The Automated Program Generator

This file documents AutoGen version 5.8. It is a tool designed for generating program files that contain repetitive text with varied substitutions. This document is very long because it is intended as a reference document. For a quick start example, See section A Simple Example.

The AutoGen distribution includes the basic generator engine and several add-on libraries and programs. Of the most general interest would be Automated Option processing, See section Automated Option Processing, which also includes stand-alone support for configuration file parsing, See section AutoOpts Features. Please see the "Add-on packages for AutoGen" section for additional programs and libraries associated with AutoGen.

This edition documents version 5.8, September 2006.

1. Introduction

AutoGen is a tool designed for generating program files that contain repetitive text with varied substitutions. Its goal is to simplify the maintenance of programs that contain large amounts of repetitious text. This is especially valuable if there are several blocks of such text that must be kept synchronized in parallel tables.

One common example is the problem of maintaining the code required for processing program options. Processing options requires a minimum of four different constructs be kept in proper order in different places in your program. You need at least:

1. The flag character in the flag string,
2. code to process the flag when it is encountered,
3. a global state variable or two, and
4. a line in the usage text.

You will need more things besides this if you choose to implement long option names, rc/ini/config file processing, environment variables and so on. All of this can be done mechanically; with the proper templates and this program. In fact, it has already been done and AutoGen itself uses it See section Automated Option Processing. For a simple example of Automated Option processing, See section Quick Start. For a full list of the Automated Option features, See section AutoOpts Features.

1.1 The Purpose of AutoGen

The idea of this program is to have a text file, a template if you will, that contains the general text of the desired output file. That file includes substitution expressions and sections of text that are replicated under the control of separate definition files.

AutoGen was designed with the following features:

1. The definitions are completely separate from the template. By completely isolating the definitions from the template it greatly increases the flexibility of the template implementation. A secondary goal is that a template user only needs to specify those data that are necessary to describe his application of a template.
2. Each datum in the definitions is named. Thus, the definitions can be rearranged, augmented and become obsolete without it being necessary to go back and clean up older definition files. Reduce incompatibilities!
3. Every definition name defines an array of values, even when there is only one entry. These arrays of values are used to control the replication of sections of the template.
4. There are named collections of definitions. They form a nested hierarchy. Associated values are collected and associated with a group name. These associated data are used collectively in sets of substitutions.
5. The template has special markers to indicate where substitutions are required, much like the ${VAR} construct in a shell here doc. These markers are not fixed strings. They are specified at the start of each template. Template designers know best what fits into their syntax and can avoid marker conflicts.

   We did this because it is burdensome and difficult to avoid conflicts using either M4 tokenization or C preprocessor substitution rules. It also makes it easier to specify expressions that transform the value. Of course, our expressions are less cryptic than the shell methods.

6. These same markers are used, in conjunction with enclosed keywords, to indicate sections of text that are to be skipped and for sections of text that are to be repeated. This is a major improvement over using C preprocessing macros. With the C preprocessor, you have no way of selecting output text because it is an unvarying, mechanical substitution process.
7. Finally, we supply methods for carefully controlling the output. Sometimes, it is just simply easier and clearer to compute some text or a value in one context when its application needs to be later. So, functions are available for saving text or values for later use.

1.2 A Simple Example

This is just one simple example that shows a few basic features. If you are interested, you also may run "make check" with the VERBOSE environment variable set and see a number of other examples in the `agen5/test/testdir' directory.

Assume you have an enumeration of names and you wish to associate some string with each name. Assume also, for the sake of this example, that it is either too complex or too large to maintain easily by hand. We will start by writing an abbreviated version of what the result is supposed to be. We will use that to construct our output templates.

In a header file, `list.h', you define the enumeration and the global array containing the associated strings:

typedef enum {
        IDX_ALPHA,
        IDX_BETA,
        IDX_OMEGA }  list_enum;

extern char const* az_name_list[ 3 ];

Then you also have `list.c' that defines the actual strings:

#include "list.h"
char const* az_name_list[] = {
        "some alpha stuff",
        "more beta stuff",
        "final omega stuff" };

First, we will define the information that is unique for each enumeration name/string pair. This would be placed in a file named, `list.def', for example.

autogen definitions list;
list = { list_element = alpha;
         list_info    = "some alpha stuff"; };
list = { list_info    = "more beta stuff";
         list_element = beta; };
list = { list_element = omega;
         list_info    = "final omega stuff"; };

The autogen definitions list; entry defines the file as an AutoGen definition file that uses a template named list. That is followed by three list entries that define the associations between the enumeration names and the strings. The order of the differently named elements inside of list is unimportant. They are reversed inside of the beta entry and the output is unaffected.

Now, to actually create the output, we need a template or two that can be expanded into the files you want. In this program, we use a single template that is capable of multiple output files. The definitions above refer to a `list' template, so it would normally be named, `list.tpl'.

It looks something like this. (For a full description, See section Template File.)

[+ AutoGen5 template h c +]
[+ CASE (suffix) +][+
   ==  h  +]
typedef enum {[+
   FOR list "," +]
        IDX_[+ (string-upcase! (get "list_element")) +][+
   ENDFOR list +] }  list_enum;

extern char const* az_name_list[ [+ (count "list") +] ];
[+

   ==  c  +]
#include "list.h"
char const* az_name_list[] = {[+
  FOR list "," +]
        "[+list_info+]"[+
  ENDFOR list +] };[+

ESAC +]

The [+ AutoGen5 template h c +] text tells AutoGen that this is an AutoGen version 5 template file; that it is to be processed twice; that the start macro marker is [+; and the end marker is +]. The template will be processed first with a suffix value of h and then with c. Normally, the suffix values are appended to the `base-name' to create the output file name.

The [+ == h +] and [+ == c +] CASE selection clauses select different text for the two different passes. In this example, the output is nearly disjoint and could have been put in two separate templates. However, sometimes there are common sections and this is just an example.

The [+FOR list "," +] and [+ ENDFOR list +] clauses delimit a block of text that will be repeated for every definition of list. Inside of that block, the definition name-value pairs that are members of each list are available for substitutions.

The remainder of the macros are expressions. Some of these contain special expression functions that are dependent on AutoGen named values; others are simply Scheme expressions, the result of which will be inserted into the output text. Other expressions are names of AutoGen values. These values will be inserted into the output text. For example, [+list_info+] will result in the value associated with the name list_info being inserted between the double quotes and (string-upcase! (get "list_element")) will first "get" the value associated with the name list_element, then change the case of all the letters to upper case. The result will be inserted into the output document.

If you have compiled AutoGen, you can copy out the template and definitions as described above and run autogen list.def. This will produce exactly the hypothesized desired output.

One more point, too. Lets say you decided it was too much trouble to figure out how to use AutoGen, so you created this enumeration and string list with thousands of entries. Now, requirements have changed and it has become necessary to map a string containing the enumeration name into the enumeration number. With AutoGen, you just alter the template to emit the table of names. It will be guaranteed to be in the correct order, missing none of the entries. If you want to do that by hand, well, good luck.

1.3 csh/zsh caveat

AutoGen tries to use your normal shell so that you can supply shell code in a manner you are accustomed to using. If, however, you use csh or zsh, you cannot do this. Csh is sufficiently difficult to program that it is unsupported. Zsh, though largely programmable, also has some anomalies that make it incompatible with AutoGen usage. Therefore, when invoking AutoGen from these environments, you must be certain to set the SHELL environment variable to a Bourne-derived shell, e.g., sh, ksh or bash.

Any shell you choose for your own scripts need to follow these basic requirements:

1. It handles trap $sig ":" without output to standard out. This is done when the server shell is first started. If your shell does not handle this, then it may be able to by loading functions from its start up files.
2. At the beginning of each scriptlet, the command \\cd $PWD is inserted. This ensures that cd is not aliased to something peculiar and each scriptlet starts life in the execution directory.
3. At the end of each scriptlet, the command echo mumble is appended. The program you use as a shell must emit the single argument mumble on a line by itself.

1.4 A User's Perspective

Alexandre wrote:
>
> I'd appreciate opinions from others about advantages/disadvantages of
> each of these macro packages.

I am using AutoGen in my pet project, and find one of its best points to be that it separates the operational data from the implementation.

Indulge me for a few paragraphs, and all will be revealed: In the manual, Bruce cites the example of maintaining command line flags inside the source code; traditionally spreading usage information, flag names, letters and processing across several functions (if not files). Investing the time in writing a sort of boiler plate (a template in AutoGen terminology) pays by moving all of the option details (usage, flags names etc.) into a well structured table (a definition file if you will), so that adding a new command line option becomes a simple matter of adding a set of details to the table.

So far so good! Of course, now that there is a template, writing all of that tedious optargs processing and usage functions is no longer an issue. Creating a table of the options needed for the new project and running AutoGen generates all of the option processing code in C automatically from just the tabular data. AutoGen in fact already ships with such a template... AutoOpts.

One final consequence of the good separation in the design of AutoGen is that it is retargetable to a greater extent. The egcs/gcc/fixinc/inclhack.def can equally be used (with different templates) to create a shell script (inclhack.sh) or a c program (fixincl.c).

This is just the tip of the iceberg. AutoGen is far more powerful than these examples might indicate, and has many other varied uses. I am certain Bruce or I could supply you with many and varied examples, and I would heartily recommend that you try it for your project and see for yourself how it compares to m4.

As an aside, I would be interested to see whether someone might be persuaded to rationalise autoconf with AutoGen in place of m4... Ben, are you listening? autoconf-3.0! `kay? =)O|

Sincerely,
        Gary V. Vaughan

2. Definitions File

This chapter describes the syntax and semantics of the AutoGen definition file. In order to instantiate a template, you normally must provide a definitions file that identifies itself and contains some value definitions. Consequently, we keep it very simple. For "advanced" users, there are preprocessing directives, sparse arrays, named indexes and comments that may be used as well.

