Next: , Previous: Miscellaneous, Up: Top

Appendix A Scan Line Formats

This appendix discusses how MH-E creates, parses, and manipulates scan lines. If you have your own MH scan or inc format files, you can teach MH-E how to handle them, but it isn't easy as you'll see.

This table lists the options in the ‘mh-scan-line-formats’ customization group.

mh-adaptive-cmd-note-flag
On means that the message number width is determined dynamically (default: ‘on’).
mh-scan-format-file
Specifies the format file to pass to the scan program (default: ‘Use MH-E scan Format’).
mh-scan-prog
Program used to scan messages (default: "scan").

There are a couple of caveats when creating your own scan format file. First, MH-E will not work if your scan lines do not include message numbers. It will work poorly if you don't dedicate a column for showing the current message and notations. It is also best to keep the first column empty to make room for the cursor and so that text isn't obscured by the current message's overlay arrow when running in a terminal. You won't be able to use the option mh-adaptive-cmd-note-flag or the threading features (see Threading).

If you've created your own format to handle long message numbers, you'll be pleased to know you no longer need it since MH-E adapts its internal format based upon the largest message number if mh-adaptive-cmd-note-flag is on (the default). If you prefer fixed-width message numbers, turn off mh-adaptive-cmd-note-flag and call mh-set-cmd-note with the width specified by your format file (see mh-scan-format-file). For example, the default width is 4, so you would use ‘(mh-set-cmd-note 4)’.

The default setting for mh-scan-format-file is ‘Use MH-E scan Format’. This means that the format string will be taken from the either mh-scan-format-mh or mh-scan-format-nmh depending on whether MH or nmh (or GNU mailutils MH) is in use. This setting also enables you to turn on the option mh-adaptive-cmd-note-flag. You can also set this option to ‘Use Default scan Format’ to get the same output as you would get if you ran scan from the shell. If you have a format file that you want MH-E to use but not MH, you can set this option to ‘Specify a scan Format File’ and enter the name of your format file.

The scan format that MH-E uses when mh-scan-format-file is set to its default of ‘Use MH-E scan Format’ is held in the variables mh-scan-format-nmh and mh-scan-format-mh depending on whether you are using nmh (or GNU mailutils MH) or not. Typically, you create your own format files rather than modifying these variables. The value of mh-scan-format-nmh is:

     (concat
      "%4(msg)"
      "%<(cur)+%| %>"
      "%<{replied}-"
      "%?(nonnull(comp{to}))%<(mymbox{to})t%>"
      "%?(nonnull(comp{cc}))%<(mymbox{cc})c%>"
      "%?(nonnull(comp{bcc}))%<(mymbox{bcc})b%>"
      "%?(nonnull(comp{newsgroups}))n%>"
      "%<(zero) %>"
      "%02(mon{date})/%02(mday{date})%<{date} %|*%>"
      "%<(mymbox{from})%<{to}To:%14(decode(friendly{to}))%>%>"
      "%<(zero)%17(decode(friendly{from}))%>  "
      "%(decode{subject})%<{body}<<%{body}%>")

The setting for mh-scan-format-mh is similar, except that MH doesn't have the function decode (which is used to decode RFC 2047 encodings).

These strings are passed to the scan program via the -format argument. The formats are identical to the defaults except that additional hints for fontification have been added to the existing notations in the fifth column (remember that in Emacs, the columns start at 0). The values of the fifth column, in priority order, are: ‘-’ if the message has been replied to, ‘t’ if an address in the ‘To:’ field matches one of the mailboxes of the current user, ‘c’ if the ‘Cc:’ field matches, ‘b’ if the ‘Bcc:’ field matches, and ‘n’ if a non-empty ‘Newsgroups:’ field is present.

The name of the program that generates a listing of one line per message is held in mh-scan-prog (default: "scan"). Unless this variable contains an absolute pathname, it is assumed to be in the mh-progs directory (see Getting Started). You may link another program to scan (see ‘mh-profile’(5)) to produce a different type of listing1.

If you change the format of the scan lines you'll need to tell MH-E how to parse the new format. As you will see, quite a lot of variables are involved to do that. Use M-x apropos <RET> mh-scan.*regexp <RET> to obtain a list of these variables. You will also have to call mh-set-cmd-note if your notations are not in column 4 (columns in Emacs start with 0). Note that unlike most of the user options described in this manual, these are variables and must be set with setq instead of in a customization buffer. For help with regular expressions, see section Syntax of Regular Expressions in The GNU Emacs Manual.

