1. BBDB

BBDB is a rolodex-like database program for GNU Emacs. BBDB stands for Insidious Big Brother Database, and is not, repeat, not an obscure reference to the Buck Rogers TV series.

It provides the following features:

• Integration with mail and news readers, with little or no interaction by the user:
  • easy (or automatic) display of the record corresponding to the sender of the current message;

  • automatic creation of records based on the contents of the current message;

  • automatic addition of data to arbitrary fields of the record corresponding to the sender of the current message.

• Listing all records which match a regular expression;

• Listing all records which match a regular expression in a particular field (`company' or `notes,' for example);

1.1 Installation

This program consists of several groups of files, organized by directory:

     lisp     - the main program code for the BBDB
     tex      - TeX support files for See section 1.8.2 bbdb-print, the BBDB
                printing utility
     texinfo  - the documentation files for the BBDB
     utils    - miscellaneous external utility programs
     misc     - things that don't fall into the above categories
     bits     - things that have been written as add-ons for BBDB
                but have not yet been merged with the main codebase

1.1.1 General Prerequisites

Various parts of the BBDB require extra packages to be available that are not part of the BBDB distribution. Please note that with one exception no extra packages (beyond those which ship with both GNU Emacs and XEmacs) are required for the use of BBDB core functionality.(1) This one exception applies to XEmacs 20.5 users - the xemacs-base package must be installed for the correct operation of the core BBDB functionality. The table below lists the requirements of the various portions of the BBDB. Please note that the absence of any of the below optional packages will not affect core BBDB functionality.

NOTES:

1. The old GNUS mail/newsreader should still work. Please keep in mind that you have a relatively recent Emacs (GNU 19.34 or later, XEmacs 19.15 or later), you are probably using Gnus.
2. As of this writing, Reportmail is available as part of the edit-utils package.
3. As of this writing, browse-url is available as part of the mail-lib package.
4. The source release of VM is currently required due to the use of macros from the VM codebase in BBDB's VM integration.

Please also note that the XEmacs package locations are as of this writing. As the XEmacs 20.5 package system is still in development, the locations may change without warning.

1.1.2 Normal User Installation

Configuring the compilation process

First of all, you should run the configure script at the toplevel of the distribution. This script will perform a number of checks on your system and generate the `Makefile''s accordingly.

The configure script also comes with a number of options that lets you customize the compilation process. These options are described below where appropriate.

Byte Compiling the Lisp files

First, you need to byte-compile the appropriate BBDB Lisp files. While this is in theory an optional step, it is virtually required in practice due to speed reasons.

In order to byte-compile the lisp files, an Emacs of some sort must be used. By default (at configure time), emacs and xemacs will be tried in that order. If you want to use a special Emacs flavor (or if you want to use xemacs at the first place), you should pass the --with-emacs=PROG option to configure.

In order to successfully compile the BBDB, the build process also needs to know the location of the various optional packages. If the directories containing these optional packages are in the default Emacs search path (the load-path variable), no other changes need be made for the build process to complete successfully.

If the optional packages are not in the default search path, the build process will not find them unless explicitly told of their location(s). To tell the build process where to find Gnus, MH-E, and/or VM, use the configure options --with-gnus-dir=DIR, --with-mhe-dir=DIR, and/or --with-vm-dir=DIR variables respectively. To tell the build process where to find any other package(s), pass the directories containing the lisp files for the package(s) to the configure option --with-other-dirs=DIRS. If multiple directories are to be added, they should be separated by spaces or colons, and should not be quoted. For example, to add the `/p/local/elisp/footnote' and `/p/local/elisp/sc' directories, call the configure script as follows:

  configure --with-other-dirs=/p/local/elisp/footnote:/p/local/elisp/sc

After configuring, run one of the following commands:

     make bbdb  - Build the core, mailer independent, components
     make gnus  - Core components plus Gnus support
     make mhe   - Core components plus MH-E support
     make rmail - Core components plus RMAIL support
     make vm    - Build the core components with VM support
     make all   - Core components plus support for all mailers
                         listed above

You can also combine the above make commands. For example, to build the BBDB with support for Gnus and VM, you can do so by typing:

make gnus vm

Moving the files to their final destination

Lisp files

As stated above, the `lisp' subdirectory contains the Emacs Lisp source files for the BBDB. Therefore, these files must be in the Emacs load-path. There are several ways of doing this, three of which are described below:

• Add the `lisp' directory from the source distribution to the load-path. This will allow you to run the BBDB in-place. This method is recommended for normal users or BBDB developers, especially if disk usage is an issue. It is not recommended for site-wide installations.

• Link the `lisp' directory into your `site-lisp' directory. This is for a site-wide installation, but it is subject to the following caveat. If you link the `lisp' directory into `site-lisp', you will make life more difficult for yourself down the road, as you will not be able to make changes to the source directory (new versions, patches, etc) without having an effect on other users who now depend on it. This directory will automatically be added to the load-path when Emacs starts.

• Make a directory whose sole purpose in life is containing the production copies of the BBDB source and byte-compiled source files. Either put this directory under `site-lisp' (or put it somewhere else and link it into `site-lisp'). This directory will automatically be added to the load-path when Emacs starts. This is the best of the three listed here, as it allows for a degree of separation between the (possibly changing) source tree and the production code.

TeX files

The `tex' subdirectory contains the TeX support files for bbdb-print, the BBDB printing utility (See section 1.8.2 bbdb-print.). The three support files, `bbdb-cols.tex', `bbdb-print.tex', and `bbdb-print-brief.tex', must be placed in a directory that is either on the default TeX search path or is listed in the TEXINPUTS environment variable. If neither of these two options is taken, TeX will not be able to process the file output by bbdb-print.

texinfo files

The `bbdb.info' file in this directory contains the documentation for the BBDB. This file should either be linked or copied to a directory on the default path for the info program or listed in the INFOPATH environment variable.

1.1.3 XEmacs Package Installation

NOTE: XEmacs packages are currently supported only under XEmacs versions after and including 20.5. If you are not running such a version of XEmacs, you should install the BBDB according to the instructions in 1.1.2 Normal User Installation.

Configuring / Byte Compiling

The configuration and byte-compilation procedures are the same as in the Normal User installation. See 1.1.2 Normal User Installation.

Moving the files to their final destination

Support is provided for the automatic installation of the BBDB in an XEmacs package directory. The following configure options are available for you:

--with-package-dir=DIR

This option sets the root of the XEmacs package directory. By default, `/usr/local/lib/xemacs/site-packages' is used.

--with-symlinks

If this option is used, the installation will be done by making symbolic links to the sources instead of copying the files.

--with-linkpath=PATH

Without this option, the installation process uses the output of pwd to determine the current directory. If something else should be used, you should provide an alternate name for the BBDB toplevel directory by using --with-linkpath. If, for example, pwd returns `/p/local/elisp/bbdb', but you prefer to use `/usr/local/elisp/bbdb/...' for the links, usr this: configure --with-linkpath=/usr/local/elisp/bbdb. This option is ignored if --with-symlinks is not used.

To perform the (un)installation, use the command make (un)install-pkg. This will compile the `lisp/auto-autoloads.el' file and will install the appropriate files to the appropriate places. The final installation tree will take the following form:

$(PACKAGEDIR)/

lisp/

bbdb/

BBDB lisp source files. This directory contains a copy of all .el and .elc files from the `lisp' source directory, or is a symbolic link to it.

info/

bbdb.info*

BBDB documentation files. These are either copies of the info files from the `texinfo' source directory, or are symbolic links to them.

etc/

bbdb/

tex/

BBDB support files for bbdb-print. This directory contains a copy of the appropriate files from the `tex' source directory, or is a symbolic link to it.

utils/

BBDB miscellaneous utilities. This directory contains a copy of the appropriate files from the `utils' source directory, or is a symbolic link to it.

1.1.4 Initial Configuration

The simplest way to configure the BBDB is to include the following forms in your Emacs configuration file:

(require 'bbdb)
(bbdb-initialize)

Note: The forms above replace the autoloads needed for previous versions of the BBDB.

This will set up the BBDB for basic querying and record manipulation (the Core Functionality referred to in the Prerequisites section). It will not enable any of the mailreader-, newsreader- or other package-specific BBDB features. To enable some or all of these features, the (bbdb-initialize) form can be enable as shown below. Alternatively, the features can be enabled manually as described in the following sections.

Modifying (bbdb-initialize)

The bbdb-initialize function can be used to enable the various package-specific BBDB functions. This feature activation is accomplished through the passing of symbols which tell the function which features to activate. These symbols are outlined below and in the Emacs documentation for the bbdb-initialize(2)

Initialization symbols for mail and news readers

gnus

Initialize support for Gnus(3). If you pass the gnus symbol, you should probably also pass the message symbol.

mh-e

Initialize support for the MH-E mail reader.

rmail

Initialize support for the RMAIL mail reader.

sendmail

Initialize support for sendmail (M-x mail)

vm

Initialize support for the VM mail reader.(4)

Initialization symbols for other packages

message

Initialize support for Message mode (the mail composition program included with Gnus).

reportmail

Initialize support for the Reportmail mail notification package.

sc

Initialize support for the Supercite message citation package. Additional initialization is required for Supercite to work with the BBDB. See section 1.1.4.8 Initializing BBDB support for Supercite.

w3

Initialize support for Web browsers.

Initialization example

To initialize support for Gnus 5.5, Message mode, Supercite, and Web browsers, the following forms would be used:

(require 'bbdb)
(bbdb-initialize 'gnus 'message 'sc 'w3)

Manual initialization

If your initialization needs exceed those provided by bbdb-initialize, refer to the following sections for a description of the procedures necessary for enabling BBDB support for the packages listed above. The procedures described are the same as those carried out by the bbdb-initialize function when passed the appropriate symbols. That is, the procedure listed in the RMAIL Prep section below is the same as than executed by bbdb-initialize when the rmail symbol is passed.

1.1.4.1 Initializing BBDB support for Gnus

To take advantage of the Gnus features of the BBDB, add one of the following forms to your Emacs configuration file:

For Gnus 3.14 or older:

(add-hook 'gnus-Startup-hook 'bbdb-insinuate-gnus)

For Gnus 3.15 or newer:

(add-hook 'gnus-startup-hook 'bbdb-insinuate-gnus)

bbdb-insinuate-gnus adds bindings for the default keys to Gnus and configures Gnus to notify the BBDB when new messages are loaded. This notification is required if the BBDB is to be able to display BBDB entries for messages displayed in Gnus.

1.1.4.2 Initializing BBDB support for MH-E

To take advantage of the MH-E features of the BBDB, add the following form to your Emacs configuration file:

(add-hook 'mh-folder-mode-hook 'bbdb-insinuate-mh)

bbdb-insinuate-mh adds bindings for the default keys to MH-E and configures MH-E to notify the BBDB when new messages are loaded. This notification is required if the BBDB is to be able to display BBDB entries for messages displayed in MH-E.

1.1.4.3 Initializing BBDB support for RMAIL

To take advantage of the RMAIL features of the BBDB, add the following form to your Emacs configuration file:

(add-hook 'rmail-mode-hook 'bbdb-insinuate-rmail)

bbdb-insinuate-rmail adds bindings for the default keys to RMAIL and configures RMAIL to notify the BBDB when new messages are loaded. This notification is required if the BBDB is to be able to display BBDB entries for messages displayed in RMAIL.