The definitions file is used to associate values with names. Every value is implicitly an array of values, even if there is only one value. Values may be either simple strings or compound collections of name-value pairs. An array may not contain both simple and compound members. Fundamentally, it is as simple as:

prog-name = "autogen";
flag = {
    name      = templ_dirs;
    value     = L;
    descrip   = "Template search directory list";
};

For purposes of commenting and controlling the processing of the definitions, C-style comments and most C preprocessing directives are honored. The major exception is that the #if directive is ignored, along with all following text through the matching #endif directive. The C preprocessor is not actually invoked, so C macro substitution is not performed.

2.1 The Identification Definition

The first definition in this file is used to identify it as a AutoGen file. It consists of the two keywords, `autogen' and `definitions' followed by the default template name and a terminating semi-colon (;). That is:

        AutoGen Definitions template-name;

Note that, other than the name template-name, the words `AutoGen' and `Definitions' are searched for without case sensitivity. Most lookups in this program are case insensitive.

Also, if the input contains more identification definitions, they will be ignored. This is done so that you may include (see section Controlling What Gets Processed) other definition files without an identification conflict.

AutoGen uses the name of the template to find the corresponding template file. It searches for the file in the following way, stopping when it finds the file:

1. It tries to open `./template-name'. If it fails,
2. it tries `./template-name.tpl'.
3. It searches for either of these files in the directories listed in the templ-dirs command line option.

If AutoGen fails to find the template file in one of these places, it prints an error message and exits.

2.2 Named Definitions

Any name may have multiple values associated with it in the definition file. If there is more than one instance, the only way to expand all of the copies of it is by using the FOR (see section FOR - Emit a template block multiple times) text function on it, as described in the next chapter.

There are two kinds of definitions, `simple' and `compound'. They are defined thus (see section Finite State Machine Grammar):

compound_name '=' '{' definition-list '}' ';'

simple_name '=' string ';'

no_text_name ';'

No_text_name is a simple definition with a shorthand empty string value. The string values for definitions may be specified in any of several formation rules.

2.2.1 Definition List

definition-list is a list of definitions that may or may not contain nested compound definitions. Any such definitions may only be expanded within a FOR block iterating over the containing compound definition. See section FOR - Emit a template block multiple times.

Here is, again, the example definitions from the previous chapter, with three additional name value pairs. Two with an empty value assigned (first and last), and a "global" group_name.

autogen definitions list;
group_name = example;
list = { list_element = alpha;  first;
         list_info    = "some alpha stuff"; };
list = { list_info    = "more beta stuff";
         list_element = beta; };
list = { list_element = omega;  last;
         list_info    = "final omega stuff"; };

2.2.2 Double Quote String

The string follows the C-style escaping (\, \n, \f, \v, etc.), plus octal character numbers specified as \ooo. The difference from "C" is that the string may span multiple lines. Like ANSI "C", a series of these strings, possibly intermixed with single quote strings, will be concatenated together.

2.2.3 Single Quote String

This is similar to the shell single-quote string. However, escapes \ are honored before another escape, single quotes ' and hash characters #. This latter is done specifically to disambiguate lines starting with a hash character inside of a quoted string. In other words,

fumble = '
#endif
';

could be misinterpreted by the definitions scanner, whereas this would not:

fumble = '
\#endif
';

As with the double quote string, a series of these, even intermixed with double quote strings, will be concatenated together.

2.2.4 Shell Output String

This is assembled according to the same rules as the double quote string, except that there is no concatenation of strings and the resulting string is written to a shell server process. The definition takes on the value of the output string.

NB The text is interpreted by a server shell. There may be left over state from previous server shell processing. This scriptlet may also leave state for subsequent processing. However, a cd to the original directory is always issued before the new command is issued.

2.2.5 An Unquoted String

A simple string that does not contain white space may be left unquoted. The string must not contain any of the characters special to the definition text (i.e., ", #, ', (, ), ,, ;, <, =, >, [, ], `, {, or }). This list is subject to change, but it will never contain underscore (_), period (.), slash (/), colon (:), hyphen (-) or backslash (\\). Basically, if the string looks like it is a normal DOS or UNIX file or variable name, and it is not one of two keywords (`autogen' or `definitions') then it is OK to not quote it, otherwise you should.

2.2.6 Scheme Result String

A scheme result string must begin with an open parenthesis (. The scheme expression will be evaluated by Guile and the value will be the result. The AutoGen expression functions are disabled at this stage, so do not use them.

2.2.7 A Here String

A `here string' is formed in much the same way as a shell here doc. It is denoted with a doubled less than character and, optionally, a hyphen. This is followed by optional horizontal white space and an ending marker-identifier. This marker must follow the syntax rules for identifiers. Unlike the shell version, however, you must not quote this marker. The resulting string will start with the first character on the next line and continue up to but not including the newline that precedes the line that begins with the marker token. No backslash or any other kind of processing is done on this string. The characters are copied directly into the result string.

Here are two examples:

str1 = <<-  STR_END
        $quotes = " ' `
        STR_END;

str2 = <<   STR_END
        $quotes = " ' `
        STR_END;
STR_END;

The first string contains no new line characters. The first character is the dollar sign, the last the back quote.

The second string contains one new line character. The first character is the tab character preceding the dollar sign. The last character is the semicolon after the STR_END. That STR_END does not end the string because it is not at the beginning of the line. In the preceding case, the leading tab was stripped.

2.2.8 Concatenated Strings

If single or double quote characters are used, then you also have the option, a la ANSI-C syntax, of implicitly concatenating a series of them together, with intervening white space ignored.

NB You cannot use directives to alter the string content. That is,

str = "fumble"
#ifdef LATER
      "stumble"
#endif
      ;

will result in a syntax error. The preprocessing directives are not carried out by the C preprocessor. However,

str = '"fumble\n"
#ifdef LATER
"     stumble\n"
#endif
';

Will work. It will enclose the `#ifdef LATER' and `#endif' in the string. But it may also wreak havoc with the definition processing directives. The hash characters in the first column should be disambiguated with an escape \ or join them with previous lines: "fumble\n#ifdef LATER....

2.3 Assigning an Index to a Definition

In AutoGen, every name is implicitly an array of values. When assigning values, they are usually implicitly assigned to the next highest slot. They can also be specified explicitly:

mumble[9] = stumble;
mumble[0] = grumble;

If, subsequently, you assign a value to mumble without an index, its index will be 10, not 1. If indexes are specified, they must not cause conflicts.

#define-d names may also be used for index values. This is equivalent to the above:

#define FIRST 0
#define LAST  9
mumble[LAST]  = stumble;
mumble[FIRST] = grumble;

All values in a range do not have to be filled in. If you leave gaps, then you will have a sparse array. This is fine (see section FOR - Emit a template block multiple times). You have your choice of iterating over all the defined values, or iterating over a range of slots. This:

[+ FOR mumble +][+ ENDFOR +]

iterates over all and only the defined entries, whereas this:

[+ FOR mumble (for-by 1) +][+ ENDFOR +]

will iterate over all 10 "slots". Your template will likely have to contain something like this:

[+ IF (exist? (sprintf "mumble[%d]" (for-index))) +]

or else "mumble" will have to be a compound value that, say, always contains a "grumble" value:

[+ IF (exist? "grumble") +]

2.4 Dynamic Text

There are several methods for including dynamic content inside a definitions file. Three of them are mentioned above (Shell Output String and see section Scheme Result String) in the discussion of string formation rules. Another method uses the #shell processing directive. It will be discussed in the next section (see section Controlling What Gets Processed). Guile/Scheme may also be used to yield to create definitions.

When the Scheme expression is preceded by a backslash and single quote, then the expression is expected to be an alist of names and values that will be used to create AutoGen definitions.

This method can be be used as follows:

\'( (name  (value-expression))
    (name2 (another-expr))  )

This is entirely equivalent to:

name  = (value-expression);
name2 = (another-expr);

Under the covers, the expression gets handed off to a Guile function named alist->autogen-def in an expression that looks like this:

(alist->autogen-def
    ( (name (value-expression))  (name2 (another-expr)) ) )

2.5 Controlling What Gets Processed

Definition processing directives can only be processed if the '#' character is the first character on a line. Also, if you want a '#' as the first character of a line in one of your string assignments, you should either escape it by preceding it with a backslash `\', or by embedding it in the string as in "\n#".

All of the normal C preprocessing directives are recognized, though several are ignored. There is also an additional #shell - #endshell pair. Another minor difference is that AutoGen directives must have the hash character (#) in column 1.

The final tweak is that #! is treated as a comment line. Using this feature, you can use: `#! /usr/local/bin/autogen' as the first line of a definitions file, set the mode to executable and "run" the definitions file as if it were a direct invocation of AutoGen. This was done for its hack value.

The ignored directives are: `#ident', `#let', `#pragma', and `#if'. Note that when ignoring the #if directive, all intervening text through its matching #endif is also ignored, including the #else clause.

The AutoGen directives that affect the processing of definitions are:

#assert `shell-script` | (scheme-expr) | <anything else>

If the shell-script or scheme-expr do not yield true valued results, autogen will be aborted. If <anything else> or nothing at all is provided, then this directive is ignored.

When writing the shell script, remember this is on a preprocessing line. Multiple lines must be backslash continued and the result is a single long line. Separate multiple commands with semi-colons.

The result is false (and fails) if the result is empty, the number zero, or a string that starts with the letters 'n' or 'f' ("no" or "false").

#define name [ <text> ]

Will add the name to the define list as if it were a DEFINE program argument. Its value will be the first non-whitespace token following the name. Quotes are not processed.

