rs@uunet.UU.NET (Rich Salz) (07/10/87)
Submitted-by: Wombat <rsk@j.cc.purdue.edu>
Posting-number: Volume 10, Issue 54
Archive-name: ease/Part04
#! /bin/sh
# This is a shell archive. Remove anything before this line, then unpack
# it by saving it into a file and typing "sh file". To overwrite existing
# files, type "sh file -c". You can also feed this as standard input via
# unshar, or by typing "sh <file", e.g.. If this archive is complete, you
# will see the following message at the end:
# "End of archive 4 (of 4)."
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test -f 'doc/ease.paper' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'doc/ease.paper'\"
else
echo shar: Extracting \"'doc/ease.paper'\" \(29063 characters\)
sed "s/^X//" >'doc/ease.paper' <<'END_OF_FILE'
X.LP
X.TL
XEase: A Configuration Language
Xfor Sendmail
X.AU
XJames S. Schoner
X.AI
XPurdue University Computing Center
XWest Lafayette, Indiana 47907
X.sp 2
X.I
X.ce
XABSTRACT
X.R
X.PP
XThe rapid expansion of computer networks and ensuing variation among mailing
Xaddress formats have made address interpretation an increasingly
Xcomplex task. In the UNIX* 4.2BSD operating system, a program named
X\fIsendmail\fR was introduced which provided a
Xgeneral internetwork mail routing facility. This facility has significantly
Xdiminished the complexity of handling address interpretation.
X.PP
X\fISendmail\fR's address interpretation is based on a rewriting
Xsystem composed of
Xa number of rewriting rules (or productions) arranged as part of a
Xconfiguration file. Unfortunately, the syntactical format of a
Xconfiguration file for \fIsendmail\fR is both terse and rigid, making it
Xrather difficult to modify. The standard format certainly serves its
Xpurpose, but, as
Xthe need to change these configurations increases in frequency, a more
Xreadable format (i.e., one that is similar to the format
Xof modern programming languages) is required to permit reasonably
Xquick modifications to the configuration. As a solution to this problem,
X\fBEase\fR
Xprovides a level of abstraction which eliminates most of the current
Xsyntactic hindrances
Xfaced by programmers who must reconfigure \fIsendmail\fR's
Xaddress parsing scheme.
X.PP
XAs a high-level specification format, \fBEase\fR is proving to be an
Xexcellent alternative to \fIsendmail\fR's cryptic
Xconfiguration file syntax. The syntactic structures of \fBEase\fR
Xare patterned after modern language constructs, making the language
Xeasy to learn and easy to remember. The format of the address rewriting
Xrule is perhaps the most significant syntactical improvement. It was
Xundoubtedly
Xthe most needed improvement. Nevertheless, every element of a configuration
Xfile is structurally enhanced through the use of \fBEase\fR.
X.FS
X* UNIX is a trademark of AT&T Bell Laboratories.
X.FE
X.sp 2
X.NH
XIntroduction
X.PP
XThe \fBEase\fR language is a high-level specification format for \fIsendmail\fR's
Xconfiguration file. The motivation for its development
Xwas to fulfill a goal of providing a readable and easily modifiable
X\fIsendmail\fR configuration file format. \fBEase\fR fulfills this goal by
Xshielding the programmer from the cryptic configuration specification required
Xby \fIsendmail\fR and providing a high-level language with which the programmer
Xmay specify all modifications to a configuration file. The development
Xof Ease coincided with
Xthe development of an \fBEase\fR translator, \fIet\fR,
Xwhich translates a configuration file written
Xin \fBEase\fR to an
Xequivalent file of the standard format accepted by \fIsendmail\fR.
X.NH
XEase in Profile
X.PP
XAs will be seen in the next section, the syntax of \fBEase\fR is quite
Xreadable and easy to learn. In order to acquire a relevant perspective
Xon this issue,
Xthe reader is advised to examine a raw configuration file for \fIsendmail\fR (the
Xoutput
Xof the \fBEase\fR translator, \fIet\fR, will do nicely). The raw syntax, while
Xquite fitting for quick translation, can prove to be a programmer's nightmare.
X.PP
XUndoubtedly, one of the more prominent features of \fBEase\fR is the ability
Xto attach
Xnames to address fields. When address field names are well-chosen, a distinct,
Xself-documenting quality becomes a visible part of the address rewriting
Xrules. Ostensibly, address field names provide a new level of semantic
Xabstraction. A brief comparison of the formats can be accomplished by examining
Xthe following equivalent representations of an address pattern:
X.DS
X user_path@host_name (\fBEase\fR format)
X $+@$- (raw format)
X.DE
XIn the above, \*Quser_path\*U represents a field of one or more address
Xtokens, and \*Qhost_name\*U represents one address token exactly. These
Xtoken fields are represented by \*Q$+\*U and \*Q$-\*U in the raw format. Clearly,
Xthe \fBEase\fR format is preferable, not only for increased readability, but
Xstructural comprehension as well.
X.PP
XOther features of \fBEase\fR include ruleset naming, long identifiers for
Xmacros and classes, flow-of-control structures, and free formatting. In
Xaddition, the C language preprocessor (cpp) can be used for file inclusion
Xand conditionally defined code constructs. The next section describes
Xthe \fBEase\fR language in complete detail.
X.NH
XEase Syntax*
X.FS
X* \fINo attempt is made to describe the complete semantic meaning
Xassociated with all of the constructs of a sendmail configuration file. Items
Xnot covered in this document include the semantic distinction among rulesets,
Xthe special uses of
Xpre-defined macros, and the method of building configuration files. To
Xobtain this information, the reader is advised to refer to
Xthe Installation and Operation Guide for Sendmail (UNIX
XProgrammer's Manual, Volume 2c), by Eric Allman.\fR
X.FE
X.PP
XAt its highest level, \fBEase\fR can be viewed as a collection of
Xblock-structures, where each block begins with a keyword and is followed by
Xzero or more related definitions and/or declarations. There are ten distinct
Xblock types. The following is
Xa list containing all ten block keywords and the block type it denotes.
X.TS
Xcenter;
Xl l .
X\fIbind\fR -ruleset identifier bindings
X\fImacro\fR -macro definitions
X\fIclass\fR -class definitions
X\fIoptions\fR -\fIsendmail\fR option definitions
X\fIprecedence\fR -precedence definitions
X\fItrusted\fR -trusted users
X\fIheader\fR -mail header definitions
X\fImailer\fR -mailer definitions
X\fIfield\fR -address field definitions
X\fIruleset\fR -address rewriting rules
X.TE
X.sp 1
XIn general,
X.TS
Xcenter ;
Xl .