1.1.4.4 Initializing BBDB support for Sendmail

To take advantage of send-mail-mode (the one invoked with M-x mail) features of the BBDB, add the following form to your Emacs configuration file:

(add-hook 'mail-setup-hook 'bbdb-insinuate-sendmail)

bbdb-insinuate-sendmail enables auto-completion in send-mail-mode.

1.1.4.5 Initializing BBDB support for VM

To take advantage of the VM features of the BBDB, either add 'vm to the parameters of the (bbdb-initialize) form, or add the following form to your `~/.vm' file:

(bbdb-insinuate-vm)

bbdb-insinuate-vm adds bindings for the default keys to VM and configures VM to notify the BBDB when new messages are loaded. This notification is required if the BBDB is to be able to display BBDB entries for messages displayed in VM.

1.1.4.6 Initializing BBDB support for Message mode

To allow the BBDB to be used in Message mode, add the following form to your Emacs initialization file:

(bbdb-insinuate-message)

bbdb-insinuate-message adds a binding for M-TAB to Message mode. This will enable completion of addressees based on BBDB records. See 1.6.1 Using the BBDB with Message Mode for more details on the operation of Message mode BBDB record completion.

1.1.4.7 Initializing BBDB support for Reportmail

To allow the Reportmail package to report information from BBDB records for new mail, add the following form to your Emacs initialization file:

(bbdb-insinuate-reportmail)

bbdb-insinuate-reportmail adds to the display-time-get-field function to allow access to BBDB records during new mail information display. See 1.6.2 Using the BBDB with Reportmail for more details on the operation of Reportmail with the BBDB.

1.1.4.8 Initializing BBDB support for Supercite

To allow the BBDB to assist in the storage of Supercite citations, add the following form to your Emacs initialization file:

(bbdb-insinuate-sc)

bbdb-insinuate-sc adds BBDB functions to two Supercite hooks - sc-post-hook and sc-attribs-postselect-hook. See 1.6.3 Using the BBDB with Supercite for more details on the operation of Supercite citation management using the BBDB.

Three other Supercite variables must be set/modified to allow the BBDB to work with it. These variables are not automatically set as it would be impossible to reliably set them without interfering with other user customizations. The modifications are:

sc-preferred-attribution-list

"sc-consult" should be added to the list. An example configuration is:

(setq sc-preferred-attribution-list '("sc-lastchoice" "x-attribution" "sc-consult" "initials" "firstname" "lastname"))

sc-attrib-selection-list

The following form should be added to sc-attrib-selection-list:

'(("sc-from-address" ((".*" . (bbdb/sc-consult-attr (sc-mail-field "sc-from-address"))))))

sc-mail-glom-frame

The following form should be added to sc-mail-glom-frame, to allow the retrieval of the name of a person who is a) in the BBDB and b) has only included their net address in the message in question.

("^$" (progn (bbdb/sc-default) (list 'abort '(step . 0))))

An example configuration is as follows:

(setq sc-mail-glom-frame '((begin (setq sc-mail-headers-start (point))) ("^x-attribution:[ \t]+.*$" (sc-mail-fetch-field t) nil t) ("^\\S +:.*$" (sc-mail-fetch-field) nil t) ("^$" (progn (bbdb/sc-default) (list 'abort '(step . 0)))) ("^[ \t]+" (sc-mail-append-field)) (sc-mail-warn-if-non-rfc822-p (sc-mail-error-in-mail-field)) (end (setq sc-mail-headers-end (point)))))

The above is also documented in `bbdb-sc.el'. The bbdb/sc-setup-variables function has been provided as an example for Supercite variable initialization. Please note that while bbdb/sc-setup-variables makes every attempt to safely configure the Supercite variables, it will not always work. Specifically, the variables sc-attrib-selection-list and sc-mail-glom-frame will not be overridden if they have already been defined.

1.1.4.9 Initializing BBDB support for Web Browsers

To allow URLs to be added to BBDB records, add the following form to your Emacs initialization file:

(bbdb-insinuate-w3)

bbdb-insinuate-w3 adds the definition of : to the W3 keymap.

The other bbdb-w3 functions, specifically the passing of URLs from BBDB records to Web browsers, do not require initialization within the BBDB. They do, however, require the configuration of the browse-url package so it knows to which Web browser URLs are to be passed. For more details on the operation of bbdb-w3, see 1.6.4 Using the BBDB with Web Browsers.

1.2 The BBDB

This section discusses the basics of the BBDB - an overview of the database's layout, and a discussion of the basic BBDB manipulation commands.

The database itself lives in a file which is named by the variable bbdb-file. If this variable is not set, the database is assumed to be in `~/.bbdb'.

1.2.1 Database Fields

The database is organized as a set of records, where each record corresponds to one person or organization. Each record has several fields, and each field is of one of several types. Below, the built-in types are listed, followed by a description of how and why some types can be used more than once in a single record:

The field types listed above can be classified into four categories, as indicated by the comments in the `Notes' column.

Field types marked as "Single value, single occurrence" may only occur once per record. Each occurrence can only have a single value. For example, there will be only one field of type name in a record. It will be named name, and will contain a single value (the person's name).

The types marked as "Multiple values through commas" are essentially the same as the single value, single occurrence field types, but with one crucial difference: they can contain multiple values in the form of a comma-separated list. So, for example, while a name field with a value of "foo, bar" would be treated as if it contained the single value "foo, bar", a net field with the same data would be thought of as having two separate values - "foo" and "bar". As in the single occurrence, single value field types, there will be only one occurrence of each "Multiple values through commas" field type, and the occurrence will have the same name as the type.

The "Multiple values through multiple occurrences" field type is the most flexible of the four listed here. There can be multiple occurrences of each type. This type necessarily does not have the name restriction imposed by the previous two types. For example, there can be multiple fields of the address type, none of which have to be named address. One could be named home, and the other could be named work.

Special properties of the notes field type: All user-defined fields that don't fit into the other predefined field types (name, company, AKA, net, address, and phone) will be created as a notes-type field. In addition, several of the user-defined field names are "special". That is, the BBDB treats the values of these "special" fields differently than it does other user-defined fields. The "special" fields are:

attribution

(Available only when the Supercite-specific BBDB functions have been enabled) Used for the storage of Supercite attributions. For initialization details, see 1.1.4.8 Initializing BBDB support for Supercite. For usage details, see 1.6.3 Using the BBDB with Supercite. The field used can be changed by changing the value of bbdb/sc-attribution-field.

aka

Used to store non-primary names associated with a given record.

face

(XEmacs only) Used for the storage of image data. This data is to be in the format output by compface, and commonly found in X-Face: headers in messages. If face support has been compiled into XEmacs, the image contained in the face header will be displayed when the record is viewed.

finger-host

Address used in place of the listed net address for fingering the entity indicated by the record. See section 1.3 BBDB Mode. The field used can be changed by changing the value of bbdb-finger-host-field.

gnus-score

Gnus scoring adjustment for this person. For initialization details, see 1.1.4.1 Initializing BBDB support for Gnus. For usage details, see 1.5.1 Gnus-specific Features. The field used can be changed by changing the value of bbdb/gnus-score-field.

mail-alias

Value used instead of name for completion. See section 1.4.3 Mail Sending Interfaces.

mail-name

(Available only when the Reportmail-specific BBDB functions have been enabled) Used for the storage of non-default names to be used in the reporting of new mail by Reportmail. For initialization details, see 1.1.4.7 Initializing BBDB support for Reportmail. For usage details, see 1.6.2 Using the BBDB with Reportmail.

mark-char

The field containing the character to be used for marking a given poster in the Gnus Summary Buffer. For usage details, see 1.5.1.2 Gnus Summary Buffer Enhancements.

tex-name

The value of this field is used in place of the name field when printing the database using bbdb-print. See section 1.8.2 bbdb-print.

www

This field contains the URL associated with the BBDB record. Common uses are with bbdb-snarf (see section 1.8.3 bbdb-snarf) and the BBDB/Web Browser functionality (for initialization details, see 1.1.4.9 Initializing BBDB support for Web Browsers. For usage details, see 1.6.4 Using the BBDB with Web Browsers).

1.2.2 Basic searching commands

You can list the contents of the database with the command M-x bbdb. You will be prompted for a regular expression, and all records which match that regexp in the name, company, network address, or any notes fields will be displayed.

A narrower search may be made by using the commands bbdb-name, bbdb-company, bbdb-net, or bbdb-notes. These commands limit their searches to the name, company, email address, and notes fields, respectively. If these commands are given a prefix argument, the listing displayed will be one line per entry; otherwise, the full database entry will be shown on multiple lines.

The functions described above are predefined to certain keys in the *BBDB* buffer. See section 1.3 BBDB Mode. for more details.

The bbdb-notes command will prompt for the notes field to search (RET for all). In this way you can limit your searches to the contents of one particular user-defined notes field. (You can add user-defined fields with the bbdb-insert-new-field command; See section 1.3 BBDB Mode.)

1.2.3 Advanced searching commands

The following functions can be used to search for records based on creation and/or modification dates. These functions will match records that have timestamp and/or creation-date fields (as appropriate. See section 1.7.3 Predefined Hooks. for more information on these fields, which are created by default.

bbdb-timestamp-older

Display all records modified before a given date. If this function is called interactively, it will prompt for a date. If it is being called non-interactively, the date should be provided as a string in `yyyy-mm-dd' format.

bbdb-timestamp-newer

Display all records modified after a given date. If this function is called interactively, it will prompt for a date. If it is being called non-interactively, the date should be provided as a string in `yyyy-mm-dd' format.

bbdb-creation-older

Display all records created before a given date. If this function is called interactively, it will prompt for a date. If it is being called non-interactively, the date should be provided as a string in `yyyy-mm-dd' format.

bbdb-creation-newer

Display all records created after a given date. If this function is called interactively, it will prompt for a date. If it is being called non-interactively, the date should be provided as a string in `yyyy-mm-dd' format.

bbdb-creation-no-change

Display all records that have not been changed since creation.

1.2.4 Manual record addition

There are several ways to add new entries to the Insidious Big Brother Database; the most straightforward is to use M-x bbdb-create, which will prompt you for all relevant information. However, the easiest way is to allow them to be added automatically by one of the mail or news-reading interfaces (See section 1.4 Interfaces.).

There is als bbdb-snarf (See section 1.8.3 bbdb-snarf), which will attempt to create a record from a text block. Note that this depends on particular formatting and may not do exactly what you want.

1.3 BBDB Mode

1.3.1 Functions bound to keys in BBDB Mode

When the `*BBDB*' buffer is active (either summoned by one of the commands in the previous section [See section 1.2 The BBDB.] or by your mail or news program), a variety of commands become available for database manipulation. Some of the commands listed below take numeric arguments. These arguments can be generated by entering the number before pressing the key(s) corresponding to the desired command. The output (if any) of the listed commands will be displayed in the `*BBDB*' buffer, and can be navigated through using the usual cursor motion commands.

e

(bbdb-edit-current-field) Edit the field on the current line. If the cursor is in the middle of a multi-line field, such as an address or comments section, then the entire field is edited, not just the current line.