After the definitions file has been processed, any remaining entries in the define list will be added to the environment.

#elif

This must follow an #if otherwise it will generate an error. It will be ignored.

#else

This must follow an #if, #ifdef or #ifndef. If it follows the #if, then it will be ignored. Otherwise, it will change the processing state to the reverse of what it was.

#endif

This must follow an #if, #ifdef or #ifndef. In all cases, this will resume normal processing of text.

#endmac

This terminates a "macdef", but must not ever be encountered directly.

#endshell

Ends the text processed by a command shell into autogen definitions.

#error [ <descriptive text> ]

This directive will cause AutoGen to stop processing and exit with a status of EXIT_FAILURE.

#if [ <ignored conditional expression> ]

#if expressions are not analyzed. Everything from here to the matching #endif is skipped.

#ifdef name-to-test

The definitions that follow, up to the matching #endif will be processed only if there is a corresponding -Dname command line option or if a #define of that name has been previously encountered.

#ifndef name-to-test

The definitions that follow, up to the matching #endif will be processed only if there is not a corresponding -Dname command line option or there was a canceling -Uname option.

#include unadorned-file-name

This directive will insert definitions from another file into the current collection. If the file name is adorned with double quotes or angle brackets (as in a C program), then the include is ignored.

#line

Alters the current line number and/or file name. You may wish to use this directive if you extract definition source from other files. getdefs uses this mechanism so AutoGen will report the correct file and approximate line number of any errors found in extracted definitions.

#macdef

This is a new AT&T research preprocessing directive. Basically, it is a multi-line #define that may include other preprocessing directives.

#option opt-name [ <text> ]

This directive will pass the option name and associated text to the AutoOpts optionLoadLine routine (see section optionLoadLine). The option text may span multiple lines by continuing them with a backslash. The backslash/newline pair will be replaced with two space characters. This directive may be used to set a search path for locating template files For example, this:

#option templ-dirs $ENVVAR/dirname

will direct autogen to use the ENVVAR environment variable to find a directory named dirname that (may) contain templates. Since these directories are searched in most recently supplied first order, search directories supplied in this way will be searched before any supplied on the command line.

#shell

Invokes $SHELL or `/bin/sh' on a script that should generate AutoGen definitions. It does this using the same server process that handles the back-quoted ` text. CAUTION let not your $SHELL be csh.

#undef name-to-undefine

Will remove any entries from the define list that match the undef name pattern.

2.6 Pre-defined Names

When AutoGen starts, it tries to determine several names from the operating environment and put them into environment variables for use in both #ifdef tests in the definitions files and in shell scripts with environment variable tests. __autogen__ is always defined. For other names, AutoGen will first try to use the POSIX version of the sysinfo(2) system call. Failing that, it will try for the POSIX uname(2) call. If neither is available, then only "__autogen__" will be inserted into the environment. In all cases, the associated names are converted to lower case, surrounded by doubled underscores and non-symbol characters are replaced with underscores.

With Solaris on a sparc platform, sysinfo(2) is available. The following strings are used:

• SI_SYSNAME (e.g., "__sunos__")
• SI_HOSTNAME (e.g., "__ellen__")
• SI_ARCHITECTURE (e.g., "__sparc__")
• SI_HW_PROVIDER (e.g., "__sun_microsystems__")
• SI_PLATFORM (e.g., "__sun_ultra_5_10__")
• SI_MACHINE (e.g., "__sun4u__")

For Linux and other operating systems that only support the uname(2) call, AutoGen will use these values:

• sysname (e.g., "__linux__")
• machine (e.g., "__i586__")
• nodename (e.g., "__bach__")

By testing these pre-defines in my definitions, you can select pieces of the definitions without resorting to writing shell scripts that parse the output of uname(1). You can also segregate real C code from autogen definitions by testing for "__autogen__".

#ifdef __bach__
  location = home;
#else
  location = work;
#endif

2.7 Commenting Your Definitions

The definitions file may contain C and C++ style comments.

/*
 *  This is a comment.  It continues for several lines and closes
 *  when the characters '*' and '/' appear together.
 */
// this comment is a single line comment

2.8 What it all looks like.

This is an extended example:

autogen definitions `template-name';
/*
 *  This is a comment that describes what these
 *  definitions are all about.
 */
global = "value for a global text definition.";

/*
 *  Include a standard set of definitions
 */
#include standards.def

a_block = {
    a_field;
    a_subblock = {
        sub_name  = first;
        sub_field = "sub value.";
    };

#ifdef FEATURE
    a_subblock = {
        sub_name  = second;
    };
#endif

};

2.9 Finite State Machine Grammar

The preprocessing directives and comments are not part of the grammar. They are handled by the scanner/lexer. The following was extracted directly from the generated defParse-fsm.c source file. The "EVT:" is the token seen, the "STATE:" is the current state and the entries in this table describe the next state and the action to take. Invalid transitions were removed from the table.

dp_trans_table[ DP_STATE_CT ][ DP_EVENT_CT ] = {

  /* STATE 0:  DP_ST_INIT */
  { { DP_ST_NEED_DEF, NULL },                       /* EVT:  autogen */

  /* STATE 1:  DP_ST_NEED_DEF */
    { DP_ST_NEED_TPL, NULL },                       /* EVT:  definitions */

  /* STATE 2:  DP_ST_NEED_TPL */
    { DP_ST_NEED_SEMI, &dp_do_tpl_name },           /* EVT:  var_name */
    { DP_ST_NEED_SEMI, &dp_do_tpl_name },           /* EVT:  other_name */
    { DP_ST_NEED_SEMI, &dp_do_tpl_name },           /* EVT:  string */

  /* STATE 3:  DP_ST_NEED_SEMI */
    { DP_ST_NEED_NAME, NULL },                      /* EVT:  ; */

  /* STATE 4:  DP_ST_NEED_NAME */
  { { DP_ST_NEED_DEF, NULL },                       /* EVT:  autogen */
    { DP_ST_DONE, &dp_do_need_name_end },           /* EVT:  End-Of-File */
    { DP_ST_HAVE_NAME, &dp_do_need_name_var_name }, /* EVT:  var_name */
    { DP_ST_HAVE_VALUE, &dp_do_end_block },         /* EVT:  } */

  /* STATE 5:  DP_ST_HAVE_NAME */
    { DP_ST_NEED_NAME, &dp_do_empty_val },          /* EVT:  ; */
    { DP_ST_NEED_VALUE, &dp_do_have_name_lit_eq },  /* EVT:  = */
    { DP_ST_NEED_IDX, NULL },                       /* EVT:  [ */

  /* STATE 6:  DP_ST_NEED_VALUE */
    { DP_ST_HAVE_VALUE, &dp_do_str_value },         /* EVT:  var_name */
    { DP_ST_HAVE_VALUE, &dp_do_str_value },         /* EVT:  other_name */
    { DP_ST_HAVE_VALUE, &dp_do_str_value },         /* EVT:  string */
    { DP_ST_HAVE_VALUE, &dp_do_str_value },         /* EVT:  here_string */
    { DP_ST_HAVE_VALUE, &dp_do_str_value },         /* EVT:  number */
    { DP_ST_NEED_NAME, &dp_do_start_block },        /* EVT:  { */

  /* STATE 7:  DP_ST_NEED_IDX */
    { DP_ST_NEED_CBKT, &dp_do_indexed_name },       /* EVT:  var_name */
    { DP_ST_NEED_CBKT, &dp_do_indexed_name },       /* EVT:  number */

  /* STATE 8:  DP_ST_NEED_CBKT */
    { DP_ST_INDX_NAME, NULL }                       /* EVT:  ] */

  /* STATE 9:  DP_ST_INDX_NAME */
    { DP_ST_NEED_NAME, &dp_do_empty_val },          /* EVT:  ; */
    { DP_ST_NEED_VALUE, NULL },                     /* EVT:  = */

  /* STATE 10:  DP_ST_HAVE_VALUE */
    { DP_ST_NEED_NAME, NULL },                      /* EVT:  ; */
    { DP_ST_NEED_VALUE, &dp_do_next_val },          /* EVT:  , */

2.10 Alternate Definition Forms

There are several methods for supplying data values for templates.

`no definitions'

It is entirely possible to write a template that does not depend upon external definitions. Such a template would likely have an unvarying output, but be convenient nonetheless because of an external library of either AutoGen or Scheme functions, or both. This can be accommodated by providing the --override-tpl and --no-definitions options on the command line. See section Invoking autogen.

`CGI'

AutoGen behaves as a CGI server if the definitions input is from stdin and the environment variable REQUEST_METHOD is defined and set to either "GET" or "POST", See section AutoGen as a CGI server. Obviously, all the values are constrained to strings because there is no way to represent nested values.

`XML'

AutoGen comes with a program named, xml2ag. Its output can either be redirected to a file for later use, or the program can be used as an AutoGen wrapper. See section Invoking xml2ag.

The introductory template example (see section A Simple Example) can be rewritten in XML as follows:

<EXAMPLE template="list.tpl"> <LIST list_element="alpha" list_info="some alpha stuff"/> <LIST list_info="more beta stuff" list_element="beta"/> <LIST list_element="omega" list_info="final omega stuff"/> </EXAMPLE>

A more XML-normal form might look like this:

<EXAMPLE template="list.tpl"> <LIST list_element="alpha">some alpha stuff</LIST> <LIST list_element="beta" >more beta stuff</LIST> <LIST list_element="omega">final omega stuff</LIST> </EXAMPLE>

but you would have to change the template list_info references into text references.

`standard AutoGen definitions'

Of course. :-)

3. Template File

The AutoGen template file defines the content of the output text. It is composed of two parts. The first part consists of a pseudo macro invocation and commentary. It is followed by the template proper.