The first variable has to do with pruning out garbage.

mh-scan-valid-regexp
This regular expression describes a valid scan line. This is used to eliminate error messages that are occasionally produced by inc2 or scan (default: "^ *[0-9]").

Next, many variables control how the scan lines are parsed.

mh-scan-body-regexp
This regular expression matches the message body fragment. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least one parenthesized expression which matches the body text as in the default of "\\(<<\\([^\n]+\\)?\\)". If this regular expression is not correct, the body fragment will not be highlighted with the face mh-folder-body.
mh-scan-cur-msg-number-regexp
This regular expression matches the current message. It must match from the beginning of the line. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least one parenthesized expression which matches the message number as in the default of "^\\( *[0-9]+\\+\\).*". This expression includes the leading space and current message marker ‘+’ within the parenthesis since it looks better to highlight these items as well. The highlighting is done with the face mh-folder-cur-msg-number. This regular expression should be correct as it is needed by non-fontification functions. See also mh-note-cur.
mh-scan-date-regexp
This regular expression matches a valid date. It must not be anchored to the beginning or the end of the line. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain only one parenthesized expression which matches the date field as in the default of "\\([0-9][0-9]/[0-9][0-9]\\)". If this regular expression is not correct, the date will not be highlighted with the face mh-folder-date.
mh-scan-deleted-msg-regexp
This regular expression matches deleted messages. It must match from the beginning of the line. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least one parenthesized expression which matches the message number as in the default of "^\\( *[0-9]+\\)D". This expression includes the leading space within the parenthesis since it looks better to highlight it as well. The highlighting is done with the face mh-folder-deleted. This regular expression should be correct as it is needed by non-fontification functions. See also mh-note-deleted.
mh-scan-good-msg-regexp
This regular expression matches “good” messages. It must match from the beginning of the line. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least one parenthesized expression which matches the message number as in the default of "^\\( *[0-9]+\\)[^D^0-9]". This expression includes the leading space within the parenthesis since it looks better to highlight it as well. The highlighting is done with the face mh-folder-msg-number. This regular expression should be correct as it is needed by non-fontification functions.
mh-scan-msg-format-regexp
This regular expression finds the message number width in a scan format. Note that the message number must be placed in a parenthesized expression as in the default of "%\\([0-9]*\\)(msg)". This variable is only consulted if mh-scan-format-file is set to ‘Use MH-E scan Format’.
mh-scan-msg-format-string
This is a format string for the width of the message number in a scan format. Use ‘0%d’ for zero-filled message numbers. This variable is only consulted if mh-scan-format-file is set to ‘Use MH-E scan Format’ (default: "%d").
mh-scan-msg-number-regexp
This regular expression extracts the message number. It must match from the beginning of the line. Note that the message number must be placed in a parenthesized expression as in the default of "^ *\\([0-9]+\\)".
mh-scan-msg-overflow-regexp
This regular expression matches overflowed message numbers (default: "^[?0-9][0-9]").
mh-scan-msg-search-regexp
This regular expression matches a particular message. It is a format string; use ‘%d’ to represent the location of the message number within the expression as in the default of "^[^0-9]*%d[^0-9]".
mh-scan-rcpt-regexp
This regular expression specifies the recipient in messages you sent. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain two parenthesized expressions. The first is expected to match the ‘To:’ that the default scan format file generates. The second is expected to match the recipient's name as in the default of "\\(To:\\)\\(..............\\)". If this regular expression is not correct, the ‘To:’ string will not be highlighted with the face mh-folder-to and the recipient will not be highlighted with the face mh-folder-address.
mh-scan-refiled-msg-regexp
This regular expression matches refiled messages. It must match from the beginning of the line. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least one parenthesized expression which matches the message number as in the default of "^\\( *[0-9]+\\)\\^". This expression includes the leading space within the parenthesis since it looks better to highlight it as well. The highlighting is done with the face mh-folder-refiled. This regular expression should be correct as it is needed by non-fontification functions. See also mh-note-refiled.
mh-scan-sent-to-me-sender-regexp
This regular expression matches messages sent to us. Note that the default setting of mh-folder-font-lock-keywords expects this expression to contain at least two parenthesized expressions. The first should match the fontification hint (see mh-scan-format-nmh) and the second should match the user name as in the default of "^ *[0-9]+.\\([bct]\\).....[ ]*\\(..................\\)". If this regular expression is not correct, the notation hints will not be highlighted with the face mh-mh-folder-sent-to-me-hint and the sender will not be highlighted with the face mh-folder-sent-to-me-sender.
mh-scan-subject-regexp
This regular expression matches the subject. It must match from the beginning of the line. Note that the default setting of ‘mh-folder-font-lock-keywords’ expects this expression to contain at least three parenthesized expressions. The first is expected to match the ‘Re:’ string, if any, and is highlighted with the face mh-folder-followup. The second matches an optional bracketed number after ‘Re:’, such as in ‘Re[2]:’ (and is thus a sub-expression of the first expression). The third is expected to match the subject line itself which is highlighted with the face mh-folder-subject. For example, the default is "^ *[0-9]+........[ ]*...................
\\([Rr][Ee]\\(\\[[0-9]+\\]\\)?:\\s-*\\)*\\([^<\n]*\\)". This regular expression should be correct as it is needed by non-fontification functions. Note that this example is broken up on two lines for readability, but is actually a single string.