;

(bbdb-edit-notes) A shortcut for editing the notes field.

d, C-k

(bbdb-delete-current-field-or-record) Delete the field on the current line. If the current line is the first line of a record, the BBDB will, after prompting the user, delete the entire record from the database. This may also be applied to multiple records at once by *.

C-o

(bbdb-insert-new-field) Inserts a new field into the current record. You are prompted (with completion) for the type of field to insert (phone, address, notes, etc); if the string you type is not a known field type, you will be asked whether to add a new field with the entered name of type notes.

If you are inserting a new phone-number field, you can control whether it is a North American or European phone number by providing a prefix argument. A prefix arg of ^U means it's to be a euronumber, and any other prefix arg means it's to be a a structured North American number. If no prefix argument is supplied, the style used is controlled by the variable bbdb-north-american-phone-numbers-p.

C-x C-t

(bbdb-transpose-fields) This is like the transpose-lines command, but it is for BBDB fields. If the cursor is on a field of a BBDB record, that field and the previous field will be transposed.

With non-zero numeric argument ARG, the previous field is moved past ARG fields. With argument 0, the field indicated by point is interchanged with the one indicated by mark.

Both fields must be in the same record, and must be of the same basic type (that is, you can use this command to change the order in which phone-number fields are listed, but you can't use it to make an address appear before a phone number; the order of field types is fixed.)

n, p

(bbdb-next-record, bbdb-prev-record) Move to the next and previous displayed record, respectively.

t

(bbdb-toggle-records-display-layout) Toggles the display layout of a record. With a numeric argument of 0, the current record will be made displayed in one line layout; with any other argument, the current record will be shown in multi-line layout.

If *t is used instead of simply t, then the state of all records will be changed instead of just the one at point. In this case, a numeric argument of 0 means that all records will unconditionally be made one-line layout; any other numeric argument means that all of the records will unconditionally be shown expanded; and no numeric argument means that the records are made to be in the opposite state of the record under point.

T

(bbdb-display-record-completely) Show all the fields of the current record. The display layout `full-multi-line' is used for this.

o

(bbdb-omit-record) Removes the current record from the display, but does not delete it from the database; it merely makes it seem as if the most recent search had not matched this record. With a numeric argument, omit the next N records. With a negative argument, go backwards.

m

(bbdb-send-mail) Begin composing mail to the person represented by the current record. The first email address is used. Normally, the mail-sending package which is used is determined by which mail-reading package is loaded; that is, if MH-E is loaded, then mh-send will be used; if VM is loaded, then vm-mail is used; if message is loaded, then it is used; otherwise, mail is used. You can override this by setting the variable bbdb-send-mail-style to one of the symbols vm, mh, message, or mail.

If *m is used instead of simply m, then mail will be sent to all of the folks listed in the `*BBDB*' buffer instead of just the person under point.

This function does not at present use the facility provided by compose-mail and mail-user-agent. In a future version of the BBDB, it will.

s, C-x C-s

(bbdb-save-db) Saves the BBDB file to disk.

r

(bbdb-refile-record) Merge the current record into some other record; that is, delete the record under point after copying all of the data within it into some other record. this is useful if you realize that somehow a redundant record has gotten into the database, and you want to merge it with another.

If both records have names and/or companies, you are asked which to use. Phone numbers, addresses, and network addresses are simply concatenated. The first record is the record under the point; the second is prompted for. Completion behavior is as dictated by the variable bbdb-completion-type.

M-d

(bbdb-dial) This command will attempt to dial the phone number currently at point, or if point is at the start of a record, the first phone number in the record. An extension, if present, is disregarded.

The method of dialling is controlled by bbdb-modem-dial. If this variable is nil, the BBDB will play touchtones corresponding to the number to be dialled. Otherwise, this variable is treated as a modem command string to be prepended to the number prior to feeding it to bbdb-modem-device.

The BBDB plays touchtones using bbdb-sound-player to play the sounds and the elements of bbdb-sound-files as the audio to be played. The first ten elements of bbdb-sound-files correspond to the touchtones for the digits `0' to `9', while the eleventh and twelfth elements correspond to `#' and `*' respectively. The default configuration assumes a Solaris[tm] installation with the demonstration sound files in /usr/demo/SOUND/sounds.

The actual number dialled depends on the following variables:

bbdb-dial-local-prefix-alist

This is a list of (SEXPR REPLACEMENT) pairs. SEXPR is evaluated to produce a regular expression which is then applied to the number. If it matches, whatever it matches is replaced by REPLACEMENT. The match and replace is performed using each item in the list that matches, in sequence, so that the output from one item may become input to another. The default value for this variable is to remove (bbdb-default-area-code) (i.e. the value of that variable, in parenthesis) from the start of the number to be dialled.

Note: If this procedure produces a transformed number then no further modifications (such as prefix additions, below) will be made to the number before dialling.

Using a prefix argument to bbdb-dial disables the processing of this variable. The other modifiers, below, are not affected by this.

bbdb-dial-local-prefix

If the number to be dialled starts with a zero, it is deemed to be a local number, and bbdb-dial-local-prefix is prepended to it (see note above concerning bbdb-dial-local-prefix-alist processing, however).

bbdb-dial-long-distance-prefix

If the number to be dialled starts with a plus sign (+), it is deemed to be a long distance number, and bbdb-dial-long-distance-prefix is prepended to it (see note above concerning bbdb-dial-local-prefix-alist processing, however).

f

(bbdb-finger) This command fingers the network address of a BBDB record. If this command is executed from the `*BBDB*' buffer, it fingers the network address of the record which is at point; otherwise, it prompts in the minibuffer (with completion) for a user to finger. With a numeric prefix argument, it fingers the Nth network address of the current record; with a prefix argument of ^U, it fingers all of them. The `*finger*' buffer is filled asynchronously, meaning that you don't have to wait around for it to finish; but fingering another user before the first finger has finished could have unpredictable results.

If this command is executed from the `*BBDB*' buffer, it may be prefixed with * (as in *f instead of simply f), meaning to finger all of the users currently listed instead of just the one under point. The numeric prefix argument has the same interpretation.

You can define a special network address to "finger" by defining a field finger-host. The name of the field to be fingered can be changed by setting bbdb-finger-host-field.

q

(bbdb-bury-buffer) Hides the `*BBDB*' buffer. Note: This command does not kill the `*BBDB*' buffer.

?

(bbdb-help) This displays a one-line help message in the minibuffer, showing some of the most common bbdb-mode commands.

i

(bbdb-info) This documentation is displayed. Please note that either `bbdb' or `bbdb.info' must be installed in one of the info directories known to Emacs for this command to work.

bbdb-info-file

If this documentation is not installed in the standard Info directory, then you should set this variable to the name of the texinfo-formatted version of this file; the bbdb-info command will use this file instead.

W

(bbdb-www) Displays the Web page listed in the www field of the current record. See section 1.6.4 Using the BBDB with Web Browsers.

P

(bbdb-print) Creates a TeX file that contains a pretty-printed version of BBDB records. If prefixed by *, only the records currently displayed will print. See section 1.8.2 bbdb-print.

h

Moves point to another window via the other-window function.

c

(bbdb-create) Create a new database record from information supplied by the user.

C

(bbdb-changed) Display all records that have been changed since the last time the database was saved.

b

(bbdb) Begin a new database search. The results of the new search will be displayed in place of the results of the old search.

S a, S c, S o, S n

(bbdb-net, bbdb-company, bbdb-notes, bbdb-name)
Begin a new database search. This search will be limited to the net address, company, notes, or name fields, respectively, of database records. See section 1.2.2 Basic searching commands. for more details.

*

bbdb-append-records will make the next display/search command to append its results to the BBDB buffer instead of replacing its content.

With an prefix arg (C-u) toggle between always append and no append. With an prefix arg that is a positive number append will be enabled for that many times. With any other argument append will be enabled once."

1.3.2 Other database manipulation functions

bbdb-kill-older

If called interactively (or with a single argument - a date in `yyyy-mm-dd' format), it will kill all records that were last modified before the given date as determined by the timestamp field. See section 1.7.3 Predefined Hooks. If called non-interactively with a date (in `yyyy-mm-dd' format), a comparison function and an action function, the comparison function is applied to the timestamp field of all records, and the action function applied to those for whom the comparison function returns true. If `nil' is supplied as the comparison function, string-lessp is used.

1.4 Interfaces

The BBDB interfaces itself with several message-handling packages, but certain parameters control its behavior depending on whether it is being used from within a mail reader or a news reader.

In all of these packages, two new keybindings will be added:

:

(bbdb/package-show-sender) Displays the BBDB entry corresponding to the author of the current message. If there is none, you will be asked whether to create one. The function called is bbdb/package-show-sender, where package is either gnus, mh, rmail, or vm, depending on the mail or news program being used when the command is invoked.

;

(bbdb/package-annotate-sender) Lets you edit the `notes' field of the BBDB record corresponding to the sender of the current message. If there is no record for the current author, you will be asked whether to create one. The function called is bbdb/package-annotate-sender, where package is either gnus, mh, rmail, or vm, depending on the mail or news program being used when the command is invoked.

These keybindings (and several other features) will not be available unless you call the appropriate "insinuation" function; See section 1.1 Installation.

It is possible to configure BBDB so that it automatically creates a record when it sees a message from a person who is not in the database. It is also possible to have text automatically added to the notes field of the corresponding record depending on the contents of the message headers. See section 1.7.2 Customization Hooks.

1.4.1 Mail Reading Interfaces

There are BBDB interfaces for the following mail readers:

• Gnus, a news- and email- reader written by Lars Magne Ingebrigtsen (based on GNUS by Mansanobu Umeda).
• MH-E, the Emacs interface to Mail Handler (MH), from the standard emacs library, but packaged separately from XEmacs since version 20.4.
• RMAIL, from the standard emacs library (packaged separately for XEmacs users as of 20.4);
• View Mail, by Kyle Jones, version 5.31 or newer;

1.4.2 News Reading Interfaces

There are BBDB interfaces for the following news readers:

• GNUS, a newsreader written by Masanobu Umeda.
• Gnus, the modern news- and email-reading incarnation of GNUS. Gnus is written by Lars Magne Ingebrigtsen.

1.4.3 Mail Sending Interfaces

When sending mail, the keystroke M-TAB is bound to the function bbdb-complete-name. This will take the string that you have typed (from point back to the preceding colon, comma, or the beginning of the line) and will complete that against the contents of the database. What you have typed may be an initial subsequence of a person's full name or network address; if it completes ambiguously, then what you have typed will be replaced with the common portion of the matches. Typing M-TAB again will show a list of possible completions. If it completes unambiguously, then an address will be inserted. The variable bbdb-completion-type controls whether completion is done on real names, or network addresses, or both. The address inserted is normally of the form User Name <email-address>; however, if User Name has an address of the form <user.name@somedomain>, only the <email-address> portion is inserted. This can be overridden by setting bbdb-dwim-net-address-allow-redundancy to t.