This pseudo macro is special. It is used to identify the file as a AutoGen template file, fixing the starting and ending marks for the macro invocations in the rest of the file, specifying the list of suffixes to be generated by the template and, optionally, the shell to use for processing shell commands embedded in the template.

AutoGen-ing a file consists of copying text from the template to the output file until a start macro marker is found. The text from the start marker to the end marker constitutes the macro text. AutoGen macros may cause sections of the template to be skipped or processed several times. The process continues until the end of the template is reached. The process is repeated once for each suffix specified in the pseudo macro.

This chapter describes the format of the AutoGen template macros and the usage of the AutoGen native macros. Users may augment these by defining their own macros, See section DEFINE - Define a user AutoGen macro.

3.1 Format of the Pseudo Macro

The pseudo macro is used to tell AutoGen how to process a template. It tells autogen:

1. The start macro marker. It consists of punctuation characters used to demarcate the start of a macro. It may be up to seven characters long and must be the first non-whitespace characters in the file.

   It is generally a good idea to use some sort of opening bracket in the starting macro and closing bracket in the ending macro (e.g. {, (, [, or even < in the starting macro). It helps both visually and with editors capable of finding a balancing parenthesis.

2. That start marker must be immediately followed by the identifier strings "AutoGen5" and then "template", though capitalization is not important.

The next several components may be intermingled:

1. Zero, one or more suffix specifications tell AutoGen how many times to process the template file. No suffix specifications mean that it is to be processed once and that the generated text is to be written to stdout. The current suffix for each pass can be determined with the (suffix) scheme function (see section `suffix' - get the current suffix).

   The suffix specification consists of a sequence of POSIX compliant file name characters and, optionally, an equal sign and a file name formatting specification. That specification may be either an ordinary sequence of file name characters with zero, one or two "%s" formatting sequences in it, or else it may be a Scheme expression that, when evaluated, produces such a string. The two string arguments allowed for that string are the base name of the definition file, and the current suffix (that being the text to the left of the equal sign). (Note: "POSIX compliant file name characters" consist of alphanumerics plus the period (.), hyphen (-) and underscore (_) characters.)

   If the suffix begins with one of these three latter characters and a formatting string is not specified, then that character is presumed to be the suffix separator. Otherwise, without a specified format string, a single period will separate the suffix from the base name in constructing the output file name.

2. Comments: blank lines, lines starting with a hash mark [#]), and edit mode comments (text between pairs of -*- strings).
3. Scheme expressions may be inserted in order to make configuration changes before template processing begins. It is used, for example, to allow the template writer to specify the shell program that must be used to interpret the shell commands in the template. It can have no effect on any shell commands in the definitions file, as that file will have been processed by the time the pseudo macro is interpreted.

   (setenv "SHELL" "/bin/sh")

   This is extremely useful to ensure that the shell used is the one the template was written to use. By default, AutoGen determines the shell to use by user preferences. Sometimes, that can be the "csh", though.

   The scheme expression can also be used to save a pre-existing output file for later text extraction (see section `extract' - extract text from another file).

   (shellf "mv -f %1$s.c %1$s.sav" (base-name))

After these must come the end macro marker:

1. The punctuation characters used to demarcate the end of a macro. Like the start marker, it must consist of seven or fewer punctuation characters.

The ending macro marker has a few constraints on its content. Some of them are just advisory, though. There is no special check for advisory restrictions.

• It must not begin with a POSIX file name character (hyphen -, underscore _ or period .), the backslash (\) or open parenthesis ((). These are used to identify a suffix specification, indicate Scheme code and trim white space.
• If it begins with an equal sign, then it must be separated from any suffix specification by white space.
• The closing marker may not begin with an open parenthesis, as that is used to enclose a scheme expression.
• It cannot begin with a backslash, as that is used to indicate white space trimming after the end macro mark. If, in the body of the template, you put the backslash character (\) before the end macro mark, then any white space characters after the mark and through the newline character are trimmed.
• It is also helpful to avoid using the comment marker (#). It might be seen as a comment within the pseudo macro.
• You should avoid using any of the quote characters double, single or back-quote. It won't confuse AutoGen, but it might well confuse you and/or your editor.

As an example, assume we want to use [+ and +] as the start and end macro markers, and we wish to produce a `.c' and a `.h' file, then the pseudo macro might look something like this:

[+ AutoGen5 template -*- Mode: emacs-mode-of-choice -*-
h=chk-%s.h
c
# make sure we don't use csh:
(setenv "SHELL" "/bin/sh")  +]

The template proper starts after the pseudo-macro. The starting character is either the first non-whitespace character or the first character after the newline that follows the end macro marker.

3.2 Naming a value

When an AutoGen value is specified in a template, it is specified by name. The name may be a simple name, or a compound name of several components. Since each named value in AutoGen is implicitly an array of one or more values, each component may have an index associated with it.

It looks like this:

comp-name-1 . comp-name-2 [ 2 ]

Note that if there are multiple components to a name, each component name is separated by a dot (.). Indexes follow a component name, enclosed in square brackets ([ and ]). The index may be either an integer or an integer-valued define name. The first component of the name is searched for in the current definition level. If not found, higher levels will be searched until either a value is found, or there are no more definition levels. Subsequent components of the name must be found within the context of the newly-current definition level. Also, if the named value is prefixed by a dot (.), then the value search is started in the current context only. No higher levels are searched.

If someone rewrites this, I'll incorporate it. :-)

3.3 Macro Expression Syntax

AutoGen has two types of expressions: full expressions and basic ones. A full AutoGen expression can appear by itself, or as the argument to certain AutoGen built-in macros: CASE, IF, ELIF, INCLUDE, INVOKE (explicit invocation, see section INVOKE - Invoke a User Defined Macro), and WHILE. If it appears by itself, the result is inserted into the output. If it is an argument to one of these macros, the macro code will act on it sensibly.

You are constrained to basic expressions only when passing arguments to user defined macros, See section DEFINE - Define a user AutoGen macro.

The syntax of a full AutoGen expression is:

[[ <apply-code> ] <value-name> ] [ <basic-expr-1> [ <basic-expr-2> ]]

How the expression is evaluated depends upon the presence or absence of the apply code and value name. The "value name" is the name of an AutoGen defined value, or not. If it does not name such a value, the expression result is generally the empty string. All expressions must contain either a value-name or a basic-expr.

3.3.1 Apply Code

The "apply code" selected determines the method of evaluating the expression. There are five apply codes, including the non-use of an apply code.

`no apply code'

This is the most common expression type. Expressions of this sort come in three flavors:

`<value-name>'

The result is the value of value-name, if defined. Otherwise it is the empty string.

`<basic-expr>'

The result of the basic expression is the result of the full expression, See section Basic Expression.

`<value-name> <basic-expr>'

If there is a defined value for value-name, then the basic-expr is evaluated. Otherwise, the result is the empty string.

`% <value-name> <basic-expr>'

If value-name is defined, use basic-expr as a format string for sprintf. Then, if the basic-expr is either a back-quoted string or a parenthesized expression, then hand the result to the appropriate interpreter for further evaluation. Otherwise, for single and double quote strings, the result is the result of the sprintf operation. Naturally, if value-name is not defined, the result is the empty string.

For example, assume that fumble had the string value, stumble:

[+ % fumble `printf '%%x\\n' $%s` +]

This would cause the shell to evaluate "printf '%x\n' $stumble". Assuming that the shell variable stumble had a numeric value, the expression result would be that number, in hex. Note the need for doubled percent characters and backslashes.

`? <value-name> <basic-expr-1> <basic-expr-2>'

Two basic-expr-s are required. If the value-name is defined, then the first basic-expr-1 is evaluated, otherwise basic-expr-2 is.

`- <value-name> <basic-expr>'

Evaluate basic-expr only if value-name is not defined.

`?% <value-name> <basic-expr-1> <basic-expr-2>'

This combines the functions of `?' and `%'. If value-name is defined, it behaves exactly like `%', above, using basic-expr-1. If not defined, then basic-expr-2 is evaluated.

For example, assume again that fumble had the string value, stumble:

[+ ?% fumble `cat $%s` `pwd` +]

This would cause the shell to evaluate "cat $stumble". If fumble were not defined, then the result would be the name of our current directory.

3.3.2 Basic Expression

A basic expression can have one of the following forms:

`'STRING''

A single quoted string. Backslashes can be used to protect single quotes ('), hash characters (#), or backslashes (\) in the string. All other characters of STRING are output as-is when the single quoted string is evaluated. Backslashes are processed before the hash character for consistency with the definition syntax. It is needed there to avoid preprocessing conflicts.

`"STRING"'

A double quoted string. This is a cooked text string as in C, except that they are not concatenated with adjacent strings. Evaluating "STRING" will output STRING with all backslash sequences interpreted.

``STRING`'

A back quoted string. When this expression is evaluated, STRING is first interpreted as a cooked string (as in `"STRING"') and evaluated as a shell expression by the AutoGen server shell. This expression is replaced by the stdout output of the shell.

`(STRING)'

A parenthesized expression. It will be passed to the Guile interpreter for evaluation and replaced by the resulting value. If there is a Scheme error in this expression, Guile 1.4 and Guile 1.6 will report the template line number where the error occurs. Guile 1.7 has lost this capability.

Additionally, other than in the % and ?% expressions, the Guile expressions may be introduced with the Guile comment character (;) and you may put a series of Guile expressions within a single macro. They will be implicitly evaluated as if they were arguments to the (begin ...) expression. The result will be the result of the last Guile expression evaluated.