Finally, there are a slew of variables that control how MH-E annotates the scan lines.

mh-cmd-note
Column for notations (default: 4). This variable should be set with the function mh-set-cmd-note. This variable may be updated dynamically if mh-adaptive-cmd-note-flag is on. The following variables contain the notational characters. Note that columns in Emacs start with 0.
mh-note-copied
Messages that have been copied are marked by this character (default: ?C).
mh-note-cur
The current message (in MH, not in MH-E) is marked by this character (default: ?+). See also mh-scan-cur-msg-number-regexp.
mh-note-deleted
Messages that have been deleted are marked by this character (default: ?D). See also mh-scan-deleted-msg-regexp.
mh-note-dist
Messages that have been redistributed are marked by this character (default: ?R).
mh-note-forw
Messages that have been forwarded are marked by this character (default: ?F).
mh-note-printed
Messages that have been printed are marked by this character (default: ?P).
mh-note-refiled
Messages that have been refiled are marked by this character (default: ?^). See also mh-scan-refiled-msg-regexp.
mh-note-repl
Messages that have been replied to are marked by this character (default: ?-).
mh-note-seq
Messages in a user-defined sequence are marked by this character (default: ?%). Messages in the ‘search’ sequence are marked by this character as well.

For example, let's say I have the following in scan.format which displays the sender, the subject, and the message number. This format places a ‘+’ after the message number for the current message according to MH; it also uses that column for notations.

     %20(decode(friendly{from})) %50(decode{subject})  %4(msg)%<(cur)+%| %>

The first thing you have to do is tell MH-E to use this file. Customize mh-scan-format-file and set its value to ‘Use Default scan Format’. If you didn't get already turn off mh-adaptive-cmd-note-flag, you'll need to do that first.

Next, tell MH-E what a valid scan line looks like so that you can at least display the output of scan in your MH-Folder buffer.

     (setq mh-scan-valid-regexp "[0-9]+[+D^ ]$")

Now, in order to get rid of the ‘Cursor not pointing to message’ message, you need to tell MH-E how to access the message number. You should also see why MH-E requires that you include a message number in the first place.

     (setq mh-scan-msg-number-regexp "^.* \\([0-9]+\\)[+D^ ]$")
     (setq mh-scan-msg-search-regexp " %d[+D^ ]$")

In order to get the next and previous commands working, add this.

     (setq mh-scan-good-msg-regexp "^.* \\([0-9]+\\)[+D^ ]$")

Note that the current message isn't marked with a ‘+’ when moving between the next and previous messages. Here is the code required to get this working.

     (set-mh-cmd-note 76)
     (setq mh-scan-cur-msg-number-regexp "^.* \\([0-9]+\\)\\+$")

Finally, add the following to delete and refile messages.

     (setq mh-scan-deleted-msg-regexp "^.* \\([0-9]+\\)D$")
     (setq mh-scan-refiled-msg-regexp "^.* \\([0-9]+\\)\\^$")

This is just a bare minimum; it's best to adjust all of the regular expressions to ensure that MH-E and highlighting perform well.


Footnotes

[1] See the section Find and Specify with scan pick Ranges Sequences in the MH book.

[2] See the section Reading Mail: inc show next prev in the MH book.