Format Program Selection

The -format string and -form formatfile specify a format string or file to read. A format string, if given, must be a single argument to the -format switch. If a format file name is passed to the -form, switch, the file is searched for using the normal nmh rules: absolute pathnames are accessed directly, tilde expansion is done on usernames, and files are searched for in the user's Mail directory as specified in their profile. If not found there, the directory “/etc/nmh” is checked.

Mode Selection and Component Specification

fmttest has four operating modes - address, raw, date, and message - which are selected by the -address, -raw, -date, and -message switches, respectively.

Address mode treats every argument as an email address to be processed by nmh's email parser using the specified format program. The parsed address is made available as a special %{text} component escape, and the output from the program is printed on standard output. If there was an error parsing the email address the error message is stored in the %{error} component escape. If no format program is given on the command line, the following default program is used:

%<{error}%{error}: %{text}%|%(putstr(proper{text}))%>

Address mode is equivalent to ap(8).

In raw mode, no processing of the specified arguments is done. Each argument is run against the specified format program with the argument text available in the %{text} component. You must specify a format with -form or -format when using raw mode.

Date mode is identical to raw mode, with one exception: if no format is specified, the following format string is used:

%<(nodate{text})error: %{text}%|%(putstr(pretty{text}))%>

Date mode is equivalent to dp(8).

In message mode the arguments to fmttest are interpreted as an optional folder and messages. fmttest will read each specified message and make all of the components in the message available to the format program. Also, the appropriate information for the %(msg), %(cur), %(unseen), and %(size) function escapes will be made available for each message. If the -file switch is given, the arguments are interpreted as filenames instead of message numbers, but otherwise the behavior is the same (except that the %(msg), %(cur), and %(unseen) function escapes will not provide any useful information).

The default format used in address mode is the default format used by scan. The following command can replicate the functionality of the repl command.

fmttest -nodupaddrs -form replcomps -outsize max [+folder] message

Regardless of the mode, other components can be provided to the format program by the use of the --component switch. For example, the following program will test the use of the “encrypted” component:

fmttest --encrypted yes -message cur

In message mode, components supplied on the command line will override components from messages.

Additional Switches

The -dupaddrs and -nodupaddrs switches control whether duplicate addresses are allowed or suppressed by the FORMATADDR instruction, which is used by the `%(formataddr)' function escape. In normal operation duplicate addresses are only suppressed by repl.

The -ccme and -noccme switches control whether or not to count the user's local mailbox as a duplicate address. This replicates the behavior of the -cc me switch to repl, and only applies if -nodupaddrs is in effect.

The -outsize switch controls the maximum number of printable characters that the format engine will produce. Characters marked as non-printing by the format engine with `%(zputlit)', characters with zero width, and extra bytes that are part of a multibyte character are not counted against this total. Two special values are supported: “max”, which means as many characters as the format engine can produce (limited by internal buffers), and “width”, which will set the value to the width of the terminal. In message mode it defaults to “width”, otherwise the default is “max”.

The -width switch controls the column width which is used by the `%(width)' function escape. It defaults to the terminal width.

The -msgnum, -msgcur, -msgsize, and the -unseen switches all control the values used, respectively, by the following function escapes: `%(num)', `%(cur)', `%(size)', and `%(unseen)'. If none are supplied, these values are taken from the message in message mode; in all other modes the default values are 0.

Compiling and Tracing Format Programs

The -dump switch outputs the complete set of format instructions for the specified format program. The -trace switch will output each format instruction as it is being executed, and show the values of the num and str registers if they have changed from the previous instruction. The output buffer is also printed if it has changed from the previous instruction.

Format Instructions

It should be noted that there is not a one-to-one correspondence between format escapes and format instructions; many instructions have side effects. Instructions prefixed with “LV” generally return a integer into the num (value) register; instructions prefixed with a “LS” return a string into the str register.