3.4 AutoGen Scheme Functions

AutoGen uses Guile to interpret Scheme expressions within AutoGen macros. All of the normal Guile functions are available, plus several extensions (see section Common Scheme Functions) have been added to augment the repertoire of string manipulation functions and manage the state of AutoGen processing.

This section describes those functions that are specific to AutoGen. Please take note that these AutoGen specific functions are not loaded and thus not made available until after the command line options have been processed and the AutoGen definitions have been loaded. They may, of course, be used in Scheme functions that get defined at those times, but they cannot be invoked.

3.4.1 `ag-function?' - test for function

Usage: (ag-function? ag-name)
return SCM_BOOL_T if a specified name is a user-defined AutoGen macro, otherwise return SCM_BOOL_F.

Arguments:
ag-name - name of AutoGen macro

3.4.2 `base-name' - base output name

Usage: (base-name)
Returns a string containing the base name of the output file(s). Generally, this is also the base name of the definitions file.

This Scheme function takes no arguments.

3.4.3 `chdir' - Change current directory

Usage: (chdir dir)
Sets the current directory for AutoGen. Shell commands will run from this directory as well. This is a wrapper around the Guile native function. It returns its directory name argument and fails the program on failure.

Arguments:
dir - new directory name

3.4.4 `count' - definition count

Usage: (count ag-name)
Count the number of entries for a definition. The input argument must be a string containing the name of the AutoGen values to be counted. If there is no value associated with the name, the result is an SCM immediate integer value of zero.

Arguments:
ag-name - name of AutoGen value

3.4.5 `def-file' - definitions file name

Usage: (def-file)
Get the name of the definitions file. Returns the name of the source file containing the AutoGen definitions.

This Scheme function takes no arguments.

3.4.6 `def-file-line' - get a definition file+line number

Usage: (def-file-line ag-name [ msg-fmt ])
Returns the file and line number of a AutoGen defined value, using either the default format, "from %s line %d", or else the format you supply. For example, if you want to insert a "C" language file-line directive, you would supply the format "# %2$d \"%1$s\"", but that is also already supplied with the scheme variable See section format file info as, "#line nn "file"". You may use it thus:

(def-file-line "ag-def-name" c-file-line-fmt)

It is also safe to use the formatting string, "%2$d". AutoGen uses an argument vector version of printf: See section Replacement for Stdio Formatting Library.

Arguments:
ag-name - name of AutoGen value
msg-fmt - Optional - formatting for line message

3.4.7 `dne' - "Do Not Edit" warning

Usage: (dne prefix [ first_prefix ] [ optpfx ])
Generate a "DO NOT EDIT" or "EDIT WITH CARE" warning string. Which depends on whether or not the --writable command line option was set. The first argument is a per-line string prefix. The optional second argument is a prefix for the first-line and, in read-only mode, activates the editor hints.

-*- buffer-read-only: t -*- vi: set ro:

The warning string also includes information about the template used to construct the file and the definitions used in its instantiation.

The optional third argument is used when the first argument is actually an invocation option and the prefix arguments get shifted. The first argument must be, specifically, "-d". That is used to signify that the date stamp should not be inserted into the output.

Arguments:
prefix - string for starting each output line
first_prefix - Optional - for the first output line
optpfx - Optional - shifted prefix

3.4.8 `error' - display message and exit

Usage: (error message)
The argument is a string that printed out as part of an error message. The message is formed from the formatting string:

DEFINITIONS ERROR in %s line %d for %s:  %s\n

The first three arguments to this format are provided by the routine and are: The name of the template file, the line within the template where the error was found, and the current output file name.

After displaying the message, the current output file is removed and autogen exits with the EXIT_FAILURE error code. IF, however, the argument begins with the number 0 (zero), or the string is the empty string, then processing continues with the next suffix.

Arguments:
message - message to display before exiting

3.4.9 `exist?' - test for value name

Usage: (exist? ag-name)
return SCM_BOOL_T iff a specified name has an AutoGen value. The name may include indexes and/or member names. All but the last member name must be an aggregate definition. For example:

(exist? "foo[3].bar.baz")

will yield true if all of the following is true:
There is a member value of either group or string type named baz for some group value bar that is a member of the foo group with index 3. There may be multiple entries of bar within foo, only one needs to contain a value for baz.

Arguments:
ag-name - name of AutoGen value

3.4.10 `find-file' - locate a file in the search path

Usage: (find-file file-name [ suffix ])
AutoGen has a search path that it uses to locate template and definition files. This function will search the same list for `file-name', both with and without the `.suffix', if provided.

Arguments:
file-name - name of file with text
suffix - Optional - file suffix to try, too

3.4.11 `first-for?' - detect first iteration

Usage: (first-for? [ for_var ])
Returns SCM_BOOL_T if the named FOR loop (or, if not named, the current innermost loop) is on the first pass through the data. Outside of any FOR loop, it returns SCM_UNDEFINED. See section FOR - Emit a template block multiple times.

Arguments:
for_var - Optional - which for loop

3.4.12 `for-by' - set iteration step

Usage: (for-by by)
This function records the "step by" information for an AutoGen FOR function. Outside of the FOR macro itself, this function will emit an error. See section FOR - Emit a template block multiple times.

Arguments:
by - the iteration increment for the AutoGen FOR macro

3.4.13 `for-from' - set initial index

Usage: (for-from from)
This function records the initial index information for an AutoGen FOR function. Outside of the FOR macro itself, this function will emit an error. See section FOR - Emit a template block multiple times.

Arguments:
from - the initial index for the AutoGen FOR macro

3.4.14 `for-index' - get current loop index

Usage: (for-index [ for_var ])
Returns the current index for the named FOR loop. If not named, then the index for the innermost loop. Outside of any FOR loop, it returns SCM_UNDEFINED. See section FOR - Emit a template block multiple times.

Arguments:
for_var - Optional - which for loop

3.4.15 `for-sep' - set loop separation string

Usage: (for-sep separator)
This function records the separation string that is to be inserted between each iteration of an AutoGen FOR function. This is often nothing more than a comma. Outside of the FOR macro itself, this function will emit an error.

Arguments:
separator - the text to insert between the output of each FOR iteration

3.4.16 `for-to' - set ending index

Usage: (for-to to)
This function records the terminating value information for an AutoGen FOR function. Outside of the FOR macro itself, this function will emit an error. See section FOR - Emit a template block multiple times.

Arguments:
to - the final index for the AutoGen FOR macro

3.4.17 `get' - get named value

Usage: (get ag-name [ alt-val ])
Get the first string value associated with the name. It will either return the associated string value (if the name resolves), the alternate value (if one is provided), or else the empty string.

Arguments:
ag-name - name of AutoGen value
alt-val - Optional - value if not present

3.4.18 `high-lim' - get highest value index

Usage: (high-lim ag-name)
Returns the highest index associated with an array of definitions. This is generally, but not necessarily, one less than the count value. (The indexes may be specified, rendering a non-zero based or sparse array of values.)

This is very useful for specifying the size of a zero-based array of values where not all values are present. For example:

tMyStruct myVals[ [+ (+ 1 (high-lim "my-val-list")) +] ];

Arguments:
ag-name - name of AutoGen value

3.4.19 `last-for?' - detect last iteration

Usage: (last-for? [ for_var ])
Returns SCM_BOOL_T if the named FOR loop (or, if not named, the current innermost loop) is on the last pass through the data. Outside of any FOR loop, it returns SCM_UNDEFINED. See section FOR - Emit a template block multiple times.

Arguments:
for_var - Optional - which for loop

3.4.20 `len' - get count of values

Usage: (len ag-name)
If the named object is a group definition, then "len" is the same as "count". Otherwise, if it is one or more text definitions, then it is the sum of their string lengths. If it is a single text definition, then it is equivalent to (string-length (get "ag-name")).

Arguments:
ag-name - name of AutoGen value

3.4.21 `low-lim' - get lowest value index

Usage: (low-lim ag-name)
Returns the lowest index associated with an array of definitions.

Arguments:
ag-name - name of AutoGen value

3.4.22 `make-header-guard' - make self-inclusion guard

Usage: (make-header-guard name)
This function will create a #ifndef/#define sequence for protecting a header from multiple evaluation. It will also set the Scheme variable header-file to the name of the file being protected and it will set header-guard to the name of the #define being used to protect it. It is expected that this will be used as follows:

[+ (make-header-guard "group_name") +]
...
#endif /* [+ (. header-guard) +]

#include "[+ (. header-file)  +]"

The #define name is composed as follows:

1. The first element is the string argument and a separating underscore.
2. That is followed by the name of the header file with illegal characters mapped to underscores.
3. The end of the name is always, "_GUARD".
4. Finally, the entire string is mapped to upper case.

The final #define name is stored in an SCM symbol named header-guard. Consequently, the concluding #endif for the file should read something like:

#endif /* [+ (. header-guard) +] */

The name of the header file (the current output file) is also stored in an SCM symbol, header-file. Therefore, if you are also generating a C file that uses the previously generated header file, you can put this into that generated file:

#include "[+ (. header-file) +]"

Obviously, if you are going to produce more than one header file from a particular template, you will need to be careful how these SCM symbols get handled.

Arguments:
name - header group name

3.4.23 `match-value?' - test for matching value

Usage: (match-value? op ag-name test-str)
This function answers the question, "Is there an AutoGen value named ag-name with a value that matches the pattern test-str using the match function op?" Return SCM_BOOL_T iff at least one occurrence of the specified name has such a value. The operator can be any function that takes two string arguments and yields a boolean. It is expected that you will use one of the string matching functions provided by AutoGen.
The value name must follow the same rules as the ag-name argument for exist? (see section `exist?' - test for value name).