X
X* Letters are distinguished by case,
X
XT{
X* An \fBEase\fR identifier is defined to be a letter, followed by zero or
Xmore letters, digits, underscores (_), or dashes (-),
XT}
X
XT{
X* A literal newline or double quotation (") character may be included in
Xany quoted string by preceding the character with a backslash (\\\\\), and
XT}
X
XT{
X* \fBEase\fR source is preprocessed by the C language preprocessor (cpp),
Xthus source comments (i.e., text enclosed by \*Q/*\*U and \*Q*/\*U) may appear
Xanywhere as part of \fBEase\fR whitespace.
XT}
X.TE
X.PP
XFor notational convenience, this document specifies all reserved
Xwords of the \fBEase\fR language in italics. In addition, quantities
Xenclosed in angle brackets (<..>) represent arbitrary
Xidentifiers, strings, or numbers.
X.NH 2
XRuleset Identifier Bindings
X.PP
XA ruleset (a set of rewriting rules) is identified solely by an integer
Xin \fIsendmail\fR's
Xconfiguration file. \fBEase\fR, however, allows each ruleset to be named with
Xa meaningful identifier. Since a special numeric association for each
Xruleset is required by the address parsing scheme of \fIsendmail\fR, a \fIbind\fR
Xblock must be present in any \fBEase\fR file which defines one or more
Xrulesets. A
X\fIbind\fR block consists of the keyword \fIbind\fR, followed by zero or more
Xstatements of the form:
X.TS
Xcenter box;
Xl .
X<ruleset-id> = \fIruleset\fR <ruleset-number> ;
X.TE
XThe following example,
X.sp 1
X\fIbind\fR
X.PP
XFINAL_RW = \fIruleset\fR 4;
X.sp 1
Xspecifies that FINAL_RW, the final rewriting ruleset, is \fIsendmail\fR's ruleset
Xnumber 4.
X.NH 2
XMacro Definitions
X.PP
XA macro is an identifier which, when referenced in the text of a program,
Xis replaced by its value, a string of zero or more characters. The value
Xof a macro may include references to other macros, but not itself! \fISendmail\fR
Xallows a maximum of 26 user-declared macros in its configuration file. In
Xaddition, there are a number of pre-declared macros which have special meaning
Xto \fIsendmail\fR (see Appendix A). \fBEase\fR macros are defined in
X\fImacro\fR blocks. \fBEase\fR allows any macro to be declared
X(which is equivalent to simply referencing it) before it is defined. A macro
Xidentifier is replaced by its value when it is preceded by the character
X\*Q$\*U. In addition, a macro reference inside a quoted string must always
Xinclude braces ({}) around the macro identifier (for delimiting purposes).
X.PP
XA \fImacro\fR block consists of the keyword \fImacro\fR, followed by zero
Xor more statements taking either of the following forms:
X.TS
Xcenter box;
Xl .
X<macro-identifier> = "<macro-value>" ;
X<macro-identifier> = \fBconditional-expression\fR ;
X.TE
XThe \fBconditional-expression\fR format will be discussed
Xlater.
X.sp 1
XThe following example,
X.sp 1
X\fImacro\fR
X.PP
Xfirst_name = "James";
X.PP
Xlast_name = "Schoner";
X.PP
Xwhole_name = "${first_name} ${second_name}";
X.sp 1
Xdefines the macros first_name, last_name, and whole_name, where whole_name
Xis the string, "James Schoner".
X.NH 2
XClass definitions
X.PP
XA class is denoted by an identifier representing a logical grouping of zero
Xor more names. Classes are used to represent the range of values a token
Xmay assume in the pattern matching of an address. Further discussion on the
Xuse of classes will be deferred until address fields are described.
X.PP
XOne identifier may be used to distinctly represent both a macro
Xand class (i.e., the set of macro identifiers and the set of class identifiers
Xmay form a non-empty intersection). A name, or class element, may
Xbe an identifier or any quoted word.
X.PP
XA \fIclass\fR block consists of the keyword \fIclass\fR, followed by zero
Xor more statements taking any of the following forms:
X.TS
Xcenter box;
Xl .
X<class-identifier> = { <name1>, <name2>, <name3>, ... } ;
X<class-identifier> = \fIreadclass\fR ( "<file-name>" ) ;
X<class-identifier> = \fIreadclass\fR ( "<file-name>", "<read-format>" ) ;
X.TE
XThe second and third forms cause \fIsendmail\fR to read the names of the class
Xfrom the named
Xfile. The third form contains a read format, which should be a \fIscanf(3)\fR
Xpattern yielding a single string.
X.sp 1
XThe following example,
X.sp 1
X\fIclass\fR
X.PP
Xcampus_hosts = { statistics, engineering, chemistry, physics, physics-2 } ;
X.PP
Xversions = { "1.0", "1.1", "4.0", "4.2", latest-and-greatest } ;
X.PP
Xphone_hosts = \fIreadclass\fR ( "/tmp/phonenet.list" ) ;
X.sp 1
Xdefines the classes campus_hosts, versions, and phone_hosts.
X.NH 2
XSendmail option definitions
X.PP
XA number of options to the \fIsendmail\fR program may be specified in
Xan \fIoptions\fR
Xblock. For a description of the various \fIsendmail\fR options and their
Xvalues, see Appendix B.
X.PP
XAn
X\fIoptions\fR block consists of the keyword \fIoptions\fR, followed by zero
Xor more statements taking any of the following forms:
X.TS
Xcenter box;
Xl l .
X<option-identifier> = "<option-value>" ;
X\fIo_delivery\fR = \fBspecial-value\fR ;
X\fIo_handling\fR = \fBspecial-value\fR ;
X.TE
XAll but two options (\fIo_delivery\fR and \fIo_handling\fR) use the first
Xform. To specify an option without a value, simply assign to it the null
Xstring (""). The \fBspecial-value\fR field of the second and third form
Xrefers to special values (non-quoted) which are specified in Appendix B.
X.sp 1
XThe following example,
X.sp 1
X\fIoptions\fR
X.PP
X\fIo_alias\fR = "/usr/lib/aliases" ;
X.PP
X\fIo_tmode\fR = "0600" ;
X.PP
X\fIo_delivery\fR = d_background ;
X.sp 1
Xsets the options \fIo_alias\fR, \fIo_tmode\fR, and \fIo_delivery\fR.
X.NH 2
XPrecedence definitions
X.PP
XMessage headers may contain a \*QPrecedence:\*U field describing the precedence
Xof the message class. Identifiers which may appear in the precedence field of
Xa message are given precedence values in a configuration file \fIprecedence\fR
Xdefinition. This association will be illustrated below in an example.
X.PP
XA \fIprecedence\fR block consists of the keyword \fIprecedence\fR, followed
Xby zero or more statements of the form:
X.KS
X.TS
Xcenter box;
Xl .