This binding is automatically set by the various insinuation functions documented earlier in this manual. (See section 1.1.4 Initial Configuration.) Briefly, the forms for these functions are:

Gnus

(add-hook 'gnus-Startup-hook 'bbdb-insinuate-gnus) for Gnus 3.14 or older
(add-hook 'gnus-startup-hook 'bbdb-insinuate-gnus) for Gnus 3.15 or newer

MH-E

(add-hook 'mh-folder-mode-hook 'bbdb-insinuate-mh)

RMAIL

(add-hook 'rmail-mode-hook 'bbdb-insinuate-rmail)

sendmail

(add-hook 'mail-setup-hook 'bbdb-insinuate-sendmail)

VM

(bbdb-insinuate-vm) Add to `~/.vm' file

The above forms should be added to your Emacs initialization file, except where otherwise noted.

You can control what "real name" is inserted with the mail-alias field: if a record has a mail-alias field, then that is used instead of their name field.

If the variable bbdb-completion-display-record is true (the default) then when you successfully complete an address with M-TAB, the corresponding record will be appended to the `*BBDB*' buffer. The buffer will not be displayed if it is not already visible, but the record will be displayed there.

When sending mail, you can use the command bbdb-yank-addresses to CC the current message to the people currently displayed in the `*BBDB*' buffer. This is useful if you are in the midst of sending or replying to a message, and you decide to add some recipients. You can use one of the M-x bbdb commands to display the set of people that you want to CC the message to, and then execute this command to add them to the list.

Mailing Lists and Mail Aliases

If you are using Jamie Zawinski's `mail-abbrevs.el' package, which uses the word-abbrev mechanism for mail aliases, then you can store your mail aliases in the BBDB instead of duplicating the information elsewhere.

If you want a mail alias to be defined for a person, simply add a mail-alias field to their record. You may have multiple aliases for the same person; simply separate them with commas.

For convenience there is the function bbdb-add-or-remove-mail-alias bound to a which adds an alias to one or multiple records when prefixed by a *. Called with a prefix argument C-u it will remove the given alias.

If more than one person has the same mail-alias, then that alias expands to the addresses of all of those people; in this way you can maintain mailing lists within the BBDB.

When you want to group aliases as in .mailrc you may just retained the group aliases in your .mailrc.

To actually define the aliases which are stored in the BBDB, call the function bbdb-define-all-aliases from your mail-setup-hook (or message-setup-hook if you use Message mode coming with Gnus). This will search the database, and call define-mail-alias to define each of the resulting aliases.

1.5 Reader-specific Features

There are features of the BBDB that are available only for specific mail- and news-readers. These features are described below.

The headers which are parsed for email addresses and what records are displayed can be controlled by the following variables: bbdb-get-addresses-from-headers controls which headers are parsed for sender addresses when calling the show-sender function of your MUA. bbdb-get-addresses-to-headers controls which headers are parsed for recipients addresses when calling the show-all-recipients function of your MUA. When using the pop up feature it will search for the addresses in bbdb-get-addresses-headers and display them. By default it will list only the first address, but by setting bbdb-get-only-first-address-p to nil one will will get records for all addresses.

If there is no MUA specific variable for ignoring certain addresses then those addresses matching bbdb-user-mail-names will be ignored.

BBDB adds the bindings : for showing all senders and ; for editing the notes of the sender.

1.5.1 Gnus-specific Features

The BBDB can be used to provide score information, or to integrate database information into the Gnus Summary buffer or the GNUS Subject List.

1.5.1.1 Scoring

The BBDB can provide scoring information to Gnus in one of two ways.

1. Articles whose authors appear in the BBDB and who have gnus-score fields will have their scores adjusted by the value contained in that field.
2. Articles whose authors appear in the BBDB but who do not have gnus-score fields will have their scores adjusted by bbdb/gnus-score-default. If bbdb/gnus-score-default is nil, no score adjustment will be made.

The BBDB by default searches the field contained in bbdb/gnus-score-field for score values. To have the BBDB use a different field, change the value of this variable.

To enable BBDB-assisted scoring, add the bbdb/gnus-score function to gnus-score-find-score-files-function. Assuming that you want to preserve the default value of this variable, use a form similar to the following:

(setq gnus-score-find-score-files-function
      '(gnus-score-find-bnews bbdb/gnus-score))

Note: The default value in Gnus 5.5 is gnus-score-find-bnews. Check your configuration before using the above code, as your values may be different.

1.5.1.2 Gnus Summary Buffer Enhancements

Gnus can use the BBDB to do one of two things:

• Mark authors in the Summary Buffer who have records in the BBDB with a user-defined mark character. See Marking Posters, below.
• For authors in the Summary Buffer who also have records in the BBDB, replace their name as listed in the Summary Buffer with their name as stored in the BBDB. See Using Names from the BBDB, below.

Marking Posters

Authors with records in the BBDB can be marked either with a user-defined mark character, or with a default one. The marking is enabled by the use of a Gnus user format code, as determined by bbdb/gnus-summary-in-bbdb-format-letter. This variable, which defaults to `b', is used to create a format code which is intended for use in gnus-summary-line-format. The format code is created by concatenating `%u' with the value of bbdb/gnus-summary-in-bbdb-format-letter. In the default case this results in the creation of the format code `%ub'.

Posts are marked as follows: If the record for the poster has the field indicated in bbdb-message-marker-field (the default is mark-char), the value of that field is used as the mark character.(5) If no such field is present, the value of bbdb/gnus-summary-known-poster-mark will be used instead. If the author is not in the BBDB, a space will be used as the mark character.

Using Names from the BBDB

The names reported for authors of posts in the Summary buffer can be altered to conform to the values present in their respective BBDB records (if any). This rewriting is enabled by the use of a Gnus user format code, as determined by bbdb/gnus-summary-user-format-letter. This variable, which defaults to `B', is used to create a format code which is intended for use in gnus-summary-line-format. The format code is created by concatenating `%u' with the value of bbdb/gnus-summary-user-format-letter. In the default case this results in the creation of the format code `%uB'. This format code is intended to replace the format code previously used in the Summary buffer format line to indicate the author and/or net address (usually `%a', `%n', and/or `$N').

The effects of this format code are in two independent parts - the marking of known posters, and the rewriting of posters names. The first, the marking of posters, occurs only when bbdb/gnus-summary-mark-known-posters is t (the default) and the posters have entries in the BBDB. When this variable is true, the marking occurs as described in the previous section, Marking Posters, above.

The poster name rewriting is done for all posters - not just for those with records in the BBDB. That said, rewriting rules for posters in the BBDB are more flexible than for those not listed. The rewriting is governed by two variables, as described below.

bbdb/gnus-summary-prefer-real-names can have one of three values - `t', `bbdb', or nil. In general, this variable governs the preference between net addresses and names. If it is `t', the name (if any) will be used. If `nil', the net address will be used. The third value, `bbdb', can be used as a method for distinguishing between authors with records in the BBDB and those without. If the variable is set to `bbdb', the name from the BBDB record will be used if the author has a record in the BBDB. If the author is not in the BBDB, the net address from the message will be printed. This variable makes little sense if bbdb/gnus-summary-prefer-bbdb-data is `nil', as no names will be printed in the Summary buffer in this case - only net addresses.

bbdb/gnus-summary-prefer-bbdb-data is used to (dis)allow use of the BBDB for author data retrieval. If it is `t', data from the BBDB will be used if available. If it is `nil', data from the BBDB will not be used.

In the following examples, assume the following:

1. Message: From: Jamie <jwz@netscape.com>
   BBDB: No record
2. Message: From: Matt <simmonmt@acm.org>
   BBDB: Name: `Matthew', Net: `simmonmt@purdue.edu'

1.5.1.3 GNUS Summary Buffer Enhancements

This section is remarkably terse, as I don't have a copy of GNUS. If anybody can provide more descriptive information, please let me know.

(autoload 'bbdb/gnus-lines-and-from "bbdb-gnus")
(setq gnus-optional-headers 'bbdb/gnus-lines-and-from)

bbdb/gnus-mark-known-posters

If t (the default), then the GNUS subject list will contain an indication of those messages posted by people who have entries in the Insidious Big Brother Database (they will be marked with an asterisk.)

You can change the character used to mark records on a record-by-record basis by adding a mark-char property to the record, whose value is be the string to display (preferably one character.)

bbdb/gnus-header-prefer-real-names

Default: nil. if t, then the GNUS subject list will display real names instead of network addresses.

bbdb/gnus-header-show-bbdb-names

Default: t. If both this variable and the bbdb/gnus-header-prefer-real-names variable are true, then for news messages from people who are in your database, the name displayed will be the primary name from the database, rather than the one from the `From:' line of the message. This doesn't affect the names of people who aren't in the database, of course.

bbdb/gnus-lines-and-from-length

Default: 18. The number of characters used to display `From:' info in GNUS, if you have set gnus-optional-headers to bbdb/gnus-lines-and-from.

1.5.2 VM-specific features

The BBDB can be used to integrate database information into the message summary.

1.5.2.1 VM Message Summary Enhancements

VM users can cause their summary buffer to display the name of the message sender according to BBDB data, instead of according to the contents of the current message's headers. In VM 5.40 or later, use the summary format control %UB" instead of "%F", and the current record name will be shown there if available. If no entry is found it behaves like "%F". See the documentation for vm-summary-format for more details. Warning, this may significantly slow down summary generation for large folders.

1.5.2.2 VM configuration of what records the BBDB buffer shows

Email addresses which match vm-summary-uninteresting-senders are ignored.

The records in the BBDB buffer are listed in the same order as found.

1.5.2.3 VM automatic setup of vm-set-auto-folder-alist

VM users can setup the vm-set-auto-folder-alist automatically by calling bbdb/vm-set-auto-folder-alist. This adds for each BBDB record containing a bbdb/vm-set-auto-folder-alist-field an entry to vm-set-auto-folder-alist.

The record field can contain a string which is used as folder name or if it starts with a ' it is treated as lisp expression returning a folder name.

1.5.2.4 VM automatic adding of labels

bbdb/vm-auto-add-label-list is a List used by bbdb/vm-auto-add-label to automatically label messages. Each element in the list is either a string or a list of two strings. If a single string then it is used as both the field value to check for and the label to apply to the message. If a list of two strings, the first is the field value to search for and the second is the label to apply.

bbdb/vm-auto-add-label-field bbdb-define-all-aliases-field is the field used by bbdb/vm-auto-add-label to automatically label messages. Value is either a single symbol or a list of symbols of bbdb fields that bbdb/vm-auto-add-label uses to check for labels to apply to messages. Defaults to bbdb-define-all-aliases-field which is typically mail-alias.

bbdb/vm-auto-add-label automatically adds labels to messages based on the bbdb/vm-auto-add-label-field bbdb-define-all-aliases-field field. Add this to bbdb-notice-hook and if using VM each message that bbdb notices will be checked. If the sender has a value in the bbdb/vm-auto-add-label-field in their BBDB record that matches a value in bbdb/vm-auto-add-label-list then a VM label will be added to the message.

This works great when `bbdb-user-mail-names' is set. As a result mail that you send to people (and copy yourself on) is labeled as well.

1.6 Using the BBDB with other packages

The BBDB adds functionality to several packages. The following sections detail these augmentations.