Arguments:
op - boolean result operator
ag-name - name of AutoGen value
test-str - string to test against

3.4.24 `out-delete' - delete current output file

Usage: (out-delete)
Remove the current output file. Cease processing the template for the current suffix. It is an error if there are push-ed output files. Use the (error "0") scheme function instead. See section Redirecting Output.

This Scheme function takes no arguments.

3.4.25 `out-depth' - output file stack depth

Usage: (out-depth)
Returns the depth of the output file stack. See section Redirecting Output.

This Scheme function takes no arguments.

3.4.26 `out-line' - output file line number

Usage: (out-line)
Returns the current line number of the output file. It rewinds and reads the file to count newlines.

This Scheme function takes no arguments.

3.4.27 `out-move' - change name of output file

Usage: (out-move new-name)
Rename current output file. See section Redirecting Output. Please note: changing the name will not save a temporary file from being deleted. It may, however, be used on the root output file.

Arguments:
new-name - new name for the current output file

3.4.28 `out-name' - current output file name

Usage: (out-name)
Returns the name of the current output file. If the current file is a temporary, unnamed file, then it will scan up the chain until a real output file name is found. See section Redirecting Output.

This Scheme function takes no arguments.

3.4.29 `out-pop' - close current output file

Usage: (out-pop [ disp ])
If there has been a push on the output, then close that file and go back to the previously open file. It is an error if there has not been a push. See section Redirecting Output.

If there is no argument, no further action is taken. Otherwise, the argument should be #t and the contents of the file are returned by the function.

Arguments:
disp - Optional - return contents of the file

3.4.30 `out-push-add' - append output to file

Usage: (out-push-add file-name)
Identical to push-new, except the contents are not purged, but appended to. See section Redirecting Output.

Arguments:
file-name - name of the file to append text to

3.4.31 `out-push-new' - purge and create output file

Usage: (out-push-new [ file-name ])
Leave the current output file open, but purge and create a new file that will remain open until a pop delete or switch closes it. The file name is optional and, if omitted, the output will be sent to a temporary file that will be deleted when it is closed. See section Redirecting Output.

Arguments:
file-name - Optional - name of the file to create

3.4.32 `out-resume' - resume suspended output file

Usage: (out-resume suspName)
If there has been a suspended output, then make that output descriptor current again. That output must have been suspended with the same tag name given to this routine as its argument.

Arguments:
suspName - A name tag for reactivating

3.4.33 `out-suspend' - suspend current output file

Usage: (out-suspend suspName)
If there has been a push on the output, then set aside the output descriptor for later reactiviation with (out-resume "xxx"). The tag name need not reflect the name of the output file. In fact, the output file may be an anonymous temporary file. You may also change the tag every time you suspend output to a file, because the tag names are forgotten as soon as the file has been "resumed".

Arguments:
suspName - A name tag for reactivating

3.4.34 `out-switch' - close and create new output

Usage: (out-switch file-name)
Switch output files - close current file and make the current file pointer refer to the new file. This is equivalent to out-pop followed by out-push-new, except that you may not pop the base level output file, but you may switch it. See section Redirecting Output.

Arguments:
file-name - name of the file to create

3.4.35 `set-option' - Set a command line option

Usage: (set-option opt)
The text argument must be an option name followed by any needed option argument. Returns SCM_UNDEFINED.

Arguments:
opt - AutoGen option name + its argument

3.4.36 `set-writable' - Make the output file be writable

Usage: (set-writable [ set? ])
This function will set the current output file to be writable (or not). This is only effective if neither the --writable nor --not-writable have been specified. This state is reset when the current suffix's output is complete.

Arguments:
set? - Optional - boolean arg, false to make output non-writable

3.4.37 `stack' - make list of AutoGen values

Usage: (stack ag-name)
Create a scheme list of all the strings that are associated with a name. They must all be text values or we choke.

Arguments:
ag-name - AutoGen value name

3.4.38 `suffix' - get the current suffix

Usage: (suffix)
Returns the current active suffix (see section Format of the Pseudo Macro).

This Scheme function takes no arguments.

3.4.39 `tpl-file' - get the template file name

Usage: (tpl-file [ full_path ])
Returns the name of the current template file. If #t is passed in as an argument, then the template file is hunted for in the template search path. Otherwise, just the unadorned name.

Arguments:
full_path - Optional - include full path to file

3.4.40 `tpl-file-line' - get the template file+line number

Usage: (tpl-file-line [ msg-fmt ])
Returns the file and line number of the current template macro using either the default format, "from %s line %d", or else the format you supply. For example, if you want to insert a "C" language file-line directive, you would supply the format "# %2$d \"%1$s\"", but that is also already supplied with the scheme variable See section format file info as, "#line nn "file"". You may use it thus:

(tpl-file-line c-file-line-fmt)

It is also safe to use the formatting string, "%2$d". AutoGen uses an argument vector version of printf: See section Replacement for Stdio Formatting Library.

Arguments:
msg-fmt - Optional - formatting for line message

3.4.41 `autogen-version' - autogen version number

This is a symbol defining the current AutoGen version number string. It was first defined in AutoGen-5.2.14. It is currently "5.8.6".

3.4.42 format file info as, "#line nn "file""

This is a symbol that can easily be used with the functions See section `tpl-file-line' - get the template file+line number, and See section `def-file-line' - get a definition file+line number. These will emit C program #line directives pointing to template and definitions text, respectively.

3.5 Common Scheme Functions

This section describes a number of general purpose functions that make the kind of string processing that AutoGen does a little easier. Unlike the AutoGen specific functions (see section AutoGen Scheme Functions), these functions are available for direct use during definition load time. The equality test (see section `string-eqv?' - caseless string match) is "overloaded" to do string equivalence comparisons. If you are looking for inequality, the Scheme/Lisp way of spelling that is, "(not (= ...))".

3.5.1 `ag-fprintf' - format to autogen stream

Usage: (ag-fprintf ag-diversion format [ format-arg ... ])
Format a string using arguments from the alist. Write to a specified AutoGen diversion. That may be either a specified suspended output stream (see section `out-suspend' - suspend current output file) or an index into the output stack (see section `out-push-new' - purge and create output file). (ag-fprintf 0 ...) is equivalent to (emit (sprintf ...)), and (ag-fprintf 1 ...) sends output to the most recently suspended output stream.

Arguments:
ag-diversion - AutoGen diversion name or number
format - formatting string
format-arg - Optional - list of arguments to formatting string

3.5.2 `bsd' - BSD Public License

Usage: (bsd prog_name owner prefix)
Emit a string that contains the Free BSD Public License. It takes three arguments: prefix contains the string to start each output line. owner contains the copyright owner. prog_name contains the name of the program the copyright is about.

Arguments:
prog_name - name of the program under the BSD
owner - Grantor of the BSD License
prefix - String for starting each output line

3.5.3 `c-string' - emit string for ANSI C

Usage: (c-string string)
Reform a string so that, when printed, the C compiler will be able to compile the data and construct a string that contains exactly what the current string contains. Many non-printing characters are replaced with escape sequences. Newlines are replaced with a backslash, an n, a closing quote, a newline, seven spaces and another re-opening quote. The compiler will implicitly concatenate them. The reader will see line breaks.

A K&R compiler will choke. Use kr-string for that compiler.

Arguments:
string - string to reformat

3.5.4 `emit' - emit the text for each argument

Usage: (emit alist ...)
Walk the tree of arguments, displaying the values of displayable SCM types.

Arguments:
alist - list of arguments to stringify and emit

3.5.5 `emit-string-table' - output a string table

Usage: (emit-string-table st-name)
Emit into the current output stream a static char const array named st-name that will have NUL bytes between each inserted string.

Arguments:
st-name - the name of the array of characters

3.5.6 `error-source-line' - display of file & line

Usage: (error-source-line)
This function is only invoked just before Guile displays an error message. It displays the file name and line number that triggered the evaluation error. You should not need to invoke this routine directly. Guile will do it automatically.

This Scheme function takes no arguments.

3.5.7 `extract' - extract text from another file

Usage: (extract file-name marker-fmt [ caveat ] [ default ])
This function is used to help construct output files that may contain text that is carried from one version of the output to the next.

The first two arguments are required, the second are optional:

• The file-name argument is used to name the file that contains the demarcated text.
• The marker-fmt is a formatting string that is used to construct the starting and ending demarcation strings. The sprintf function is given the marker-fmt with two arguments. The first is either "START" or "END". The second is either "DO NOT CHANGE THIS COMMENT" or the optional caveat argument.
• caveat is presumed to be absent if it is the empty string (""). If absent, "DO NOT CHANGE THIS COMMENT" is used as the second string argument to the marker-fmt.
• When a default argument is supplied and no pre-existing text is found, then this text will be inserted between the START and END markers.

The resulting strings are presumed to be unique within the subject file. As a simplified example:

[+ (extract "fname" "// %s - SOMETHING - %s" ""
"example default") +]

will result in the following text being inserted into the output:

// START - SOMETHING - DO NOT CHANGE THIS COMMENT
example default
// END   - SOMETHING - DO NOT CHANGE THIS COMMENT

The "example default" string can then be carried forward to the next generation of the output, provided the output is not named "fname" and the old output is renamed to "fname" before AutoGen-eration begins.

NB:

You can set aside previously generated source files inside the pseudo macro with a Guile/scheme function, extract the text you want to keep with this extract function. Just remember you should delete it at the end, too. Here is an example from my Finite State Machine generator:

[+ AutoGen5 Template  -*- Mode: text -*-
h=%s-fsm.h   c=%s-fsm.c
(shellf
"[ -f %1$s-fsm.h ] && mv -f %1$s-fsm.h .fsm.head
[ -f %1$s-fsm.c ] && mv -f %1$s-fsm.c .fsm.code" (base-name)) +]

This code will move the two previously produced output files to files named ".fsm.head" and ".fsm.code". At the end of the 'c' output processing, I delete them.