X<precedence-identifier> = <precedence-integer> ;
X.TE
X.KE
XThe following example,
X.sp 1
X\fIprecedence\fR
X.PP
Xspecial-delivery = 100;
X.PP
Xjunk = -100;
X.sp 1
Xdefines the precedence level for the names \*Qspecial-delivery\*U and
X\*Qjunk\*U. Thus, whenever the name \*Qjunk\*U appears in
Xa \*QPrecedence:\*U field, the corresponding message class will be set to -100.
X.NH 2
XTrusted users
X.PP
X\fISendmail\fR's \fB-f\fR flag allows trusted users to override the sender's
Xmachine address. Trusted users are listed in \fItrusted\fR blocks. A
X\fItrusted\fR block consists of the keyword \fItrusted\fR, followed
Xby zero or more sets of users taking the form:
X.TS
Xcenter box;
Xl .
X{ <user1>, <user2>, <user3>, ... } ;
X.TE
XThe following example,
X.sp 1
X\fItrusted\fR
X.PP
X{ root, uucp, network } ;
X.PP
X{ acu, kcs, jss } ;
X.sp 1
Xspecifies that the users root, uucp, network, acu, kcs, and jss can be trusted
Xto use the \fIsendmail\fR flag, \fB-f\fR.
X.NH 2
XMail header definitions
X.PP
XThe format of the message headers inserted by \fIsendmail\fR is defined in one
Xor more \fIheader\fR blocks in the configuration file. A \fIheader\fR block
Xconsists of the keyword \fIheader\fR, followed by zero or more statements
Xtaking any of the following forms:
X.TS
Xcenter box;
Xl
Xl
Xl
Xl
Xl
Xl
Xl
Xl
Xl
Xl
Xl .
X\fIfor\fR ( <mailer-flag1>, <mailer-flag2>, ... )
X \fIdefine\fR ( "<header-title>" , \fBheader-value\fR ) ;
X
X\fIfor\fR ( <mailer-flag1>, <mailer-flag2>, ... ) {
X \fIdefine\fR ( "<header-title>" , \fBheader-value\fR ) ;
X \fIdefine\fR ( "<header-title>" , \fBheader-value\fR ) ;
X .
X .
X} ;
X
X\fIdefine\fR ( "<header-title>" , \fBheader-value\fR ) ;
X.TE
XThe first form is used to define one header for one or more mailer
Xflags. The second form differs from the first in that more than one
Xheader may be defined for a given set of flags. The third form is used to
Xdefine a header,
Xregardless of mailer flags. Refer to Appendix C for a list of \fBEase\fR
Xidentifiers representing mailer flags. The header title is a simple
Xstring of characters (no macro references), whereas the \fBheader-value\fR can
Xbe either a string of characters (possibly containing macro references) or
Xa \fBconditional-expression\fR (discussed later).
X.sp 1
XThe following example,
X.DS
X\fIheader\fR
X
X \fIdefine\fR ( "Subject:", "") ;
X
X \fIfor\fR ( \fIf_return\fR )
X \fIdefine\fR ( "Return-Path:", "<${\fIm_sreladdr\fR}>" ) ;
X
X \fIfor\fR ( \fIf_date\fR ) {
X \fIdefine\fR ( "Resent-Date:", "${\fIm_odate\fR}" ) ;
X \fIdefine\fR ( "Date:", "${\fIm_odate\fR}" );
X } ;
X.DE
Xdefines a \*QSubject\*U field for all mailers, regardless of their flags, a
X\*QReturn-Path\*U field for mailers whose definition specifies
Xthe flag, \fIf_return\fR, and the headers, \*QResent-Date\*U and \*QDate\*U,
Xfor mailers whose definition specifies the flag, \fIf_date\fR.
X.NH 2
XMailer Definitions
X.PP
X\fISendmail\fR's definition of a mailer (or an interface to one) occurs in a
X\fImailer\fR block. A \fImailer\fR block consists of the keyword \fImailer\fR,
Xfollowed by zero or more statements of the form:
X.TS
Xcenter box;
Xl .
X<mailer-identifier> { \fBmailer-spec\fR } ;
X.TE
XThe field, \fBmailer-spec\fR, is a list of zero or more of the
Xfollowing attribute assignments (where successive assignment statements are
Xseparated by commas):
X.TS
Xcenter ;
Xl l
Xl l
Xl l
Xl l
Xl l
Xl l
Xl l .
X\fIPath\fR = \fBstring-attribute\fR
X\fIArgv\fR = \fBstring-attribute\fR
X\fIEol\fR = \fBstring-attribute\fR
X\fIMaxsize\fR = \fBstring-attribute\fR
X\fIFlags\fR = { <mailer-flag1>, <mailer-flag2>, ... }
X\fISender\fR = <sender-ruleset-id>
X\fIRecipient\fR = <recipient-ruleset-id>
X.TE
XThe \fBstring-attribute\fR value can take the form of a quoted string
X(possibly containing macro references) or a \fBconditional-expression\fR
X(discussed later).
X.sp 1
XThe following example,
X.sp 1
X\fImailer\fR
X.DS
X local {
X \fIPath\fR = "/bin/mail",
X \fIFlags\fR = { \fIf_from\fR, \fIf_locm\fR },
X \fISender\fR = Sender_RW,
X \fIRecipient\fR = Recip_RW,
X \fIArgv\fR = "mail -d ${\fIm_ruser\fR}",
X \fIMaxsize\fR = "200000"
X } ;
X.DE
Xdefines a mailer named \*Qlocal\*U.
X.NH 2
XAddress field definitions
X.PP
X\fISendmail\fR's address parsing scheme treats an address as a group of tokens
X(an address token is precisely defined in the Arpanet protocol RFC822). In
Xgeneral, \fIsendmail\fR divides an address into tokens based on a list of
Xcharacters assigned as a string to the special macro \fIm_addrops\fR. These
Xcharacters will individually be considered as tokens and will separate tokens
Xwhen parsing is performed.
X.PP
XFor
Xthe \fBEase\fR language, there is a distinct set of address tokens (defined
Xbelow) which are used in combination to represent generic forms of
Xaddresses. In
Xaddition to literal address tokens, the pattern to be matched in a rewriting
Xrule (often refered to as the LHS) may
Xinclude field identifiers which match one of five possibilities:
X.DS
X - zero or more tokens
X - one or more tokens
X - one token exactly
X - one token which is an element of an aribitrary class \fBX\fR
X - one token which is not an element of an aribitrary class \fBX\fR
X.DE
XA particular field type may be assigned to one or more identifiers. Each
Xfield identifier is associated with (or defined to be) a field type in
Xa \fIfield\fR declarations block. A \fIfield\fR declarations block consists
Xof the keyword \fIfield\fR, followed by zero or more field definitions of
Xthe form:
X.TS
Xcenter box;
Xl .