1.6.1 Using the BBDB with Message Mode

At this time, the only feature the BBDB adds to Message mode is the binding to M-TAB which allows for BBDB record completion.

1.6.2 Using the BBDB with Reportmail

The BBDB can modify the `reportmail.el' package to use information from BBDB records when identifying the senders or recipients of e-mail messages.

In normal operation, Reportmail displays the name and net address sender and recipient of incoming messages. The BBDB can be configured to intercept and rewrite this information before it appears in the Emacs mode-line. It first attempts to rewrite the sender and/or recipient information by substituting those addresses with information from the BBDB. Replacement information is first sought from the mail-name field in the respective BBDB records. If no such field is found, the name field is returned. If no BBDB record is found, no rewriting is performed.

The BBDB-Reportmail augmentation is accomplished through the advising of the Reportmail display-time-get-field function in order to do a-posteriori modification of the returned value. The augmentation uses the bbdb/reportmail-alternate-full-name function to retrieve data from the BBDB for use in rewriting.

1.6.3 Using the BBDB with Supercite

The BBDB can be used with Supercite to store attributions with BBDB records. Normally, when a non-default attribution is entered for a given message, the entered attribution is used for that message, and is then discarded. When the BBDB-Supercite augmentation is enabled, the non-default attribution will be added to the record (if any) for the entity being cited. This poor explanation sounds complicated, but it's not. If a message from `Jamie Zawinski <jwz@netscape.com>' is being replied to, Supercite will, by default, suggest the citation `Jamie'. If the non-default citation `jwz' is entered, Supercite can save it with the BBDB record for `Jamie Zawinski' in the attribution field.

The field used can be changed by changing the value of bbdb/sc-attribution-field.

1.6.4 Using the BBDB with Web Browsers

The BBDB/Web Browser integration is in two parts, one which is automatically enabled, and one which must be manually enabled (see section 1.1.4.9 Initializing BBDB support for Web Browsers). The first feature added is the ability to display the URL associated with a given record in a Web Browser. The second is the ability to add URLs to BBDB records from within W3, the Emacs Web Browser.

Pressing W in the *BBDB* buffer while the cursor is positioned over a record with a www field will cause the first URL in the field to be loaded in a Web Browser. This functionality uses browse-url to display URLs - see the documentation for browse-url for information on selecting the browser to be used.

If W3 is used, and if the BBDB/W3 functionality has been enabled as described in 1.1.4.9 Initializing BBDB support for Web Browsers, pressing the : key will add the URL currently being displayed in W3 to a user-specified BBDB record.

1.7 Options

There are many variables which control the behavior of the Insidious Big Brother Database, and there are many hook-variables which can be used to modify its behavior in more complex ways. Several pieces of functionality are included which use the hooks in this way.

1.7.1 Customization Parameters

bbdb-file

The name of the file which contains your personal database. Default: `~/.bbdb'.

bbdb-default-area-code

The default area code to use when prompting for a new phone number. Default: 415. This must be a number, not a string.

bbdb-north-american-phone-numbers-p

Whether syntax-checking of phone numbers should be enforced. Default: t. This only works for Bell-system phone numbers. If this is true, then you can't enter invalid phone numbers, and all phone numbers are pretty-printed in the same way. European phone numbers don't have as strict a syntax, however, so this is a harder problem for them (on which I am punting).

You can have both styles of phone number in your database by providing a prefix argument to the bbdb-insert-new-field command.

bbdb-check-zip-codes-p

Whether syntax-checking of zip codes should be enforced. Default: t. If this is true, you can't enter invalid zip codes. A zip code is valid if it matches one of the regular expressions in the variable bbdb-legal-zip-codes.

bbdb-address-formatting-alist

Controls the display of addresses in the buffer. Each entry in this list consists of an identifying function and a formatting function. The identifying function must accept an address and return t if the associated formatting function is to be used. The formatting function must insert the formatted address in the current buffer. Identifying functions usually base their decision on the zip code format or on the country name. The default entries will format an address using continental style if the zip code matches bbdb-continental-zip-regexp. If the zip code does not match, addresses are formatted in US style.

bbdb-continental-zip-regexp

Decides whether an address should be formatted using US or European style. If the zip code of an address matches the regular expression, the European style is used. This works only if the expression (bbdb-address-is-continental . bbdb-format-address-continental) is part of bbdb-address-formatting-alist.

bbdb-electric-p

Whether bbdb mode should be "electric" like electric-buffer-list. Default: t. What this means is that the BBDB buffer which pops up when you use it can be disposed of by pressing the space bar, at which point your window configuration will be restored to what it was before you invoked the db list. (The bbdb-mode commands still work as well.)

There are some problems with electric modes; for example, keyboard macros and incremental search don't work. (This is not a bug in BBDB, but in `electric.el'.)

bbdb-case-fold-search

Default: the same as case-fold-search. case-fold-search is bound to this by M-x bbdb and related commands. This variable lets the case-sensitivity of ^S and of the bbdb searching commands be different.

bbdb/mail-auto-create-p

If this is t (the default), then VM, MH, and RMAIL will automatically create new bbdb records for people you receive mail from. If this variable is a function name or lambda expression, then it is called with no arguments to decide whether an entry should be automatically created. You can use this to, for example, not create records for messages which have reached you through a particular mailing list, or to only create records automatically if the mail has a particular subject. See the variables bbdb-ignore-most-messages-alist and bbdb-ignore-some-messages-alist (See section 1.7.3 Predefined Hooks.)

bbdb/news-auto-create-p

If this is t (default: nil), then GNUS will automatically create new BBDB records for people you read messages by. If this is a function name or lambda expression, then it is called with no arguments to decide whether an entry should be automatically created. You can use this to, for example, create or not create messages which have a particular subject. See the variable bbdb-auto-notes-alist (See section 1.7.3 Predefined Hooks.).

If you want to autocreate messages based on the current newsgroup, it's probably a better idea to set this variable to t or nil from your gnus-select-group-hook instead.

To automatically remember users in certain groups, you can do something like

(setq gnus-select-group-hook '(lambda () (setq bbdb/news-auto-create-p (or (string= "some.news.group" gnus-newsgroup-name) (string= "other.news.group" gnus-newsgroup-name)))))

bbdb-quiet-about-name-mismatches

If this is false (the default), then BBDB will prompt you when it notices a name change, that is, when the "real name" in a message doesn't correspond to a record already in the database with the same network address. As in, "John Smith <jqs@frob.com>" versus "John Q. Smith <jqs@frob.com>". If this is true, then you will not be asked if you want to change it (and it will not be changed.) If a number then it is the number of seconds to sit-for while displaying the name mismatch.

bbdb-use-alternate-names

If this is false, then the BBDB will not use the AKA field. Otherwise (the default) then the mail and news interfaces will ask you if you want to add an alternate name when a name-change is noticed, and will ask you whether the new name should be made the primary one. Note that if bbdb-quiet-about-name-mismatches is true, you will not be asked any questions about alternate names.

bbdb-readonly-p

If this is true (default: nil), then nothing will attempt to change the database implicitly, and you will be prevented from doing it explicitly. If you have more than one emacs running at the same time, you might want to arrange for this to be set to t in all but one of them.

bbdb-auto-revert-p

If this variable is true (default: nil) and the BBDB file is noticed to have changed on disk, it will be automatically reverted without prompting you first. Otherwise you will be asked. (But if the file has changed and you have made changes in memory as well, you will always be asked.)

bbdb-notice-auto-save-file

If this is true (default: nil), then the BBDB will notice when its auto-save file is newer than the file is was read from, and will offer to revert.

bbdb-use-pop-up

If true (the default), display a continuously-updating BBDB window while in VM, MH, RMAIL, or GNUS. Each time a new message is selected, the record corresponding to that message's sender will be displayed in another window. The buffer in this other window will be in bbdb-mode, and all corresponding commands will be available.

This buffer will be positioned on the screen by finding the tallest of the windows present, and splitting it such that the bottom bbdb-pop-up-target-lines lines of the window display the `*BBDB*' buffer. With the default configurations of VM, MH, RMAIL, and GNUS, this means that the bbdb-list buffer will be just below the message-body buffer.

If this is the symbol horiz, then the BBDB window will be stacked horizontally instead of vertically, if there is room to do that tastefully.

bbdb-pop-up-target-lines

Desired number of lines in a VM/MH/RMAIL/GNUS pop-up BBDB window, default 5.

bbdb-completion-type

Controls the behavior of the bbdb-complete-name command. If nil (the default), completion is done across the set of all full-names and user-ids in the database; if the symbol name, completion is done on real-names only; if the symbol net, completion is done on network addresses only; if it is primary, then completion is done only across the set of primary network addresses (the first address in the list of addresses for a given user). If it is primary-or-name, completion is done across primaries and real names.

bbdb-expand-mail-aliases

If non-nil, expand mail aliases in `bbdb-complete-name'.

bbdb-complete-name-allow-cycling

Whether to allow cycling of email addresses when calling `bbdb-complete-name' on a completed address in a composition buffer."

bbdb-complete-name-full-completion

Show full expanded completion rather than partial matches. If t then do it always, if a number then just is the number of completions for a specific match is below that number.

bbdb-user-mail-names

A regular expression identifying the addresses that belong to you. If a message from an address matching this is seen, the BBDB record for the `To:' line will be shown instead of the one for the `From:' line. If this is nil, it will default to the value of (user-login-name).

bbdb-always-add-addresses

If this is t, then whenever the Insidious Big Brother Database notices a new email address corresponding to a person who is in the database, it will add it to the database. If this is nil (the default), then whenever a new network address is noticed for a person in the database, you will be asked whether to add the address. If this is the symbol never (really if it is not t and not nil) then new network addresses will never be automatically added.

bbdb-new-nets-always-primary

If this is t, then when the Insidious Big Brother Database adds a new address to a record, it will always add it to the front of the list of addresses, making it the primary address. If this is nil (the default), then you will be asked. If this is the symbol never (really if it is not t and not nil) then new network addresses will always be added to the end of the list.

bbdb-canonicalize-redundant-nets-p

If this is non-nil, redundant network addresses will be ignored. If a record has an address of the form foo@baz.com, setting this to t will cause subsequently-noticed addresses like foo@bar.baz.com to be ignored (since we already have a more general form of that address.) This is similar in function to one of the possible uses of the variable bbdb-canonicalize-net-hook but is somewhat more automatic. (This can't quite be implemented in terms of the canonicalize-net-hook because it needs access to the database to determine whether an address is redundant, and the canonicalize-net-hook is purely a textual manipulation which is performed before any database access.)

bbdb-message-caching-enabled

Whether caching of the message->bbdb-record association should be used for the interfaces which support it (VM, MH, and RMAIL). This can speed things up a lot. One implication of this variable being true (the default) is that the bbdb-notice-hook will not be called each time a message is selected, but only the first time. Likewise, if selecting a message would generate a question (whether to add an address, change the name, etc) you will only be asked that question the very first time the message is selected.

bbdb-offer-save

If t (the default), then certain actions will cause the BBDB to ask you whether you wish to save the database. If nil, then the offer to save will never be made. If not t and not nil, then any time it would ask you, it will just save it without asking.

