (artitem)
+ "Returns the Retrieval Status Value (RSV, score) from the ARTITEM."
+ (elt artitem 2))
+
+(defun nnir-artlist-artitem-rsv (artlist n)
+ "Returns from ARTLIST the Retrieval Status Value of the Nth artitem
+\(counting from 1)."
+ (nnir-artitem-rsv (nnir-artlist-article artlist n)))
+
+;; unused?
+(defun nnir-artlist-groups (artlist)
+ "Returns a list of all groups in the given ARTLIST."
+ (let ((res nil)
+ (with-dups nil))
+ ;; from each artitem, extract group component
+ (setq with-dups (mapcar 'nnir-artitem-group artlist))
+ ;; remove duplicates from above
+ (mapc (function (lambda (x) (add-to-list 'res x)))
+ with-dups)
+ res))
+
+
+;; The end.
+(provide 'nnir)
+
+;;; arch-tag: 9b3fecf8-4397-4bbb-bf3c-6ac3cbbc6664
projects / emacs.git / commitdiff