X\fBfield-id-list\fR : \fBfield-type\fR ;
X.TE
XA \fBfield-id-list\fR is a list of one or more identifiers, each separated by
Xa comma. A \fBfield-type\fR, on the other hand, is a representation of
Xone of the five fields
Xdescribed above. The syntax for each of the five forms follows:
X.DS
X \fImatch\fR ( 0* )
X \fImatch\fR ( 1* )
X \fImatch\fR ( 1 )
X \fImatch\fR ( 1 ) in <class-X>
X \fImatch\fR ( 0 ) in <class-X>
X.DE
XThe star in the first two forms means: "or more". Thus, the first
Xform would read: "match zero or more tokens". The fourth form describes
Xa field where one token is matched from an arbitrary class (class-X), whereas
Xthe fifth form describes a field where one token is matched if it is not of the
Xgiven class (class-X).
X.sp 1
XThe following example,
X.sp 1
X.DS
X\fIfield\fR
X anypath : \fImatch\fR ( 0* );
X recipient_host : \fImatch\fR ( 1 );
X local_site : \fImatch\fR ( 1 ) in \fIm_sitename\fR;
X remote_site : \fImatch\fR ( 0 ) in \fIm_sitename\fR;
X.DE
Xdefines the fields anypath, recipient_host, local_site, and remote_site.
X.NH 2
XAddress rewriting rules
X.PP
XAddress rewriting rules are grouped according to the function they perform. For
Xexample, it is desirable to form a distinct group for those rewriting rules
Xwhich perform transformations on recipient addresses.
X.PP
XSets of rewriting rules are defined in \fIruleset\fR blocks. A \fIruleset\fR
Xblock consists of the keyword \fIruleset\fR, followed by zero or more
Xruleset definitions of the form:
X.TS
Xcenter box;
Xl .
X<ruleset-id> { <rewriting-rule1> <rewriting-rule2> ... }
X.TE
XThe ruleset identifier, ruleset-id, must be defined in a \fIbind\fR block, as
Xdescribed earlier. The rewriting rules have the form:
X.DS
X \fIif\fR ( <match-pattern> )
X <match-action> ( <rewriting-pattern> ) ;
X.DE
Xwhere match-pattern, rewriting-pattern, and match-action are described below.
X.NH 3
XMatch-patterns
X.PP
XA match-pattern is a sequence of Ease address elements representing an
Xaddress format. If the address being rewritten matches the pattern
X\*Qmatch-pattern\*U,
Xthen the address is reformatted using the pattern \*Qrewriting-pattern\*U, and
Xthe corresponding
Xaction (\*Qmatch-action\*U) is performed. The five distinct Ease address
Xelements which may constitute a match-pattern are as follows:
X.TS
Xcenter ;
Xl .
X1. Field Identifiers (refer to previous section)
XT{
X2. Non-alphanumeric characters (the exception is the case for literal
Xdouble quotes, which must be preceded by a backslash (\\\\\)
XT}
X3. Macro references
X4. Quoted strings ("...")
X5. \fBConditional-expressions\fR (discussed later)
X.TE
XBelow are two sample match-patterns, each describing the same address format:
X.DS
X user-id @ hostname . $arpa_suffix
X user-id @ hostname ".ARPA"
X.DE
Xwhere user-id and hostname are field identifiers, and arpa_suffix is a
Xuser-defined macro with the value \*QARPA\*U.
X.NH 3
XRewriting-patterns
X.PP
XA rewriting-pattern specifies the form in which to rewrite a matched
Xaddress. The seven distinct elements which may be used to form
Xa rewriting-pattern are as follows:
X.TS
Xcenter ;
Xl .
X
XT{
X1. Non-alphanumeric characters (the exception is the case for literal
Xdouble quotes, left parentheses, or right parentheses, each of which
Xmust be preceded by a backslash (\\\\\).
XT}
X
XT{
X2. A call to another ruleset. This is used to perform rewrites
Xon a suffix of the rewriting-pattern. The proper use of this
Xfeature will be demonstrated by example below.
XT}
X
X3. Quoted strings ("...").
X
X4. \fBConditional-expressions\fR (discussed later).
X
X5. A macro reference.
X
XT{
X6. A positional reference in the matched address. A positional
Xreference takes the form: $<integer-position>. For example,
X$3 references the value of the third \fBEase\fR address
Xelement in the matched address.
XT}
X
XT{
X7. Canonicalized host names of the form \fIcanon\fR (<id-token>),
Xwhere id-token is a regular identifier, a quoted identifier (with
Xdouble quotes), a macro reference yielding an identifier, or a
Xpositional reference in the matched address. The canonicalization of
Xa host name is simply a mapping to its canonical (or official) form.
XT}
X
X.TE
XBelow are two sample rewriting-patterns:
X.DS
X $1 % $2 < @ $3 ".ARPA" >
X OLDSTYLE_RW ( $1 )
X.DE
XThe first form specifies an address such as a%b<@c.ARPA>, where a, b, and c
Xrepresent matched identifiers or paths. The second form specifies a call to
Xthe ruleset \*QOLDSTYLE_RW\*U, for old-style rewriting on the parameter
X$1, which probably references the entire matched address. This will become
Xclear in later examples.
X.NH 3
XMatch-actions
X.PP
XWhen a ruleset is called, the address to be rewritten is compared (or matched)
Xsequentially against the match-address of each rewriting rule. When a
Xmatch-address describes the address \fIsendmail\fR is attempting to rewrite, the
Xaddress is rewritten (or reformatted) using the rule's
Xrewriting-pattern. Following this rewrite, the corresponding match-action
Xis performed. There are four match-actions:
X.TS
Xcenter ;
Xl l .
X\fIretry\fR T{
X-a standard action which causes the rewritten address
Xto be again compared to the match-address of the current rule.
XT}
X
X\fInext\fR T{
X-an action which causes the rewritten address to be
Xcompared to the match-address of the next rewriting rule of the current
Xruleset. If the end of the list is reached, the ruleset returns the
Xrewritten address.
XT}
X
X\fIreturn\fR T{
X-an action which causes an immediate return of the
Xruleset with the current rewritten address.
XT}
X
X\fIresolve\fR T{
X-an action which specifies that the address has been
Xcompletely resolved (i.e., no further rewriting is necessary). The
X\fIresolve\fR action is described in more detail below.
XT}
X.TE
X.PP
XThe match-action, \fIresolve\fR, is special in that it terminates
Xthe address rewriting altogether. The semantic structure of \fIsendmail\fR's
Xrewriting scheme requires that a \fIresolve\fR action appear only in the
Xruleset whose numerical binding is to the number zero. The \fIresolve\fR action
Xmust specify three parameters: \fImailer\fR, \fIhost\fR, and \fIuser\fR. If
Xthe \fImailer\fR is local, the \fIhost\fR parameter may be omitted. The
X\fImailer\fR argument must be specified as a single word, macro, or positional
Xreference in the matched address. The \fIhost\fR argument may be specified as
Xa single word or as an expression which expands to a single word (i.e.,
X\fIhost\fR ($1 ".ARPA")). In addition, the \fIhost\fR argument may be a
Xcanonicalization (as described above) or a numeric internet specification. The
Xkeyword \fIhostnum\fR is used for numeric internet specifications, as in
X\fIhostnum\fR ("00.00.00.00") or \fIhostnum\fR ( $2 ). The \fIuser\fR
Xspecification is a rewriting-pattern, as described above.