Instruction	Description
COMP	Output component
COMPF	Formatted output component
LIT	Output literal text
CHAR	Output single character
NUM	Output the num register
NUMF	Formatted output of the num register
STR	Output the str register
STRF	Formatted output of the str register
STRFW	Not used
PUTADDR	Output address list in str register
STRLIT	Output str, no space compression
STRLITZ	Like STRLIT, but not counted against width
LS_COMP	Write component to str register
LS_LIT	Write literal to str register
LS_GETENV	Write environment var to str register
LS_DECODECOMP	Decode RFC 2047 encoded component to str register
LS_DECODE	Decode RFC 2047 encoded string to str register
LS_TRIM		Trim trailing whitespace from str register
LV_COMP		Convert component to integer, store in num register
LV_COMPFLAG	Set num to 1 if TRUE set in component
LV_LIT	Load literal value into num register
LV_DAT	Load value from dat array into num register (see note)
LV_STRLEN	Set num to the length of str
LV_PLUS_L	Add value to num register
LV_MINUS_L	Subtract value from num register
LV_DIVIDE_L	Divide num register by value
LV_MODULO_L	num modulo value
LV_CHAR_LEFT	Store remaining number of printable chars in num
LS_MONTH	Write short name of month to str from date component
LS_LMONTH	Write long name of month to str from date component
LS_ZONE		Write time zone offset to str from date component
LS_DAY		Write short name of day of week to str from date component
LS_WEEKDAY	Write long name of day of week to str from date component
LS_822DATE	Write RFC 822 compatible date to str from date component
LS_PRETTY	Write date with “pretty” timezone to str
LV_SEC	Write date component seconds to num
LV_MIN	Write date component minutes to num
LV_HOUR	Write date component hour to num
LV_MON	Write date component numeric month to num (start at 1)
LV_YEAR	Write date component year to num
LV_YDAY	Write date component Julian day to num
LV_WDAY	Write date component day of week to num (0 == Sunday)
LV_ZONE	Write date component time zone offset to num
LV_CLOCK	Write date component in Unix epoch time to num
LV_RCLOCK	Write offset of date component from current time to num
LV_DAYF	Write 1 to num if day of week is explicit
LV_DST	Write 1 to num if DST is in effect for date component
LV_ZONEF	Write 1 to num if timezone is explicit
LS_ADDR	Write email address of addr component to str
LS_PERS	Write personal name of addr component to str
LS_MBOX	Write mailbox (username) of addr component to str
LS_HOST	Write host of addr component to str
LS_PATH	Write host route of addr component to str
LS_GNAME	Write group name of addr component to str
LS_NOTE	Write note portion of addr component to str
LS_822ADDR	Write “proper” RFC 822 version of addr component to str
LS_FRIENDLY	Write friendly (name or note) of address component to str
LS_UNQUOTE	Remove RFC 2822 quotes from string
LV_HOSTTYPE	Set num to type of host (0=local, 1=network)
LV_INGRPF	Set num to 1 if address was inside of group
LV_NOHOSTF	Set num to 1 of no host was present in address component
LOCALDATE	Convert date component to local timezone
GMTDATE	Convert date component to GMT
PARSEDATE	Parse date component
PARSEADDR	Parse address component
FORMATADDR	Add address component to list in str
CONCATADDR	Like FORMATADDR, but will not suppress duplicates
MYMBOX	Set num if address component is a local address
SAVESTR		Save str register temporarily
DONE	End program
NOP	No operation
GOTO	Jump to new instruction
IF_S_NULL	Branch if str is NULL
IF_S	Branch if str is not NULL
IF_V_EQ	Branch if num is equal to value
IF_V_NE	Branch if num is not equal to value
IF_V_GT	Branch if num is greater than value
IF_MATCH	Branch if str contains string
IF_AMATCH	Branch if str starts with string
S_NULL	Set num to 1 if str is NULL
S_NONNULL	Set num to 1 if str is not NULL
V_EQ	Set num to 1 if num equals value
V_NE	Set num to 1 if num does not equal value
V_GT	Set num to 1 if num is greater than value
V_MATCH	Set num to 1 if str contains string
V_AMATCH	Set num to 1 if str starts with string

The LV_DAT instruction is a bit special. Callers of the format library pass in an array of integers that are used by certain format escapes. The current list of format escapes and the indexes they use are:

dat[0]	%(num)
dat[1]	%(cur)
dat[2]	%(size)
dat[3]	%(width)
dat[4]	%(unseen)