diff options
Diffstat (limited to 'lisp/mh-e/mh-index.el')
| -rw-r--r-- | lisp/mh-e/mh-index.el | 290 |
1 files changed, 162 insertions, 128 deletions
diff --git a/lisp/mh-e/mh-index.el b/lisp/mh-e/mh-index.el index cde630236b0..7a52b94dd2b 100644 --- a/lisp/mh-e/mh-index.el +++ b/lisp/mh-e/mh-index.el @@ -1,6 +1,6 @@ ;;; mh-index -- MH-E interface to indexing programs -;; Copyright (C) 2005 Free Software Foundation, Inc. +;; Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc. ;; Author: Satyaki Das <satyaki@theforce.stanford.edu> ;; Maintainer: Bill Wohler <wohler@newt.com> @@ -31,6 +31,7 @@ ;;; swish-e ;;; mairix ;;; namazu +;;; pick ;;; grep ;;; ;;; (2) To use this package, you first have to build an index. Please read @@ -352,43 +353,74 @@ they are concatenated to construct the base name." (defun* mh-index-search (redo-search-flag folder search-regexp &optional window-config) "Perform an indexed search in an MH mail folder. -Use a prefix argument to repeat the search, as in REDO-SEARCH-FLAG below. - -If REDO-SEARCH-FLAG is non-nil and the current folder buffer was generated by a -index search, then the search is repeated. Otherwise, FOLDER is searched with -SEARCH-REGEXP and the results are presented in an MH-E folder. If FOLDER is -\"+\" then mail in all folders are searched. Optional argument WINDOW-CONFIG -stores the window configuration that will be restored after the user quits the -folder containing the index search results. - -Four indexing programs are supported; if none of these are present, then grep -is used. This function picks the first program that is available on your -system. If you would prefer to use a different program, set the customization -variable `mh-index-program' accordingly. - -The documentation for the following functions describes how to generate the -index for each program: +Use a prefix argument to repeat the search. + +Unlike regular searches, the prompt for the folder to search can be `all' to +search all folders; in addition, the search works recursively on the listed +folder. The search criteria are entered in an MH-Pick buffer as described in +`mh-search-folder'. + +To perform the search, type \\<mh-pick-mode-map>\\[mh-do-search]. Another +difference from the regular searches is that because the search operates on +more than one folder, the messages that are found are put in a temporary +sub-folder of `+mhe-index' and are displayed in an MH-Folder buffer. This +buffer is special because it displays messages from multiple folders; each set +of messages from a given folder has a heading with the folder name. + +In addition, the \\<mh-folder-mode-map>\\[mh-index-visit-folder] command can +be used to visit the folder of the message at point. Initially, only the +messages that matched the search criteria are displayed in the folder. While +the temporary buffer has its own set of message numbers, the actual messages +numbers are shown in the visited folder. Thus, the \\[mh-index-visit-folder] +command is useful to find the actual message number of an interesting message, +or to view surrounding messages with the \\[mh-rescan-folder] command. + +Because this folder is temporary, you'll probably get in the habit of killing +it when you're done with \\[mh-kill-folder]. + +If you have run the \\[mh-search-folder] command, but change your mind while +entering the search criteria and actually want to run an indexed search, then +you can use the \\<mh-pick-mode-map>\\[mh-index-do-search] command in the +MH-Pick buffer. + +The \\<mh-folder-mode-map>\\[mh-index-search] command runs the command defined +by the `mh-index-program' option. The default value is \"Auto-detect\" which +means that MH-E will automatically choose one of \"swish++\", \"swish-e\", +\"mairix\", \"namazu\", \"pick\" and \"grep\" in that order. If, for example, +you have both \"swish++\" and \"mairix\" installed and you want to use +\"mairix\", then you can set this option to \"mairix\". + + *NOTE* + + The \"pick\" and \"grep\" commands do not perform a recursive search on + the given folder. + +This command uses an \"X-MHE-Checksum:\" header field to cache the MD5 +checksum of a message. This means that if an incoming message already contains +an \"X-MHE-Checksum:\" field, that message might not be found by this command. +The following \"procmail\" recipe avoids this problem by renaming the existing +header field: + + :0 