1.7.2 Customization Hooks

All of the hooks variables described below may be set to a symbol or lambda expression, which will be funcalled; or may be set to a list of symbols or lambda expressions, each of which will be funcalled in turn. Almost all hooks in Emacs work this way. But notice that some of the hooks described below are called with arguments.

bbdb-list-hook

Hook or hooks invoked after the bbdb-list-buffer is filled in. Invoked with no arguments.

bbdb-create-hook

Hook or hooks invoked each time a new bbdb-record is created. Invoked with one argument, the new record. This is called before the record is added to the database. Note that bbdb-change-hook will be called as well.

bbdb-change-hook

Hook or hooks invoked each time a bbdb-record is altered. Invoked with one argument, the record. This is called before the database buffer is modified. Note that if a new bbdb record is created, both this hook and bbdb-create-hook will be called.

bbdb-mode-hook

Hook or hooks invoked when the `*BBDB*' buffer is created.

bbdb-notice-hook

Hook or hooks invoked each time a bbdb-record is "noticed," that is, each time it is displayed by the news or mail interfaces. Invoked with one argument, the new record. The record need not have been modified for this to be called - use bbdb-change-hook for that. You can use this to, for example, add something to the notes field based on the subject of the current message. It is up to your hook to determine whether it is running in GNUS, VM, MH, or RMAIL, and to act appropriately.

Also note that bbdb-change-hook will not be called as a result of any modifications you may make to the record inside this hook.

Beware that if the variable bbdb-message-caching-enabled is true (a good idea) then when you are using VM, MH, or RMAIL, this hook will be called only the first time that message is selected. (The GNUS interface does not use caching.) When debugging the value of this hook, it is a good idea to set caching-enabled to nil.

bbdb-after-read-db-hook

Hook or hooks invoked (with no arguments) just after the Insidious Big Brother Database is read in. Note that this can be called more than once if the BBDB is reverted. One possible use for this is to rename the `.bbdb' buffer; for example

(setq bbdb-after-read-db-hook '(lambda () (rename-buffer " bbdb")))

will cause the buffer visiting the bbdb-file to be called " bbdb". The leading space in its name will prevent it from showing up in the buffer list.

bbdb-load-hook

Hook or hooks invoked (with no arguments) when the Insidious Big Brother Database code is first loaded. WARNING: Slow functions should not be put on this hook, as the BBDB code will, if not loaded before, be loaded during the first use of BBDB-related Customization functions. Slow functions should be put on bbdb-initialize-hook.

bbdb-initialize-hook

Hook or hooks invoked (with no arguments) when the bbdb-initialize function is called.

bbdb-canonicalize-net-hook

If this is non-nil, it should be a function of one argument: a network address string. (Note that, unlike the other hook-variables described above, this may not be a list of functions.) Whenever the Insidious Big Brother Database "notices" a message, the corresponding network address will be passed to this function first, as a kind of "filter" to do whatever transformations upon it you like before it is compared against or added to the database. For example: it is the case that CS.CMU.EDU is a valid return address for all mail originating at a machine in the .CS.CMU.EDU domain. So, if you wanted all such addresses to be canonically hashed as user@CS.CMU.EDU, instead of as user@somehost.CS.CMU.EDU, you might set this variable to a function like this:

(setq bbdb-canonicalize-net-hook '(lambda (addr) (cond ((string-match "\\`\\([^@]+@\\).*\\.\\(CS\\.CMU\\.EDU\\)\\'" addr) (concat (substring addr (match-beginning 1) (match-end 1)) (substring addr (match-beginning 2) (match-end 2)))) (t addr))))

You could also use this function to rewrite UUCP-style addresses into domain-style addresses, or any number of other things.

This function will be called repeatedly until it returns a value EQ to the value passed in. So multiple rewrite rules might apply to a single address.

There is an example of the use of this variable in the file `bbdb-hooks.el': the function sample-bbdb-canonicalize-net-hook.

The bbdb-canonicalize-net-hook is powerful in that it allows arbitrary rewriting of addresses, however, in many cases that is overkill. The function bbdb-delete-redundant-nets can be used as a value of bbdb-change-hook to cause network addresses which appear to be "redundant" to be deleted each time a modification is made to a record.

This works as follows: suppose one gets mail from user@foo.bar.com, and then later gets mail from user@bar.com. At this point, one can generally delete the user@foo.bar.com address, since the user@bar.com address is more general. (See also the variable `bbdb-canonicalize-redundant-nets-p', which has the effect of ignoring subsequent addresses from user@quux.bar.com if the address user@bar.com is already known.)

1.7.3 Predefined Hooks

If the variable bbdb-change-hook is set to the symbol bbdb-timestamp-hook (the default), then every record in the database will have a field named `timestamp', which will always contain the date and time at which this record was created or last modified.

If the variable bbdb-create-hook is set to the symbol bbdb-creation-date-hook (the default), then every record in the database will have a field named `creation-date', which will contain the date and time at which this record was added to the database.

If the variable bbdb/mail-auto-create-p is set to the symbol bbdb-ignore-most-messages-hook, then the variable bbdb-ignore-most-messages-alist will determine which messages should have records automatically created for them. The format of this alist is

(( HEADER-NAME . REGEXP ) ... )

(("From" . "@.*\\.maximegalon\\.edu")
 ("Subject" . "time travel"))

will cause BBDB entries to be made only for messages sent by people at Maximegalon U., or (that's or) people posting about time travel.

There may be only one entry per header in this alist: that is,

(("From" . "addr1\\|addr2") ... )

is legal, but

(("From" . "addr1") ("From" . "addr2") ... )

is not.

If the variable bbdb/mail-auto-create-p is set to the symbol bbdb-ignore-some-messages-hook, then the variable bbdb-ignore-some-messages-alist will determine which messages should have records automatically created for them. This is the exact inverse of the semantics of the bbdb-ignore-most-messages-alist: the alist specifies which messages should not have records automatically created for them, instead of which should. For example,

(("From" . "mailer.daemon")
 ("To" . "mailing-list-1\\|mailing-list-2")
 ("CC" . "mailing-list-1\\|mailing-list-2"))

will cause BBDB entries to not be made for messages from any mailer daemon, or messages sent to or CCed to either of two mailing lists.

The variable bbdb/news-auto-create-p may be set to either of the above-mentioned functions as well, to get this behavior for netnews messages instead of mail messages.

If the variable bbdb-notice-hook is set to the symbol bbdb-auto-notes-hook, then the variable bbdb-auto-notes-alist may be used to automatically add text to the notes fields of the records corresponding to certain messages. The format of this alist is

(( HEADER-NAME
   (REGEXP . STRING) ... )
    ... )

for example,

(("To" ("-vm@" . "VM mailing list"))
 ("Subject" ("sprocket" . "mail about sprockets")
            ("you bonehead" . "called me a bonehead")))

will cause the text "VM mailing list" to be added to the notes field of the record corresponding to anyone you get mail from via one of the VM mailing lists. If, that is, bbdb/mail-auto-create-p is set such that the record would have been created, or if the record already existed.

The format of elements of this list may also be

(REGEXP FIELD-NAME STRING)

(REGEXP FIELD-NAME STRING REPLACE-P)

meaning add the given string to the named field. The field-name may not be `name', `aka', `address', `phone', or `net' (builtin fields) but must be either `notes', `company', or the name of a user-defined note-field.

("pattern" . "string to add")

is equivalent to

("pattern" notes "string to add")

STRING can contain \& or \N escapes like in the function replace-match. For example, to automatically add the contents of the organization field of a message to the company field of a BBDB record, you can use this:

("Organization" (".*" company "\\&"))

(Note you need two \ to get a single \ into a lisp string literal.)

If STRING is an integer N, the Nth matching subexpression is used, so the above example could be written more efficiently as

("Organization" (".*" company 0))

If STRING is neither a string or an integer, it should be a function which is called with the contents of the field, and the result of the function call is used.

If the REPLACE-P flag is true, then the string replaces the old contents instead of being appended to it.

If multiple clauses match the message, all of the corresponding strings will be added.

If the string is being appended (REPLACE-P is false or not provided) then the new string is appended to the end of the existing field value, with an intervening newline. So each piece of text automatically added to this field will go on its own line.

You can control what the separator is by putting a field-separator property on the symbol naming the field. For example, to make text automatically added to a field named newsgroups be separated by commas, you could do

(put 'newsgroups 'field-separator "; ")

This variable works for news as well. You might want to arrange for this to have a different value when in mail as when in news.

There may be only one entry per header in this alist: that is,

 (("Subject" ("\\bfoo\\b" . "Foo!!")
             ("bar" . "Bar!")))

will work, but

 (("Subject" ("\\bfoo\\b" . "Foo!!"))
  ("Subject" ("bar" . "Bar!")))

will not.

Here's a more complicated example: some people include bitmaps of themselves in their mail messages in an X-Face: header field. You can capture this field into the `*BBDB*' with the following:

(setq bbdb-auto-notes-alist
      (append bbdb-auto-notes-alist
              (list "x-face"
                    (list (concat "[ \t\n]*\\([^ \t\n]*\\)"
                                  "\\([ \t\n]+\\([^ \t\n]+\\)\\)?"
                                  "\\([ \t\n]+\\([^ \t\n]+\\)\\)?"
                                  "\\([ \t\n]+\\([^ \t\n]+\\)\\)?"
                                  )
                          'face
                          "\\1\\3\\5\\7"))))

(The calls to list and concat are just for readability, it could easily be a constant.) The tricky bit here is that it strips out the newlines and whitespace used for header continuation, which are not actually a part of the face data. So though the mail message may have the face data on multiple lines, the entry in the `*BBDB*' will be just one line.

bbdb-auto-notes-ignore is an alist of headers and regexps to ignore in bbdb-auto-notes-hook. Each element looks like

(HEADER . REGEXP)

for example,

("Organization" . "^Gatewayed from\\|^Source only")

would exclude the phony Organization: headers in GNU mailing-lists gatewayed to the gnu.* newsgroups. Note that this exclusion applies only to a single field, not to the entire message. For that, use the variable bbdb-auto-notes-ignore-all.

bbdb-auto-notes-ignore-all is an alist of headers and regexps which cause the entire message to be ignored in bbdb-auto-notes-hook. Each element looks like

(HEADER . REGEXP)

for example,

("From" . "BLAT\\.COM")

would exclude any notes recording for message coming from BLAT.COM. Note that this is different from bbdb-auto-notes-ignore, which applies only to a particular header field, rather than the entire message.

1.8 Utilities

This section describes BBDB functionality that does not fit neatly into other sections.

1.8.1 bbdb-ftp

The bbdb-ftp utility enables the storage of FTP sites as BBDB records. The bbdb-create-ftp-site function is used to create a BBDB record for an FTP site. The command will prompt for information needed to create the record. The FTP site for a given record can be accessed with the bbdb-ftp command.

1.8.2 bbdb-print

bbdb-print is a utility for pretty-printing entries from the BBDB using TeX. It is invoked by pressing P in the *BBDB* buffer - this will cause all records in the BBDB to be printed as governed by the variables described below. If P is prefixed by a *, only the currently-displayed records will be printed. Once invoked, bbdb-print will generate the TeX output in a buffer named `~/bbdb.tex' (controlled by bbdb-print-file-name). The generated output will be shown, and must be manually saved.