X.PP
XIn general, the format of a \fIresolve\fR action will be as follows:
X.DS
X \fIresolve\fR ( \fImailer\fR ( <mailer-name> ),
X \fIhost\fR ( <host-name> ),
X \fIuser\fR ( <user-address>) );
X.DE
XExamples of the match-action statement are shown below:
X.DS
X\fIfield\fR
X anypath : \fImatch\fR (0*);
X usr, path : \fImatch\fR (1*);
X hostname : \fImatch\fR (1);
X phone_host : \fImatch\fR (1) in phonehosts;
X.DE
X.DS
X\fIruleset\fR
X
X EXAMPLE_RW {
X
X \fIif\fR ( anypath < path > anypath ) /* basic RFC821/822 parse */
X \fIretry\fR ( $2 );
X \fIif\fR ( usr " at " path ) /* \*Qat\*U -> \*Q@\*U */
X \fInext\fR ( $1 @ $2 );
X \fIif\fR ( @path: usr )
X \fIreturn\fR ( LOCAL_RW ( < @$1 > : $2 ) );
X \fIif\fR ( anypath < @phone_host".ARPA" > anypath )
X \fIresolve\fR ( \fImailer\fR ( tcp ),
X \fIhost\fR ( csnet-relay ),
X \fIuser\fR ( $1 % $2 < @"csnet-relay" > $3 ) );
X }
X.DE
X.PP
XThe example above defines the ruleset \*QEXAMPLE_RW\*U, which contains four
Xrewriting rules. The first rewriting rule discards all tokens of an address
Xwhich lie on either side of a pair of angle brackets (<>), thereby
Xrewriting the address as
Xthe sequence of tokens contained within the angle brackets ($2). Following the
Xaddress rewrite, the rule is applied again (\fIretry\fR). When the first rule
Xfails to match the address being rewritten, the second rule is applied.
X.PP
XThe second
Xrule simply replaces the word \*Qat\*U by the symbol \*Q@\*U. The \*Q\fInext\fR\*U
Xaction specifies that if a match is made, a rewrite is performed and
Xmatching continues at the next (or following) rule.
X.PP
XThe third rule illustrates
Xthe use of the \*Q\fIreturn\fR\*U action, which is executed if the
Xpattern \*Q@path: usr\*U
Xdescribes the current format of the address being rewritten. In this example,
Xthe \fIreturn\fR action returns the result of a call to ruleset \*QLOCAL_RW\*U,
Xwhich rewrites the address \*Q<@$1>:$2\*U, where $1 and $2 are substituted
Xwith the token(s) matched respectively by \*Qpath\*U and \*Qusr\*U.
X.PP
XThe fourth (and final) rule signals a resolution (and termination) of the
Xrewriting process if the given pattern is matched. The resolution specifies
Xthat the mailer \*Qtcp\*U will be used to deliver the message to the host
X\*Qcsnet-relay\*U. The \fIuser\fR parameter specifies the final form of the address
Xwhich \fIsendmail\fR has just resolved.
X.sp 2
X.PP
XThe \fBEase\fR construct which remains to be examined is the
X\fBconditional-expression\fR. The \fBconditional-expression\fR provides a
Xmethod for
Xconstructing strings based on the condition that some test macro is (or is not)
Xset. The general form begins with the concatenation of a string and a
X\fBstring-conditional\fR:
X.DS
X \fIconcat\fR ( <quoted-string>, \fBstring-conditional\fR )
X \fIconcat\fR ( \fBstring-conditional\fR, <quoted-string> )
X.DE
XA \fBstring-conditional\fR assumes either of the following forms:
X.DS
X \fIifset\fR ( <macro-name>, <ifset-string> )
X \fIifset\fR ( <macro-name>, <ifset-string>, <notset-string> )
X.DE
XA \fBstring-conditional\fR of the first form evaluates to \*Qifset-string\*U
Xif the macro \*Qmacro-name\*U has been assigned a value; otherwise it
Xevaluates to the null string. The second form behaves similarly, except
Xthat the \fBstring-conditional\fR evaluates to \*Qnotset-string\*U, instead
Xof the null string, if the macro \*Qmacro-name\*U has no value.
X.sp 1
XThe following \fBconditional-expression\fR,
X.DS
X \fIconcat\fR ( "New ", \fIifset\fR ( city, "York", "Jersey" ) )
X.DE
Xevaluates to the string "New York", if the macro \*Qcity\*U is set. Otherwise,
Xthe \fBconditional-expression\fR evaluates to the string "New Jersey".
X.NH
XEase Translation
X.PP
XIt is important to note that \fBEase\fR is translated by a stand-alone
Xtranslator to the raw configuration file format. No modifications were
Xmade to the \fIsendmail\fR program itself. As a result, syntactical verification
Xof a configuration file can be performed without invoking \fIsendmail\fR.
X.PP
XThe \fBEase\fR language is translated by invoking
Xthe C language preprocessor (cpp) with \fBEase\fR source as input, then piping
Xthe output as input to the \fBEase\fR translator (\fIet\fR). The \fBEase\fR
Xtranslator may be invoked on the command line in one of four ways:
X.TS
Xcenter box ;
Xl l .
X\fIet\fR <input-file> <output-file> [read from a file, write to a file]
X\fIet\fR <input-file> [read from a file, write to standard output]
X\fIet\fR - <output-file> [read from standard input, write to a file]
X\fIet\fR [read from standard input, write to standard output]
X.TE
X.NH
XConclusion
X.PP
X\fBEase\fR is currently in use at the Purdue University Computing
XCenter. Source code for the \fBEase\fR translator (\fIet\fR) may be
Xobtained on request by writing to:
X.DS
XU.S. Mail:
X James S. Schoner
X c/o Kevin S. Braunsdorf
X Purdue University Computing Center
X Purdue University
X West Lafayette, Indiana 47907
X
XElectronic Mail:
X ksb@j.cc.purdue.edu
X.DE
X.PP
XMuch of the success of this project is attributable to the constant support
Xand insight offered by Mark Shoemaker. To him, I owe a debt of gratitude. In
Xaddition, I would like to thank Kevin Smallwood, Paul Albitz, and Rich Kulawiec
Xfor their many notable suggestions and valuable insight.