wf + | formail -R \"X-MHE-Checksum\" \"X-Old-MHE-Checksum\" + +The documentation for the following commands describe how to set up the +various indexing programs to use with MH-E. The \"pick\" and \"grep\" commands +do not require additional configuration. - `mh-swish++-execute-search' - `mh-swish-execute-search' - `mh-mairix-execute-search' - `mh-namazu-execute-search' - -If none of these programs are present then we use pick. If desired grep can be -used instead. Details about these methods can be found in: - - `mh-pick-execute-search' - `mh-grep-execute-search' -This and related functions use an X-MHE-Checksum header to cache the MD5 -checksum of a message. This means that already present X-MHE-Checksum headers -in the incoming email could result in messages not being found. The following -procmail recipe should avoid this: - - :0 wf - | formail -R \"X-MHE-Checksum\" \"Old-X-MHE-Checksum\" - -This has the effect of renaming already present X-MHE-Checksum headers." +In a program, if REDO-SEARCH-FLAG is non-nil and the current folder buffer was +generated by a index search, then the search is repeated. Otherwise, FOLDER is +searched with SEARCH-REGEXP and the results are presented in an MH-E folder. +If FOLDER is \"+\" then mail in all folders are searched. Optional argument +WINDOW-CONFIG stores the window configuration that will be restored after the +user quits the folder containing the index search results." (interactive (list current-prefix-arg (progn @@ -998,8 +1030,8 @@ Unlike the other index search programs \"pick\" only searches messages present in the folder itself and does not descend into any sub-folders that may be present. -FOLDER-PATH is the directory containing the mails to be searched and -SEARCH-REGEXP is the pattern that pick gets." +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP is used +to search." (set-buffer (get-buffer-create mh-index-temp-buffer)) (erase-buffer) (setq mh-index-pick-folder @@ -1029,7 +1061,13 @@ SEARCH-REGEXP is the pattern that pick gets." (defun mh-grep-execute-search (folder-path search-regexp) "Execute grep and read the results. -FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." + +Unlike the other index search programs \"grep\" only searches messages present +in the folder itself and does not descend into any sub-folders that may be +present. + +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP is used +to search." (set-buffer (get-buffer-create mh-index-temp-buffer)) (erase-buffer) (call-process mh-grep-binary nil '(t nil) nil @@ -1081,34 +1119,34 @@ other matches left then return nil. If the current record is invalid return (defun mh-mairix-execute-search (folder-path search-regexp-list) "Execute mairix and read the results. -In the examples below replace /home/user/Mail with the path to your MH +In the examples below, replace \"/home/user/Mail\" with the path to your MH directory. -First create the directory /home/user/Mail/.mairix. Then create the file -/home/user/Mail/.mairix/config with the following contents: +First create the directory \"/home/user/Mail/.mairix\". Then create the file +\"/home/user/Mail/.mairix/config\" with the following contents: - # This should contain the same thing as your `mh-user-path' - base=/home/user/Mail + base=/home/user/Mail - # List of folders that should be indexed. 3 dots at the end means there are - # subfolders within the folder - mh_folders=archive...:inbox:drafts:news:sent:trash + # List of folders that should be indexed. 3 dots at the end means there + # are subfolders within the folder + mh=archive...:inbox:drafts:news:sent:trash - vfolder_format=raw - database=/home/user/Mail/mairix/database + vfolder_format=raw + database=/home/user/Mail/mairix/database Use the following command line to generate the mairix index. Run this daily from cron: - mairix -f /home/user/Mail/.mairix/config + mairix -f /home/user/Mail/.mairix/config -FOLDER-PATH is the directory in which SEARCH-REGEXP-LIST is used to search." +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP-LIST is used +to search." (set-buffer (get-buffer-create mh-index-temp-buffer)) (erase-buffer) (unless mh-mairix-binary (error "Set mh-mairix-binary appropriately")) (apply #'call-process mh-mairix-binary nil '(t nil) nil - "-f" (format "%s%s/config" mh-user-path mh-mairix-directory) + "-r" "-f" (format "%s%s/config" mh-user-path