Arguments:
file-name - name of file with text
marker-fmt - format for marker text
caveat - Optional - warn about changing marker
default - Optional - default initial text

3.5.8 `format-arg-count' - count the args to a format

Usage: (format-arg-count format)
Sometimes, it is useful to simply be able to figure out how many arguments are required by a format string. For example, if you are extracting a format string for the purpose of generating a macro to invoke a printf-like function, you can run the formatting string through this function to determine how many arguments to provide for in the macro. e.g. for this extraction text:

 /*=fumble bumble
  * fmt: 'stumble %s: %d\n'
 =*/

You may wish to generate a macro:

 #define BUMBLE(a1,a2) printf_like(something,(a1),(a2))

You can do this by knowing that the format needs two arguments.

Arguments:
format - formatting string

3.5.9 `fprintf' - format to a file

Usage: (fprintf port format [ format-arg ... ])
Format a string using arguments from the alist. Write to a specified port. The result will NOT appear in your output. Use this to print information messages to a template user.

Arguments:
port - Guile-scheme output port
format - formatting string
format-arg - Optional - list of arguments to formatting string

3.5.10 `gperf' - perform a perfect hash function

Usage: (gperf name str)
Perform the perfect hash on the input string. This is only useful if you have previously created a gperf program with the make-gperf function See section `make-gperf' - build a perfect hash function program. The name you supply here must match the name used to create the program and the string to hash must be one of the strings supplied in the make-gperf string list. The result will be a perfect hash index.

See the documentation for gperf(1GNU) for more details.

Arguments:
name - name of hash list
str - string to hash

3.5.11 `gpl' - GNU General Public License

Usage: (gpl prog-name prefix)
Emit a string that contains the GNU General Public License. It takes two arguments: prefix contains the string to start each output line, and prog_name contains the name of the program the copyright is about.

Arguments:
prog-name - name of the program under the GPL
prefix - String for starting each output line

3.5.12 `hide-email' - convert eaddr to javascript

Usage: (hide-email display eaddr)
Hides an email address as a java scriptlett. The 'mailto:' tag and the email address are coded bytes rather than plain text. They are also broken up.

Arguments:
display - display text
eaddr - email address

3.5.13 `html-escape-encode' - encode html special characters

Usage: (html-escape-encode str)
This function will replace replace the characters '&', '<' and '>' characters with the HTML/XML escape-encoded strings ("&", "<", and ">", respectively).

Arguments:
str - string to make substitutions in

3.5.14 `in?' - test for string in list

Usage: (in? test-string string-list ...)
Return SCM_BOOL_T if the first argument string is found in one of the entries in the second (list-of-strings) argument.

Arguments:
test-string - string to look for
string-list - list of strings to check

3.5.15 `join' - join string list with separator

Usage: (join separator list ...)
With the first argument as the separator string, joins together an a-list of strings into one long string. The list may contain nested lists, partly because you cannot always control that.

Arguments:
separator - string to insert between entries
list - list of strings to join

3.5.16 `kr-string' - emit string for K&R C

Usage: (kr-string string)
Reform a string so that, when printed, a K&R C compiler will be able to compile the data and construct a string that contains exactly what the current string contains. Many non-printing characters are replaced with escape sequences. New-lines are replaced with a backslash-n-backslash and newline sequence,

Arguments:
string - string to reformat

3.5.17 `lgpl' - GNU Library General Public License

Usage: (lgpl prog_name owner prefix)
Emit a string that contains the GNU Library General Public License. It takes three arguments: prefix contains the string to start each output line. owner contains the copyright owner. prog_name contains the name of the program the copyright is about.

Arguments:
prog_name - name of the program under the LGPL
owner - Grantor of the LGPL
prefix - String for starting each output line

3.5.18 `license' - an arbitrary license

Usage: (license lic_name prog_name owner prefix)
Emit a string that contains the named license. The license text is read from a file named, lic_name.lic, searching the standard directories. The file contents are used as a format argument to printf(3), with prog_name and owner as the two string formatting arguments. Each output line is automatically prefixed with the string prefix.

Arguments:
lic_name - file name of the license
prog_name - name of the licensed program or library
owner - Grantor of the License
prefix - String for starting each output line

3.5.19 `make-gperf' - build a perfect hash function program

Usage: (make-gperf name strings ...)
Build a program to perform perfect hashes of a known list of input strings. This function produces no output, but prepares a program named, `gperf_<name>' for use by the gperf function See section `gperf' - perform a perfect hash function.

This program will be obliterated as AutoGen exits. However, you may incorporate the generated hashing function into your C program with commands something like the following:

[+ (shellf "sed '/^int main(/,$d;/^#line/d' ${gpdir}/%s.c"
name ) +]

where name matches the name provided to this make-perf function. gpdir is the variable used to store the name of the temporary directory used to stash all the files.

Arguments:
name - name of hash list
strings - list of strings to hash

3.5.20 `makefile-script' - create makefile script

Usage: (makefile-script text)
This function will take ordinary shell script text and reformat it so that it will work properly inside of a makefile shell script. Not every shell construct can be supported; the intent is to have most ordinary scripts work without much, if any, alteration.

The following transformations are performed on the source text:

1. Trailing whitespace on each line is stripped.
2. Except for the last line, the string, " ; \\" is appended to the end of every line that does not end with a backslash, semi-colon, conjunction operator or pipe. Note that this will mutilate multi-line quoted strings, but make renders it impossible to use multi-line constructs anyway.
3. If the line ends with a backslash, it is left alone.
4. If the line ends with one of the excepted operators, then a space and backslash is added.
5. The dollar sign character is doubled, unless it immediately precedes an opening parenthesis or the single character make macros '*', '<', '@', '?' or '%'. Other single character make macros that do not have enclosing parentheses will fail. For shell usage of the "$@", "$?" and "$*" macros, you must enclose them with curly braces, e.g., "${?}". The ksh construct $(<command>) will not work. Though some makes accept ${var} constructs, this function will assume it is for shell interpretation and double the dollar character. You must use $(var) for all make substitutions.
6. Double dollar signs are replaced by four before the next character is examined.
7. Every line is prefixed with a tab, unless the first line already starts with a tab.
8. The newline character on the last line, if present, is suppressed.
9. Blank lines are stripped.

This function is intended to be used approximately as follows:

$(TARGET) : $(DEPENDENCIES)
<+ (out-push-new) +>
....mostly arbitrary shell script text....
<+ (makefile-script (out-pop #t)) +>

Arguments:
text - the text of the script

3.5.21 `max' - maximum value in list

Usage: (max list ...)
Return the maximum value in the list

Arguments:
list - list of values. Strings are converted to numbers

3.5.22 `min' - minimum value in list

Usage: (min list ...)
Return the minimum value in the list

Arguments:
list - list of values. Strings are converted to numbers

3.5.23 `prefix' - prefix lines with a string

Usage: (prefix prefix text)
Prefix every line in the second string with the first string.

For example, if the first string is "# " and the second contains:

two
lines

The result string will contain:

# two
# lines

Arguments:
prefix - string to insert at start of each line
text - multi-line block of text

3.5.24 `printf' - format to stdout

Usage: (printf format [ format-arg ... ])
Format a string using arguments from the alist. Write to the standard out port. The result will NOT appear in your output. Use this to print information messages to a template user. Use "(sprintf ...)" to add text to your document.

Arguments:
format - formatting string
format-arg - Optional - list of arguments to formatting string

3.5.25 `raw-shell-str' - single quote shell string

Usage: (raw-shell-str string)
Convert the text of the string into a singly quoted string that a normal shell will process into the original string. (It will not do macro expansion later, either.) Contained single quotes become tripled, with the middle quote escaped with a backslash. Normal shells will reconstitute the original string.

Notice: some shells will not correctly handle unusual non-printing characters. This routine works for most reasonably conventional ASCII strings.

Arguments:
string - string to transform

3.5.26 `shell' - invoke a shell script

Usage: (shell command)
Generate a string by writing the value to a server shell and reading the output back in. The template programmer is responsible for ensuring that it completes within 10 seconds. If it does not, the server will be killed, the output tossed and a new server started.

Arguments:
command - shell command - the result value is stdout

3.5.27 `shell-str' - double quote shell string

Usage: (shell-str string)
Convert the text of the string into a double quoted string that a normal shell will process into the original string, almost. It will add the escape character \\ before two special characters to accomplish this: the backslash \\ and double quote ".

NOTE: some shells will not correctly handle unusual non-printing characters. This routine works for most reasonably conventional ASCII strings.

WARNING:
This function omits the extra backslash in front of a backslash, however, if it is followed by either a backquote or a dollar sign. It must do this because otherwise it would be impossible to protect the dollar sign or backquote from shell evaluation. Consequently, it is not possible to render the strings "\\$" or "\\`". The lesser of two evils.

All others characters are copied directly into the output.

The sub-shell-str variation of this routine behaves identically, except that the extra backslash is omitted in front of " instead of `. You have to think about it. I'm open to suggestions.