END_OF_FILE
if test 29063 -ne `wc -c <'doc/ease.paper'`; then
echo shar: \"'doc/ease.paper'\" unpacked with wrong size!
fi
# end of 'doc/ease.paper'
fi
if test -f 'maketd/maketd.c' -a "${1}" != "-c" ; then
echo shar: Will not clobber existing file \"'maketd/maketd.c'\"
else
echo shar: Extracting \"'maketd/maketd.c'\" \(16381 characters\)
sed "s/^X//" >'maketd/maketd.c' <<'END_OF_FILE'
X/* maketd - MAKE Transitive Dependencies.
X * (This is a lie - the dependencies are not transitive, but "all"
X * dependencies are correctly made.)
X *
X * Based loosely on a shell script by Stephan Bechtolsheim, svb@purdue
X * Other Makefile related features have been added or merged in from
X * other programs.
X *
X * Written & hacked by Stephen Uitti, PUCC staff, ach@pucc-j, 1985
X * maketd is copyright (C) Purdue University, 1985
X *
X * removed some of Steve's good, but unnecessary, options in favor
X * of more compile time flags & better implicit rules in the makefile
X * dinked: -q -e -E -k
X * Kevin S Braunsdorf, PUCC UNIX Group 1986 (ksb@j.cc.purdue.edu)
X *
X * Permission is hereby given for its free reproduction and
X * modification for non-commercial purposes, provided that this
X * notice and all embedded copyright notices be retained.
X * Commercial organisations may give away copies as part of their
X * systems provided that they do so without charge, and that they
X * acknowledge the source of the software.
X */
X#include <sys/types.h>
X#include <sys/param.h>
X#include <sys/file.h> /* for access */
X#ifdef BSD2_9
X#include <ndir.h> /* for MAXPATHLEN, MAXNAMLEN */
X#endif
X#ifdef BSD4_2
X#include <sys/dir.h> /* for MAXNAMLEN in 4.2 */
X#endif
X#include <ctype.h> /* for isupper */
X#include <stdio.h>
Xextern char *rindex(), *index(), *strcat(), *strcpy();
X
X#include "srtunq.h"
X#include "abrv.h"
X#include "nshpopen.h"
X#include "maketd.h"
X
X#ifndef CPP
X#define CPP "/lib/cpp "
X#endif CPP not in Makefile
X
X/* forward functions */
Xvoid msoio(); /* open output file */
Xvoid rdwr(); /* read old Makefile into new */
Xvoid mkdepend(); /* does the real work */
X
X/* globals */
Xchar *prgnm; /* our program name */
XFILE *makefd; /* makefile stream */
Xint alldep = FALSE; /* -a all - /usr/include too */
Xchar *targetname = NULL; /* -t target name for next file */
Xchar *destsuffix = ".o"; /* -s suffix for targets */
Xint header = TRUE; /* print header & trailer */
Xint usestdout = FALSE; /* -d use stdout for makefile */
Xint forcehead = FALSE; /* -f force header/trailer */
Xint makenseen = FALSE; /* output file has been specified */
Xchar *makename = "makefile"; /* -m default file for edit */
Xint backedup = FALSE; /* for interupt recovery */
Xchar backupfn[MAXNAMLEN+1]; /* backup file name */
Xint ismakeopen = FALSE; /* if the output file is open */
Xchar objpath[MAXPATHLEN+1]; /* -o prepended to .o's */
Xint nonlocalo = FALSE; /* -nonlocalo objects in source dir */
Xint replace = FALSE; /* -r replace depends in Makefile */
Xchar cppflags[BUFSIZ]; /* -D, -I, -U flags to pass to cpp */
Xint shortincl = TRUE; /* -x do abreviations */
Xint verbose = FALSE; /* -v verbage for the debugger */
Xstatic char sopts[] = "abdfhrxv"; /* single char opts */
Xstatic SRTUNQ u; /* unique include files */
X
Xchar usage[] =
X"Usage: maketd [-a -b -d -f -h -mMAKEFILE -nonlocalo -oDIR -r -sSUFFIX\n\
X -tTARGETNAME -x -v -Iincludedir -Ddefine -Uundefine file...]\n";
Xchar helptext[] =
X"-a\tdo all dependencies, including /usr/include\n\
X-b\tgenerate binary, rather than object related dependencies\n\
X-d\tdependencies to stdout, rather than Makefile\n\
X-f\tforce header/trailer (use with -d)\n\
X-h\thelp (this text)\n\
X-m\tspecify MAKEFILE for edit\n\
X-nonlocalo Objects live in source directory\n\
X-o\tprepend DIR to target: DIR/a.o: foo.h\n\
X-r\treplace dependencies for a target\n\
X-s\tchange suffix target's SUFFIX: a.SUFFIX: foo.h\n\
X-t\tchange target's basename: TARGET.o: foo.h\n\
X-x\tdon't abbreviate includes\n\
X-v\tprint extra verbose (debug) output to stderr\n\
X-I\tspecify include directory, as in /lib/cpp\n\
X-D\tspecify defines, as in /lib/cpp\n\
X-U\tspecify undefines, as in /lib/cpp\n";
X
Xchar deplin[] = "# DO NOT DELETE THIS LINE - make depend DEPENDS ON IT\n";
Xchar searchdep[] = "# DO NOT DELETE THIS LINE";
Xchar trailer[] = "\n# *** Do not add anything here - It will go away. ***\n";
X
X/* some init & argv parsing */
Xmain(argc, argv)
Xregister int argc;
Xregister char **argv;
X{
X register a; /* argv subscript */
X register i; /* tmp */
X register len; /* length of current argument */
X register files = FALSE; /* files ever seen */
X register char *q; /* tmp */
X
X /* prgnm = program name, for error printing */
X if ((prgnm = rindex(argv[0], '/')) == NULL)
X prgnm = argv[0];
X else
X prgnm++;
X catchsig(); /* init signal traps */
X srtinit(&abrv); /* init abbreviations tree */
X for (a = 1; a < argc; a++) { /* argv prepass: find all -I's */
X if (argv[a][0] == '-' && argv[a][1] == 'I' && strlen(&argv[a][2]) > 2)
X if ((q = srtin(&abrv, hincl(&argv[a][2]), lngsrt)) != NULL)
X fprintf(stderr, "%s: %s - %s\n", prgnm, q, &argv[a][2]);
X }
X cppflags[0] = '\0'; /* terminate cpp flags string */
X objpath[0] = '\0'; /* init object path */
X srtinit(&u); /* init sorting database tag */
X for (a = 1; a < argc; a++) {
X len = strlen(argv[a]);
X if (argv[a][0] == '-' && len > 2 && index(sopts, argv[a][1]) != NULL)
X err("options must be listed seperately - '%s'\n", argv[a]);
X if (len > 1 && argv[a][0] == '-') {
X switch (argv[a][1]) {
X case 'D': /* /lib/cpp flags to pass */
X case 'I':
X case 'U':
X if (strlen(cppflags) + strlen(argv[a]) + 2 > BUFSIZ)