mh-mairix-directory) search-regexp-list) (goto-char (point-min)) (setq mh-mairix-folder @@ -1167,10 +1205,10 @@ REGEXP-LIST is an alist of fields and values." (let ((expr-list (cdr conjunct)) (expr-string "")) (dolist (e expr-list) - (setq expr-string (concat expr-string "+" + (setq expr-string (concat expr-string "," (if (atom e) "" "~") (if (atom e) e (cadr e))))) - (setq final (concat final "," (substring expr-string 1))))) + (setq final (concat final "/" (substring expr-string 1))))) (substring final 1))) result))) result)) @@ -1327,52 +1365,49 @@ space-separated list of FOLDERS, or nothing to search all folders." (defun mh-swish-execute-search (folder-path search-regexp) "Execute swish-e and read the results. -In the examples below, replace /home/user/Mail with the path to your MH -directory. - -First create the directory /home/user/Mail/.swish. Then create the file -/home/user/Mail/.swish/config with the following contents: - - IndexDir /home/user/Mail - IndexFile /home/user/Mail/.swish/index - IndexName \"Mail Index\" - IndexDescription \"Mail Index\" - IndexPointer \"http://nowhere\" - IndexAdmin \"nobody\" - #MetaNames automatic - IndexReport 3 - FollowSymLinks no - UseStemming no - IgnoreTotalWordCountWhenRanking yes - WordCharacters abcdefghijklmnopqrstuvwxyz0123456789- - BeginCharacters abcdefghijklmnopqrstuvwxyz - EndCharacters abcdefghijklmnopqrstuvwxyz0123456789 - IgnoreLimit 50 1000 - IndexComments 0 - FileRules pathname contains /home/user/Mail/.swish - FileRules pathname contains /home/user/Mail/mhe-index - FileRules filename is index - FileRules filename is \\..* - FileRules filename is #.* - FileRules filename is ,.* - FileRules filename is .*~ +In the examples below, replace \"/home/user/Mail\" with the path to your +MH directory. + +First create the directory \"/home/user/Mail/.swish\". Then create the file +\"/home/user/Mail/.swish/config\" with the following contents: + + DefaultContents TXT* + IndexDir /home/user/Mail + IndexFile /home/user/Mail/.swish/index + IndexName \"Mail Index\" + IndexDescription \"Mail Index\" + IndexPointer \"http://nowhere\" + IndexAdmin \"nobody\" + #MetaNames automatic + IndexReport 3 + FollowSymLinks no + UseStemming no + IgnoreTotalWordCountWhenRanking yes + WordCharacters abcdefghijklmnopqrstuvwxyz0123456789- + BeginCharacters abcdefghijklmnopqrstuvwxyz + EndCharacters abcdefghijklmnopqrstuvwxyz0123456789 + IgnoreLimit 50 1000 + IndexComments 0 + FileRules filename contains \\D + FileRules pathname contains /home/user/Mail/.swish + FileRules pathname contains /home/user/Mail/mhe-index + +This configuration does not index the folders that hold the results of your +searches in \"+mhe-index\" since they tend to be ephemeral and the original +messages are indexed anyway. If there are any directories you would like to ignore, append lines like the -following to config: - - FileRules pathname contains /home/user/Mail/scripts +following to \"config\": -You do not want to index the folders that hold the results of your searches -since they tend to be ephemeral and the original messages are indexed anyway. -The configuration file above assumes that the results are found in sub-folders -of `mh-index-folder' which is +mhe-index by default. + FileRules pathname contains /home/user/Mail/scripts -Use the following command line to generate the swish index. Run this -daily from cron: +Use the following command line to generate the swish index. Run this daily +from cron: - swish-e -c /home/user/Mail/.swish/config + swish-e -c /home/user/Mail/.swish/config -FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP is used to +search." (set-buffer (get-buffer-create mh-index-temp-buffer)) (erase-buffer) (unless mh-swish-binary @@ -1407,7 +1442,8 @@ FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." (list (let* ((s (buffer-substring-no-properties start (1+ (point))))) (unless (string-match mh-swish-folder s) (return 'error)) - (if (string-match mh-user-path s) + (if (and (string-match mh-user-path s) + (< (match-end 0) (1- (length s)))) (format "+%s" (substring s (match-end 0) (1- (length s)))) (return 'error))) @@ -1431,34 +1467,35 @@ FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." (defun mh-swish++-execute-search (folder-path search-regexp) "Execute swish++ and read the results. -In the examples below, replace /home/user/Mail with the path to your MH +In the examples below, replace \"/home/user/Mail\" with the path to your MH directory. -First create the directory /home/user/Mail/.swish++. Then create the file -/home/user/Mail/.swish++/swish++.conf with the following contents: +First create the directory \"/home/user/Mail/.swish++\". Then create the file +\"/home/user/Mail/.swish++/swish++.conf\" with the following contents: - IncludeMeta Bcc Cc Comments Content-Description From Keywords - IncludeMeta Newsgroups Resent-To Subject To - IncludeMeta Message-Id References In-Reply-To - IncludeFile Mail * - IndexFile /home/user/Mail/.swish++/swish++.index + IncludeMeta Bcc Cc Comments Content-Description From Keywords + IncludeMeta Newsgroups Resent-To Subject To + IncludeMeta Message-Id References In-Reply-To + IncludeFile Mail * + IndexFile /home/user/Mail/.swish++/swish++.index -Use the following command line to generate the swish index. Run this -daily from cron: +Use the following command line to generate the swish index. Run this daily +from cron: - find /home/user/Mail -path /home/user/Mail/mhe-index -prune \\ - -o -path /home/user/Mail/.swish++ -prune \\ - -o -name \"[0-9]*\" -print \\ - | index -c /home/user/Mail/.swish++/swish++.conf /home/user/Mail + find /home/user/Mail -path /home/user/Mail/mhe-index -prune \\ + -o -path /home/user/Mail/.swish++ -prune \\ + -o -name \"[0-9]*\" -print \\ + | index -c /home/user/Mail/.swish++/swish++.conf - -You do not want to index the folders that hold the results of your searches -since they tend to be ephemeral and the original messages are indexed anyway. -The command above assumes that the results are found in sub-folders of -`mh-index-folder' which is +mhe-index by default. +This command does not index the folders that hold the results of your searches +in \"+mhe-index\" since they tend to be ephemeral and the original messages +are indexed anyway. -On some systems (Debian GNU/Linux, for example), use index++ instead of index. +On some systems (Debian GNU/Linux, for example), use \"index++\" instead of +\"index\". -FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP is used to +search." (set-buffer (get-buffer-create mh-index-temp-buffer)) (erase-buffer) (unless mh-swish++-binary @@ -1512,32 +1549,29 @@ REGEXP-LIST is an alist of fields and values." (defun mh-namazu-execute-search (folder-path search-regexp) "Execute namazu and read the results. -In the examples below, replace /home/user/Mail with the path to your MH +In the examples below, replace \"/home/user/Mail\" with the path to your MH directory. -First create the directory /home/user/Mail/.namazu. Then create the file -/home/user/Mail/.namazu/mknmzrc with the following contents: +First create the directory \"/home/user/Mail/.namazu\". Then create the file +\"/home/user/Mail/.namazu/mknmzrc\" with the following contents: - package conf; # Don't remove this line! - $ADDRESS = 'user@localhost'; - $ALLOW_FILE = \"[0-9]*\"; - $EXCLUDE_PATH = \"^/home/user/Mail/(mhe-index|spam)\"; + package conf; # Don't remove this line! + $ADDRESS = 'user@localhost'; + $ALLOW_FILE = \"[0-9]*\"; + $EXCLUDE_PATH = \"^/home/user/Mail/(mhe-index|spam)\"; -In the above example configuration, none of the mail files contained in the -directories /home/user/Mail/mhe-index and /home/user/Mail/spam are indexed. +This configuration does not index the folders that hold the results of your +searches in \"+mhe-index\" since they tend to be ephemeral and the original +messages are indexed anyway. -You do not want to index the folders that hold the results of your searches -since they tend to be ephemeral and the original messages are indexed anyway. -The configuration file above assumes that the results are found in sub-folders -of `mh-index-folder' which is +mhe-index by default. - -Use the following command line to generate the namazu index. Run this -daily from cron: +Use the following command line to generate the namazu index. Run this daily +from cron: - mknmz -f /home/user/Mail/.namazu/mknmzrc -O /home/user/Mail/.namazu \\ - /home/user/Mail + mknmz -f /home/user/Mail/.namazu/mknmzrc -O /home/user/Mail/.namazu \\ + /home/user/Mail -FOLDER-PATH is the directory in which SEARCH-REGEXP is used to search." +In a program, FOLDER-PATH is the directory in which SEARCH-REGEXP is used to +search." (let ((namazu-index-directory (format "%s%s" mh-user-path mh-namazu-directory))) (unless (file-exists-p namazu-index-directory) |