For each record printed, bbdb-print will look for the presence of the tex-name field. If this field is found, its value will be printed instead of the name field for the record in question. This field is intended to allow the storage of names with accents or other characters that would be illegal in the address portion of a message. While other fields have special characters quoted by bbdb-print as described below, the contents of the tex-name field (if present) are used verbatim.

The following variables govern the printing of records (and of the printing of the fields therein):

bbdb-print-omit-fields

This variable should be set to a list of the fields that are not to be printed by bbdb-print. It defaults to:

(omit tex-name aka mail-alias)

bbdb-print-file-name

The name of the file where generated TeX output is to be stored. The default is `~/bbdb.tex'.

bbdb-print-require

The fields required for printing a record. This allows, for example, only records with phone numbers to be printed. The value of the value of the variable will be evaluated once for each record, and the record will be printed only if the evaluation returns a non-nil value. The symbols name, company, net, phone, address, and notes will be set to the appropriate values during evaluation; they will be nil if the field does not exist or is elided.

The value of this variable can be any lisp expression, but typically it will be used for a boolean combination of the field variables, as in the following simple examples:

• Print only people whose phone numbers are known:
  
  (setq bbdb-print-require 'phone)
• Print people whose names AND companies are known:
  
  (setq bbdb-print-require '(and name company))
• Print people whose names, and either addresses OR phone numbers are known:
  
  (setq bbdb-print-require '(and name (or address phone))).

  The default value is:
  

  (or address phone)

bbdb-print-alist

Formatting options for all formats of bbdb-print. The value is an alist of the form

((OPTION . VALUE) (OPTION . VALUE) ...)

Separate settings can be configured using bbdb-print-brief-alist and bbdb-print-full-alist. Settings in these variables will override the ones in bbdb-print-alist.

The possible options and legal settings are:

columns

1, 2, 3, 4, quad (4 little 2-column pages per sheet) or grid (12 credit-card-sized pages per sheet).

separator

0-7, the style of heading for each letter. 0=none, 1=line, 2=boxed letters, 3=large boxed letters, 4=large letters, 5=letters with lines, 6=letters with suits, 7=boxed letters with suits.

omit-area-code

A regular expression matching area codes to omit.

phone-on-first-line

If t, the first phone number will be put on the same line as the name. If nil, the name will be the only text on the line. If set to a string, the field of type phone whose name matches the string will be used. The string should be a valid regular expression.

n-phones

Maximum number of phone numbers to include.

n-addresses

Maximum number of addresses to include.

include-files

List of TeX files to \input. If the filenames are not absolute, the files must be located somewhere in the TeX input path.

ps-fonts

Postscript fonts (TimesNewRoman and Courier) will be used if the value is non-nil. Standard TeX fonts (ec-fonts) will be used otherwise.

font-size

The integer point size of the font to be used.

hsize

The horizontal dimension of the pages. The value must be a string, and must be a valid TeX dimension. Alternatively, the TeX default will be used if the value is nil.

vsize

The vertical dimension of the pages. The value must be a string, and must be a valid TeX dimension. Alternatively, the TeX default will be used if the value is nil.

hoffset

The TeX output will be offset to the right by the value of this option. The value must be either a string containing a valid TeX dimension or nil or 0 to use the default TeX horizontal offset.

voffset

The TeX output will be offset downward by the value of this option. The value must be either a string containing a valid TeX dimension or nil or 0 to use the default TeX vertical offset.

quad-hsize

Horizontal size to be used for the individual pages in the quad format. The value must be a string containing a valid TeX dimension.

quad-vsize

Vertical size to be used for the individual pages in the quad format. The value must be a string containing a valid TeX dimension.

The default value is

((omit-area-code . "(AREA-CODE)") (phone-on-first-line . "^[ \t]*$") (ps-fonts . nil) (font-size . 6) (quad-hsize . "3.15in") (quad-vsize . "4.5in"))

Where AREA-CODE is the value of bbdb-default-area-code or `000' if there is no default area code.

bbdb-print-brief-alist

Extra options for the bbdb-print brief format. The value(s) of this variable can either supplement or override the values in bbdb-print-alist. The format and possible values of this variable are as in bbdb-print-alist, described above.

The default value is

((columns . 1) (separator . 1) (n-phones . 2) (n-addresses . 1) (include-files "bbdb-print-brief" "bbdb-cols"))

bbdb-print-full-alist

Extra options for the bbdb-print non-brief format. The value(s) of this variable can either supplement or override the values in bbdb-print-alist. The format and possible values of this variable are as in bbdb-print-alist, described above.

The default value is

((columns . 3) (separator . 2) (include-files "bbdb-print" "bbdb-cols"))

bbdb-print-prolog

TeX statements to include at the beginning of the bbdb-print output file.

bbdb-print-epilog

TeX statements to include at the end of the bbdb-print output file.

1.8.3 bbdb-snarf

bbdb-snarf provides the ability to generate BBDB records from raw text. If invoked as bbdb-snarf, it attempts to generate the record from the information around point (the cursor). If invoked as bbdb-snarf-region(6), the active region is used to generate the record.

Restrictions:

1. bbdb-snarf currently recognizes only US-style phone numbers.
2. bbdb-snarf works best with things that look like mailing addresses.

Example of an address that bbdb-snarf will recognize:

another test person
1234 Gridley St.
Los Angeles, CA 91342
555-1212
test@person.net
http://www.foo.bar/
other stuff about this person

1.8.4 bbdb-srv

bbdb-srv provides the ability to initiate the display of BBDB records from outside of Emacs. This allows external programs to cause the BBDB record for a given person to appear in the running Emacs when, for example, mail is recieved from that person. One specific application, described below, is the integration of Netscape and the BBDB, allowing the display of BBDB records corresponding to Mail and/or News messages displayed in Netscape.

bbdb-srv is composed of two parts - an external Perl script (`utils/bbdb-srv.pl'), and an Emacs Lisp file (`lisp/bbdb-srv.el'). The external portion is used to send commands to the internal portion.

NOTE: bbdb-srv requires gnuserv and itimer, both included with XEmacs. gnuserv must be started with the gnuserv-start command before bbdb-srv can be used.

In it's most basic form, mail or news headers are passed to the Perl script. The Perl script then causes the BBDB record (if any) corresponding to the passed headers to be displayed in the running Emacs. While bbdb-srv will operate with just a `From:' field, it works better when passed both `From:' and `To:' headers. When bbdb-srv notices that the logged-in user is named in the `From:' header, it will attempt to display the record (if any) for the person named in the `To:' header. If no `To:' header is passed, it falls back on the record (if any) for the logged-in user (the person named in the `From:' header).

An example manual invocation of bbdb-srv is as follows:

% cat |bbdb-srv.pl
From: Jamie Zawinski <jwz@netscape.com>
To: Matt Simmons <simmonmt@acm.org>
CTRL-D

If the invoking user is Jamie Zawinski, the record for Matt Simmons (if any) will be displayed. If the invoking user is not Jamie Zawinski, the record for Jamie Zawinski (if any) will be displayed.

As mentioned above bbdb-srv can be used with Netscape Mail and Netscape News. Please note that it can only be used with the UNIX versions 3.0b2 and greater of these applications. To allow Netscape to use bbdb-srv, set the `NS_MSG_DISPLAY_HOOK' variable to bbdb-srv.pl(7) as follows:

Bourne Shell (`/bin/sh') and variants:

# NS_MSG_DISPLAY_HOOK bbdb-srv.pl
# export NS_MSG_DISPLAY_HOOK

C-Shell (`/bin/csh') and variants:

% setenv NS_MSG_DISPLAY_HOOK bbdb-srv.pl

The following variables can be used to customize the behavior of bbdb-srv:

bbdb/srv-auto-create-p

This variable is similar to bbdb/news-auto-create-p and bbdb/mail-auto-create-p. That is, when headers are passed in to bbdb-srv, a new BBDB record can be created if none exists, depending on the value of this variable. Possible values are:

t

Automatically create new BBDB records if the headers passed in do not correspond to an already-existing record.

nil

Do not automatically create new BBDB records.

FUNCTION

FUNCTION is called. If it returns t, a record will be created for the person named in the `From:' header. If it returns nil, no record will be created.

A suggested function for use is bbdb/srv-auto-create-mail-news-dispatcher. This function will attempt to determine the source of the passed headers - whether they were part of a mail message or of a news article. The action (if any) dictated by the value of either bbdb/mail-auto-create-p or bbdb/news-auto-create-p, based on the determined source of the passed headers.

bbdb/srv/display-delay

bbdb-srv pauses between displaying the records corresponding to each passed set of headers. This variable controls the length of time (in seconds) of the delay between the display of different records. Note when setting this variable that only one set of headers can be queued at a time. If three sets of headers are passed to bbdb-srv in less than the delay time, only the first and last will be displayed.

1.9 Internals

This section is currently a dumping ground for things that should eventually go here, but were found elsewhere in the file.

INFORMATION IN THIS SECTION IS FOR INFORMATIONAL PURPOSES ONLY. IT SHOULD NOT BE TAKEN AS DOCUMENTATION OF AN EXTERNAL API. EVERYTHING LISTED BELOW IS SUBJECT TO CHANGE WITHOUT NOTICE

The first time you use one of the BBDB commands, this file is read into an emacs buffer, and remains there. As you make changes to the database, this buffer is changed as well, ensuring that if it is auto-saved, it will be saved in its most current state.

1.9.1 BBDB data file format

The data file is arranged in a hierarchical fashion. At the top level are vectors, with one vector per database record. It is very important that each vector be on its own line, as the BBDB builds and stores markers based on this layout. The markers are then used to increase the speed of database modifications (more on this later). The record vectors contain the individual fields of the record. These fields can be of any type, but are currently integers, strings, lists of strings, alists, vectors, or lists of vectors. In the case of fields that contain one or more vectors, they can be further broken down in terms of the fields of their component vectors.

In an effort to provide a more concrete example to illustrate the above, and to provide a reference for database accessor and modifier functions, we describe the database format below. This description starts with the fields of the individual record vectors, and drills down through the vectors used by some of the fields.

1.9.1.1 Record Vectors

The phone, address and cache vector fields are described below. Please note that, as indicated in the table above, the cache is present only in the internal version of the database - it is not written out as part of the `.bbdb' file.

In addition, the accessor and modifier functions for the notes alist are described.

1.9.1.2 Phone Vectors

To access the fields in the below table, you must first get the list of phone vectors using the bbdb-record-phones function. Note that if you alter the phones field with the bbdb-record-set-phones function, you are altering the entire phones list for the given record. Use the modifier functions below for modifications to individual phone vectors.

1.9.1.3 Address Vectors

To access the fields in the below table, you must first get the list of address vectors using the bbdb-record-addresses function. Note that if you alter the addresses field with the bbdb-record-set-addresses function, you are altering the entire addresses list for the given record. Use the modifier functions below for modifications to individual address vectors.

1.9.1.4 Cache Vector

This vector is present only in the internal database representation. It is not written out to the database file because it contains information aggregated from the rest of the record that is reconstructed when the database is read. To write the cache information to the database file would increase the risk of database inconsistency, and would violate the principles of normalization.