X err("too many cpp flags - buffer overflow");
X strcat(cppflags, argv[a]);
X strcat(cppflags, " "); /* add a space separator */
X break;
X case 'a': /* /usr/include deps too */
X alldep = TRUE;
X break;
X case 'b': /* target has no suffix */
X destsuffix = "";
X break;
X case 'd': /* don't edit Makefile */
X if (ismakeopen)
X err("Makefile already open, -d must precede filenames");
X if (makenseen)
X err("Conflict - check -d and -m options");
X makenseen = TRUE;
X usestdout = TRUE;
X if (!forcehead)
X header = FALSE; /* don't do header & trailer */
X break;
X case 'f': /* force header/trailer */
X forcehead = TRUE;
X header = TRUE;
X break;
X case 'h': /* help */
X fputs(usage, stdout);
X fputs(helptext, stdout);
X exit(0);
X break;
X case 'm': /* specify makefile name for edit */
X if (ismakeopen)
X err("Makefile already open, -m must precede filenames");
X if (makenseen)
X err("Conflict, check -m and -d options.");
X if (strlen(makename) == 0)
X err("-m option requires file name.");
X makenseen = TRUE;
X makename = &argv[a][2];
X break;
X case 'n': /* objects reside with sources */
X if (strcmp("nonlocalo", &argv[a][1]) != 0)
X err("bad -n option"); /* what a crock of an option */
X if (objpath[0] != '\0')
X err("nonlocalo conflict - check -o's");
X nonlocalo = TRUE;
X break;
X case 'o':
X if (nonlocalo)
X err("object path conflict - check -o's and -nonlocalo's");
X strcpy(objpath, &argv[a][2]);
X i = strlen(objpath);
X if (i == 0)
X err("-o requires path string.");
X if (i >= MAXPATHLEN)
X err("Object path too long: max is %d.", MAXPATHLEN);
X if (objpath[i - 1] != '/')
X strcat(objpath, "/");
X break;
X case 'r': /* replace mode */
X if (ismakeopen)
X err("Makefile already open, -r must precede filenames");
X replace = TRUE;
X break;
X case 's': /* destination suffix */
X destsuffix = &argv[a][2];
X break;
X case 't': /* set target's basename */
X targetname = &argv[a][2];
X if (len <= 2)
X err("target option requires name.");
X break;
X case 'v': /* user wants to hear noise */
X verbose = TRUE;
X break;
X case 'x': /* don't abbrev. */
X if (files)
X err("-x option must preceed all files.");
X shortincl = FALSE;
X break;
X default:
X err("Unknown option %s.", argv[a]);
X } /* end switch */
X } else { /* must be a filename */
X if (verbose)
X fprintf(stderr, "%s: working on %s.\n", prgnm, argv[a]);
X files = TRUE; /* at least one file seen */
X if (replace && a != argc - 1)
X err("Only one file allowed with -r (edit aborted)");
X mkdepend(argv[a]); /* file to process */
X targetname = NULL; /* affect only one file */
X } /* if option */
X } /* for argv */
X if (ismakeopen && header)
X fputs(trailer, makefd); /* do not delete... */
X#if DEL_BACKUP
X if (backedup)
X if (unlink(backupfn))
X err("Can't delete backup file %s on completion", backupfn);
X#endif DEL_BACKUP
X if (!files)
X err("No files to process, use -h for full help.\n%s", usage);
X exit(0); /* exit status - good */
X}
X
X/* msoio - Make Sure Output Is Open.
X * Interacts strongly via globals: makefd, backedup, backupfn, makename,
X * header, ismakeopen
X */
Xvoid
Xmsoio(targ)
Xchar *targ; /* if -r, is target name */
X{
X FILE *tmpfd; /* temp file desc for -d */
X char buf[BUFSIZ]; /* for reading the makefile */
X
X if (ismakeopen)
X return;
X ismakeopen = TRUE; /* will be: all errs are fatal */
X if (usestdout) {
X makefd = stdout;
X if (header) {
X fputc('\n', makefd); /* one blank line */
X fputs(deplin, makefd); /* the first line */
X }
X /* scan "makefile" or "Makefile" for include defines */
X if (access(makename, R_OK) != 0) {
X makename[0] = 'M'; /* try Makefile */
X if (access(makename, R_OK) != 0)
X return; /* just punt */
X }
X if ((tmpfd = fopen(makename, "r")) == NULL)
X return; /* just punt */
X while (fgets(buf, BUFSIZ, tmpfd) != NULL)
X srchincl(buf); /* scan whole file */
X fclose(tmpfd);
X return;
X } /* ... if standard out */
X/* !makenseen means (default) try "makefile" then "Makefile" */
X if (!makenseen && access(makename, F_OK) != 0)
X makename[0] = 'M'; /* try Makefile */
X/* side effect: "Makefile" will be created if neither exist */
X if (access(makename, F_OK) == 0) { /* if makefile exits */
X strcpy(backupfn, makename); /* get rid of .bak */
X strcat(backupfn, ".bak");
X if (access(backupfn, F_OK) == 0) {
X if (unlink(backupfn))
X err("Can't remove %s for pre-edit\n", backupfn);
X }
X if (link(makename, backupfn)) /* mv makefile to .bak */
X err("Can't link %s to %s.", backupfn, makename);
X backedup = TRUE; /* for interupt status */
X if (unlink(makename))
X err("Can't unlink %s during rename.", makename);
X } else {
X backupfn[0] = '\0'; /* no copy (no makefile) */
X }
X if ((makefd = fopen(makename, "w")) == NULL)
X err("Can't open output file '%s' for write.", makename);
X if (backupfn[0] != '\0') /* if no .bak file - done */
X rdwr(targ); /* read/write Makefile */
X else
X fputs(deplin, makefd); /* must start with this */
X}
X
X/* create beginging of new Makefile by reading old one */
Xvoid
Xrdwr(targ)
Xchar *targ;
X{
X register FILE *oldfd; /* file pointer for .old */
X char rwbuf[BUFSIZ]; /* temp for read/write */
X register tlen; /* targ length */
X register puntln = FALSE; /* punt current line? */
X register contln = FALSE; /* previous line ended with '\'? */
X register blankln = 0; /* number of blank lines seen */
X register srchsn = FALSE; /* search line seen? */
X
X if ((oldfd = fopen(backupfn, "r")) == NULL)
X err("Can't open backup copy of %s\n", makename);
X tlen = strlen(targ);
X while (fgets(rwbuf, BUFSIZ, oldfd) != NULL) { /* until EOF */
X if (!srchsn) {
X if (strncmp(searchdep, rwbuf, (sizeof searchdep) - 1) == 0) {