To access the cache fields using the functions listed below that begin with bbdb-cache-, you must first get the cache vector using the bbdb-record-cache function. The functions that begin with bbdb-record- get the cache vector internally. Note that if you alter the cache field in the high-level record with the bbdb-record-set-cache function, you are altering the entire cache vector for the given record. Use the modifier functions below for modifications to individual cache fields.

The functions listed above will return nil if their respective cache fields are not set. The functions listed below will return the value of their cache fields if set, but will also build (and set) the correct field values if the fields are unset:

bbdb-record-name

Return the name in the Name Cache field of the cache (if set). If the name has not been built yet (if the field is nil), the name is built, stored in the Name Cache field, and returned.

bbdb-record-sortkey

Return the name it the Sort Key field of the cache (if set). If the Sort Key field has not yet been set (if the field is nil), the Sort Key is built, stored in the Sort Key field, and returned.

1.9.1.5 Notes String or Alist

If there is only a single note for a given record, the notes field for that record will be a string. If there is more than one note, the notes field will be an association list (alist) with elements of the form

where NAME is the symbol for the name of the note, and VALUE is the value of the note.

1.9.1.6 Example BBDB record

1.10 Mailing Lists

(If you are looking for a way to create mailing lists with BBDB, you should be looking at the section on See section 1.4.3 Mail Sending Interfaces.)

There are three mailing lists for the BBDB. bbdb-info@lists.sourceforge.net gets moderate traffic, and is intended for the discussion and distribution of development versions of the BBDB. Users of development versions of the BBDB should be subscribed to this list. It is also to this list that bugs should be reported. See section 1.13.1 Known Bugs. for instructions on submitting bug reports.

The second mailing list bbdb-announce@lists.sourceforge.net has very low volume. Any user of the BBDB should consider subscribing to this list, as new releases and security issues will be posted here.

A third mailing list, bbdb-cvs@lists.sourceforge.net, is intended for developers to follow the changes made to the BBDB development version. Developers of the BBDB should consider to subscribe to this list.

1.11 Changes in this Version

1.11.1 Major Changes

Database File Version Change

(the following version-migration text will move in a future version. It is in this section currently because new users will likely not experience it)

There has been a version change in the BBDB database file. The new version supports non-US zip codes, and an additional "Country" field in addresses.

An automatic version-migration mechanism has been implemented that allows older version `.bbdb' files to either be migrated to the new version, or used as-is without migrating. When the BBDB detects a database file with an old version, it will display the features that have been introduced in the database file from the time of the older version's implementation. It will offer the choice of migration or use of the BBDB with the older-version file.

If migration is chosen, the database file will be automatically changed to the new format. If migration is declined, the file will kept in the older format in the .bbdb buffer, but will be stored internally in the new format. When changes need to be made to the .bbdb buffer, changed records will be reverse-migrated from their internal version to that of the disk file.

1.11.2 Other Changes

TeX Output

By default, ec fonts are used for TeX output instead of cm fonts. With the ps-fonts option set in bbdb-print-alist, TimesNewRoman and Courier fonts are used.

1.12 The Latest Version

Released versions of the BBDB can be found at the following site:

• WWW: http://bbdb.sourceforge.net
• FTP: ftp://ftp.sourceforge.net/pub/bbdb

Development versions of the BBDB can be obtained in the following ways:

• WWW: http://bbdb.sourceforge.net
• Anonymous CVS: See http://bbdb.sourceforge.net for instructions.

Users of development versions of the BBDB should subscribe to the bbdb-info mailing list. See section 1.10 Mailing Lists.

1.13 The Future

The future consists of Bugs and Features.

1.13.1 Known Bugs

1. M-TAB conflicts with ispell. Workaround: The binding installed by the BBDB for address completion/expansion conflicts with that used by ispell. The suggested workarounds are to rebind the ispell key (the BBDB binding is not configurable at this time), to manually invoke ispell via M-x, or to not use ispell completion functionality in BBDB-enabled message composition buffers. The following is an example of such a rebinding, supplied by Kai Großjohan:

   (defun my-message-mode-keys () (define-key message-mode-map (kbd "M-TAB") 'bbdb-complete-name)) (add-hook 'message-mode-hook 'my-message-mode-keys)

2. BBDB and abbrev expansion is inconsistent. Workaround: Currently, M-TAB must be used to expand/complete against BBDB names and net addresses, and TAB must be used to expand abbrevs (the values in the mail-alias field). Unification is planned for a future version.

3. The *BBDB* buffer does not always come up when the first article in a Gnus Summary Buffer is selected. Workaround: Pressing g to reload the article. This will cause the *BBDB* buffer to be displayed.

4. Expansion will fail when the name to be expanded is a subset of the name for another record. For example, if you have entries for `John' and `Johnathan', you will not be able to expand the name for `John'. Workaround: Use the net address for the subset name (`John' in this example).

It is commonly known that there are no bugs in the BBDB. Bugs found in defiance of this rule should be submitted using M-x bbdb-submit-bug-report. These bug reports will bbe sent to the bbdb-info mailing list (see section 1.10 Mailing Lists) and are available from public archives. Other big brothers may be reading your bug reports.

1.13.2 TODO List

The Near Future

• Add bbdb/MUA-delete-sender-record

• Configurable completion. Should allow user to specify "complete on names first, then nets", etc.

• More variables for upgrading. Specifically a variable that lets users specify extra fields for upgrading (an alist `(a . b)' that says field `a' should be upgraded the same way as field `b'.

• Change all functions that switch on MUAs to use compose-mail (ex: bbdb-send-mail-internal).

• Soren Dayton's method for generically extending the BBDB with special-purpose fields

• Sorting records on alternate keys. bbdb-sort-by from Boris Goldowsky.

• Sorting individual types of fields - Sam Steingold's method.

• Button 3 menus (Mark Moll and Soren Dayton)

• Conditionalized erasure of properties when text is pasted into the BBDB.

• Should notice when there are `Reply-To:' addresses.

• Should have a command for merging together two divergent copies of a `.bbdb' file (in case you read mail on one machine and news on another, for instance.)

• The `*BBDB*' buffer should be resized to exactly fit what it's displaying, even when not in "electric" mode.

• It should be possible to do completion on last names as well as first names.

• The BBDB buffer is left at the top of the stack when GNUS is exited because GNUS runs its exit-hooks too early. This should be fixed.

• String area codes (German area codes can begin with zeroes) patch from Michael Sperber <sperber@informatik.uni-tuebingen.de>

• Default country variable, similar to bbdb-default-area-code.

• Make format self-describing in comment

• ISO-8859-x characters in records for printing.

• Prefix for W (bbdb-www) command to allow selection of different addresses.

• Generalized buttons (via extents) for fields. Example: `(a . b)' means create button that calls `b' for each entry in the `a' field.

• Remove support for GNUS. Start with lisp Makefile (remove nntp and gnus loads).

• More flexible auto-addition. Conditionalizing of addition (conditions or supplied function). Prompt if multiple records that meet criteria exist.

• Different output formats. See Toby Speight's <s8iurdodvm.fsf@plato.ansa.co.uk> and Bin Mu's <199801221605.KAA23663@DerivaTech.Com>.

• Generalized area-code-split program that could split, for example, based on input copied (or straight fetch of page) from the Bellcore NANP page.

• Print multivalue (comma-separated) fields with one value per line

• Easier BBDB extension. See <xcdyb30f3hb.fsf@ra.cs.uchicago.edu> from Soren Dayton.

• Take birthdays from the BBDB, add them to calendar. From Boris Goldowsky.

• Make mail aliases file for other mailers. From Boris Goldowsky.

• Various patches from Boris Goldowsky in `bbdb-ext'.

• Various other patches:
  • bbdb-filters-0.2
  • bbdb-frame.el
  • bbdb-letter-1.0
  • bbdb-plz
  • bbdb-query
  • country
  • country-info
  Note that these files have not been investigated. They may or may not be incorporated.

• Ability to remove all properties from copied strings.

Not-So-Near Future

• Fix Gnus scoring so it rebuilds when gnus-score disappears

• Multiline note fields

• Change key to be some kind of unique number

• There should be better support for non-American addresses and phone numbers. This might be Near Future if somebody volunteers to send me patches.

• Should reimplement "electric" mode to not be so broken.

• The *C-o keystroke should add a field to all displayed records. Perhaps *; should append some text to an arbitrary field of all displayed records.

• Multiple `.bbdb' files with precedence relationships. See Wes Hardaker's
  <sdu3awz75a.fsf@oakdale.ucdavis.edu>

• Automatically grab information about a person from their sig. See Graham Clark's info-bbdb post <6282.199706161624@havra.dcs.ed.ac.uk> and Adrian Aichner's info-bbdb post <rxsiuzebpgp.fsf@midnight.ecf.teradyne.com>. Would like to have bbdb-snarf attack the sig then compare the snarfed data with the header data.

Thoughts

• Are there enough hooks?

• The interfaces should share more code.

• The bbdb-create-internal function should be more forgiving.

• More * commands in general, including *d.

1.13.3 End of Life (EOL) Statements

The items in the following list describe items for which support will be removed in coming versions of the BBDB. The items listed are guaranteed to be supported and present only until the EOL date. They may be removed without warning at any time thereafter.

1. advertized-bbdb-delete-current-field-or-record
   Support for this function will be removed for version 2.2. It is recommended that all code depending on this variable be switched to use bbdb-delete-current-field-or-record. The two functions have the same calling conventions and effects. This EOL statement was added for version 2.1.

2. Support for the GNUS (not Gnus) newsreader
   The GNUS-specific parts of the BBDB will be actively removed for the 2.2 release. No further maintenance and/or bugfixes are planned for GNUS code at this time. This EOL statement was added for version 2.1.

1.14 Thanks

Thanks to everyone on the info-bbdb mailing list for many useful suggestions. This hack would be far less insidious without their input!

Thanks list for versions after 2.00.06.

Thanks to Alex Schroeder, Ronan Waide, Thomas DeWeese, Robert Fenk, Didier Verna, Bill Carpenter.

Thanks list for versions after 1.51 prior to and including 2.00.06.

Thanks to Adrian Aichner, Kees de Bruin, David Carlton, Soren Dayton, Brian Edmonds, Boris Goldowsky, Seth Golub, John Heidemann, Christopher Kline, Carsten Leonhardt, Hrvoje Niksic, Jens-Ulrik Hoger Petersen, Colin Rafferty, Matt Simmons, Sam Steingold, Marco Walther, Christoph Wedler.

Thanks list for versions prior to and including 1.51.

And special thanks to Sebastian Kremer, Joe Wells, Todd Kaufmann, Andy Norman, Ivan Vazquez, Stewart Clamen, Roland McGrath, Dave Brennan, Kimball Collins, Dirk Grunwald, Philippe Queinnec, Boris Putanec, Dave Disser, Francois Felix Ingrand, Sean Owens, Guido Bosch, Lance Brown, Tom Emerson, George Hartzell, Luis Miguel Silveira, Kimmo Suominen, Derek Upham, David Zuhn, Rod Whitby, Richard Mlynarik.

Last, but not least, thanks to Jamie Zawinski for writing BBDB in the first place.

This document was generated by XEmacs shared group account on December, 19 2009 using texi2html 1.65.