X srchsn = TRUE;
X fputs(deplin, makefd); /* re-write this line */
X if (!replace)
X break;
X continue; /* don't print this line */
X }
X } else {
X if (strcmp("\n", rwbuf) == 0) {
X if (!puntln)
X blankln++;
X contln = FALSE;
X puntln = FALSE;
X continue; /* don't output this blank line */
X }
X if (!contln) {
X if (strncmp(targ, rwbuf, tlen) == 0)
X puntln = TRUE;
X else if (strcmp(&trailer[1], rwbuf) == 0)
X puntln = TRUE;
X }
X if (lastlnch(rwbuf) == '\\')
X contln = TRUE;
X else
X contln = FALSE;
X } /* if srchsn */
X if (!puntln) {
X srchincl(&rwbuf[0]); /* search this line for defines */
X if (blankln != 0) { /* compress mult blank lines to one */
X putc('\n', makefd);
X blankln = 0;
X }
X fputs(rwbuf, makefd); /* non targ lines */
X }
X } /* while fgets */
X if (!srchsn) /* deplin never found */
X fputs(deplin, makefd); /* so write one */
X (void) fclose(oldfd); /* close the .old file for gigles */
X}
X
X#define MAXCOL 78 /* output width max for makefile */
X
X/* mkdepend - name is historical, but does the "real work" */
Xvoid
Xmkdepend(infile)
Xchar *infile;
X{
X register char *p; /* temp pointer */
X register char *q; /* temp pointer */
X register char *r; /* temp pointer */
X register FILE *cppfd; /* file desc for /lib/cpp */
X char buf[BUFSIZ]; /* temp buff */
X char basename[MAXNAMLEN+1]; /* just the file name */
X register oplen; /* length target & path */
X register le; /* length of current output line */
X register char *targ; /* target's name */
X register i; /* tmp for index */
X register firstln; /* first line of a list */
X
X if ((p = rindex(infile, '/')) == NULL) /* past path */
X p = infile;
X else
X p++;
X if (nonlocalo && p != infile) { /* objpath = source path */
X for (q = objpath, r = infile; r < p;)
X *q++ = *r++;
X *q = '\0'; /* null terminate */
X }
X strcpy(basename, p);
X if ((p = rindex(basename, '.')) != NULL)
X *p = '\0'; /* remove trailing ".*" */
X if (targetname != NULL) /* set up target's name */
X targ = targetname;
X else
X targ = basename;
X if (makename == NULL) {
X makename = "Makefile";
X if (access(makename, F_OK) != 0)
X makename[0] = 'm'; /* not a real check */
X }
X msoio(targ); /* Make Sure Output Is Open */
X abrvsetup(); /* create abrev table, write defs. */
X if (access(infile, R_OK) != 0) {
X fprintf(stderr, "%s: Can't open input file '%s', skipped.\n",
X prgnm, infile);
X return;
X }
X (void)strcpy(buf, CPP); /* build cpp cmd line */
X#if CPP_M
X strcat(buf, "-M "); /* -M flag - does dependencies */
X#endif
X if (strlen(buf) + strlen(cppflags) + strlen(infile) + 1 > BUFSIZ)
X err("cpp command line buffer overflow");
X (void)strcat(buf, cppflags); /* add command flags */
X (void)strcat(buf, infile); /* add file name */
X srtfree(&u); /* init insertion sorter */
X if (verbose)
X fprintf(stderr, "%s: cpp line is '%s'\n", prgnm, buf);
X if ((cppfd = nshpopen(buf, "r")) == NULL)
X err("Can't open pipe for %s", buf);
X#if CPP_M
X while (fgets(buf, BUFSIZ, cppfd) != NULL) {
X if ((p = index(buf, ':')) == NULL)
X err("cpp -M format error - colon");
X p++; /* pass colon */
X if (*p++ != SPC)
X err("cpp -M format error - space");
X p = hincl(p); /* skip any uglies in include path */
X if (!alldep && strncmp("/usr/include", p, 12) == 0)
X continue; /* ignore /usr/include... stuff */
X if (index(p, '\n') != NULL) /* replace newline with EOS */
X *index(p, '\n') = '\0';
X if ((q = srtin(&u, p, (int (*)())0)) != NULL) /* insert into list */
X fprintf(stderr, "%s: %s - %s\n", prgnm, q, p); /* warning */
X }
X#else
X while (fgets(buf, BUFSIZ, cppfd) != NULL) {
X if (buf[0] != '#') /* must start with '#' */
X continue;
X if ((p = index(buf, '"')) == NULL) /* find first double quote */
X continue;
X p++;
X p = hincl(p); /* skip any uglies in include path */
X if (index(p, '"') != NULL) /* terminate the file name */
X *index(p, '"') = '\0';
X if (!alldep && strncmp("/usr/include", p, 12) == 0)
X continue; /* ignore /usr/include... stuff */
X if ((q = srtin(&u, p, (int (*)())0)) != NULL) /* insert into list */
X fprintf(stderr, "%s: %s - %s\n", prgnm, q, p); /* warning */
X }
X#endif
X srtgti(&u); /* init for srtgets */
X /* 2 for colon space */
X oplen = strlen(objpath) + strlen(targ) + strlen(destsuffix) + 2;
X le = MAXCOL; /* force new line output */
X firstln = TRUE; /* first line of a file entry */
X while ((p = srtgets(&u)) != NULL) { /* write out the entries */
X if (shortincl)
X if ((i = findabr(p)) != MXABR) /* i = found index or MXABR */
X p += abrvlen[i] - 2;
X if (le + strlen(p) >= MAXCOL) {
X if (firstln) {
X le = oplen;
X fprintf(makefd, "\n%s%s%s: ", objpath, targ, destsuffix);
X firstln = FALSE;
X } else {
X le = 8;
X fprintf(makefd, " \\\n\t");
X }
X } else {
X fputc(SPC, makefd);
X }
X if (shortincl && i != MXABR) {
X fprintf(makefd, "$%c", 'A' + i);
X p += 2; /* right place */
X le += 2;
X }
X fputs(p, makefd);
X le += 1 + strlen(p);
X }
X fputc('\n', makefd); /* end with newline */
X nshpclose(cppfd); /* end of that file */
X}
END_OF_FILE
if test 16381 -ne `wc -c <'maketd/maketd.c'`; then
echo shar: \"'maketd/maketd.c'\" unpacked with wrong size!
fi
# end of 'maketd/maketd.c'
fi
echo shar: End of archive 4 \(of 4\).
cp /dev/null ark4isdone
MISSING=""
for I in 1 2 3 4 ; do
if test ! -f ark${I}isdone ; then
MISSING="${MISSING} ${I}"
fi
done
if test "${MISSING}" = "" ; then
echo You have unpacked all 4 archives.
rm -f ark[1-9]isdone
else
echo You still need to unpack the following archives:
echo " " ${MISSING}
fi
## End of shell archive.
exit 0
--
Rich $alz "Anger is an energy"
Cronus Project, BBN Labs rsalz@bbn.com
Moderator, comp.sources.unix sources@uunet.uu.net