Meanwhile, the best way to document is with a detailed output example. If the backslashes make it through the text processing correctly, below you will see what happens with three example strings. The first example string contains a list of quoted foos, the second is the same with a single backslash before the quote characters and the last is with two backslash escapes. Below each is the result of the raw-shell-str, shell-str and sub-shell-str functions.

foo[0]           ''foo'' 'foo' "foo" `foo` $foo
raw-shell-str -> \'\''foo'\'\'' '\''foo'\'' "foo" `foo` $foo'
shell-str     -> "''foo'' 'foo' \"foo\" `foo` $foo"
sub-shell-str -> `''foo'' 'foo' "foo" \`foo\` $foo`

foo[1]           \'bar\' \"bar\" \`bar\` \$bar
raw-shell-str -> '\'\''bar\'\'' \"bar\" \`bar\` \$bar'
shell-str     -> "\\'bar\\' \\\"bar\\\" \`bar\` \$bar"
sub-shell-str -> `\\'bar\\' \"bar\" \\\`bar\\\` \$bar`

foo[2]           \\'BAZ\\' \\"BAZ\\" \\`BAZ\\` \\$BAZ
raw-shell-str -> '\\'\''BAZ\\'\'' \\"BAZ\\" \\`BAZ\\` \\$BAZ'
shell-str     -> "\\\\'BAZ\\\\' \\\\\"BAZ\\\\\" \\\`BAZ\\\` \\\$BAZ"
sub-shell-str -> `\\\\'BAZ\\\\' \\\"BAZ\\\" \\\\\`BAZ\\\\\` \\\$BAZ`

There should be four, three, five and three backslashes for the four examples on the last line, respectively. The next to last line should have four, five, three and three backslashes. If this was not accurately reproduced, take a look at the agen5/test/shell.test test. Notice the backslashes in front of the dollar signs. It goes from zero to one to three for the "cooked" string examples.

Arguments:
string - string to transform

3.5.28 `shellf' - format a string, run shell

Usage: (shellf format [ format-arg ... ])
Format a string using arguments from the alist, then send the result to the shell for interpretation.

Arguments:
format - formatting string
format-arg - Optional - list of arguments to formatting string

3.5.29 `sprintf' - format a string

Usage: (sprintf format [ format-arg ... ])
Format a string using arguments from the alist.

Arguments:
format - formatting string
format-arg - Optional - list of arguments to formatting string

3.5.30 `string-capitalize' - capitalize a new string

Usage: (string-capitalize str)
Create a new SCM string containing the same text as the original, only all the first letter of each word is upper cased and all other letters are made lower case.

Arguments:
str - input string

3.5.31 `string-capitalize!' - capitalize a string

Usage: (string-capitalize! str)
capitalize all the words in an SCM string.

Arguments:
str - input/output string

3.5.32 `string-contains-eqv?' - caseless substring

Usage: (*=* text match)
string-contains-eqv?: Test to see if a string contains an equivalent string. `equivalent' means the strings match, but without regard to character case and certain characters are considered `equivalent'. Viz., '-', '_' and '^' are equivalent.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.33 `string-contains?' - substring match

Usage: (*==* text match)
string-contains?: Test to see if a string contains a substring. "strstr(3)" will find an address.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.34 `string-downcase' - lower case a new string

Usage: (string-downcase str)
Create a new SCM string containing the same text as the original, only all the upper case letters are changed to lower case.

Arguments:
str - input string

3.5.35 `string-downcase!' - make a string be lower case

Usage: (string-downcase! str)
Change to lower case all the characters in an SCM string.

Arguments:
str - input/output string

3.5.36 `string-end-eqv-match?' - caseless regex ending

Usage: (*~ text match)
string-end-eqv-match?: Test to see if a string ends with a pattern. Case is not significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.37 `string-end-match?' - regex match end

Usage: (*~~ text match)
string-end-match?: Test to see if a string ends with a pattern. Case is significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.38 `string-ends-eqv?' - caseless string ending

Usage: (*= text match)
string-ends-eqv?: Test to see if a string ends with an equivalent string.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.39 `string-ends-with?' - string ending

Usage: (*== text match)
string-ends-with?: Test to see if a string ends with a substring. strcmp(3) returns zero for comparing the string ends.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.40 `string-equals?' - string matching

Usage: (== text match)
string-equals?: Test to see if two strings exactly match.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.41 `string-eqv-match?' - caseless regex match

Usage: (~ text match)
string-eqv-match?: Test to see if a string fully matches a pattern. Case is not significant, but any character equivalences must be expressed in your regular expression.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.42 `string-eqv?' - caseless string match

Usage: (= text match)
string-eqv?: Test to see if two strings are equivalent. `equivalent' means the strings match, but without regard to character case and certain characters are considered `equivalent'. Viz., '-', '_' and '^' are equivalent. If the arguments are not strings, then the result of the numeric comparison is returned.

This is an overloaded operation. If the arguments are not both strings, then the query is passed through to scm_num_eq_p().

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.43 `string-has-eqv-match?' - caseless regex contains

Usage: (*~* text match)
string-has-eqv-match?: Test to see if a string contains a pattern. Case is not significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.44 `string-has-match?' - contained regex match

Usage: (*~~* text match)
string-has-match?: Test to see if a string contains a pattern. Case is significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.45 `string-match?' - regex match

Usage: (~~ text match)
string-match?: Test to see if a string fully matches a pattern. Case is significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.46 `string-start-eqv-match?' - caseless regex start

Usage: (~* text match)
string-start-eqv-match?: Test to see if a string starts with a pattern. Case is not significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.47 `string-start-match?' - regex match start

Usage: (~~* text match)
string-start-match?: Test to see if a string starts with a pattern. Case is significant.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.48 `string-starts-eqv?' - caseless string start

Usage: (=* text match)
string-starts-eqv?: Test to see if a string starts with an equivalent string.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.49 `string-starts-with?' - string starting

Usage: (==* text match)
string-starts-with?: Test to see if a string starts with a substring.

Arguments:
text - text to test for pattern
match - pattern/substring to search for

3.5.50 `string-substitute' - multiple global replacements

Usage: (string-substitute source match repl)
match and repl may be either a single string or a list of strings. Either way, they must have the same structure and number of elements. For example, to replace all less than and all greater than characters, do something like this:

(string-substitute source
("&"     "<"    ">")
("&amp;" "&lt;" "&gt;"))

Arguments:
source - string to transform
match - substring or substring list to be replaced
repl - replacement strings or substrings

3.5.51 `string-table-add' - Add an entry to a string table

Usage: (string-table-add st-name str-val)
Check for a duplicate string and, if none, then insert a new string into the string table. In all cases, returns the character index of the beginning of the string in the table.

The returned index can be used in expressions like:

string_array + <returned-value>

that will yield the address of the first byte of the inserted string. See the `strtable.test' AutoGen test for a usage example.

Arguments:
st-name - the name of the array of characters
str-val - the (possibly) new value to add

3.5.52 `string-table-new' - create a string table

Usage: (string-table-new st-name)
This function will create an array of characters. The companion functions, (See section `string-table-add' - Add an entry to a string table, and see section `emit-string-table' - output a string table) will insert text and emit the populated table, respectively.

With these functions, it should be much easier to construct structures containing string offsets instead of string pointers. That can be very useful when transmitting, storing or sharing data with different address spaces.

Here is a brief example copied from the strtable.test test:

[+ (string-table-new "scribble")
   (out-push-new)
   (define ix 0)
   (define ct 1)  +][+

FOR str IN that was the week that was +][+
  (set! ct (+ ct 1))
  (set! ix (string-table-add "scribble" (get "str")))
+]
    scribble + [+ (. ix) +],[+
ENDFOR  +]
    NULL };
[+ (out-suspend "main")
   (emit-string-table "scribble")
   (ag-fprintf 0 "\nchar const *ap[%d] = {" ct)
   (out-resume "main")
   (out-pop #t) +]

Some explanation:

I added the (out-push-new) because the string table text is diverted into an output stream named, "scribble" and I want to have the string table emitted before the string table references. The string table references are also emitted inside the FOR loop. So, when the loop is done, the current output is suspended under the name, "main" and the "scribble" table is then emitted into the primary output. (emit-string-table inserts its output directly into the current output stream. It does not need to be the last function in an AutoGen macro block.) Next I ag-fprintf the array-of-pointer declaration directly into the current output. Finally I restore the "main" output stream and (out-pop #t)-it into the main output stream.

Here is the result. Note that duplicate strings are not repeated in the string table:

static char const scribble[18] =
    "that\0" "was\0"  "the\0"  "week\0";

char const *ap[7] = {
    scribble + 0,
    scribble + 5,
    scribble + 9,
    scribble + 13,
    scribble + 0,
    scribble + 5,
    NULL };

These functions use the global name space stt-* in addition to the function names.

Arguments:
st-name - the name of the array of characters

3.5.53 `string->c-name!' - map non-name chars to underscore

Usage: (string->c-name! str)
Change all the graphic characters that are invalid in a C name token into underscores. Whitespace characters are ignored. Any other character type (i.e. non-graphic and non-white) will cause a failure.

Arguments:
str - input/output string

3.5.54 `string-tr' - convert characters with new result

Usage: (string-tr source match translation)
This is identical to string-tr!, except that it does not over-write the previous value.

Arguments:
source - string to transform
match - characters to be converted
translation - conversion list

3.5.55 `string-tr!' - convert characters

Usage: (string-tr! source match translation)
This is the same as the tr(1) program, except the string to transform is the first argument. The second and third arguments are used to construct mapping arrays for the transformation of the first argument.

It is too bad this little program has so many different and incompatible implementations!

Arguments:
source - string to transform
match - characters to be converted
translation - conversion list

3.5.56 `string-upcase' - upper case a new string

Usage: (string-upcase str)
Create a new SCM string containing the same text as the original, only all the lower case letters are changed to upper case.

Arguments:
str - input string

3.5.57 `string-upcase!' - make a string be upper case

Usage: (string-upcase! str)
Change to upper case all the characters in an SCM string.

Arguments:
str - input/output string

3.5.58 `sub-shell-str' - back quoted (sub-)shell string

Usage: (sub-shell-str string)
This function is substantially identical to shell-str, except that the quoting character is ` and the "leave the escape alone" character is ".

Arguments:
string - string to transform

3.5.59 `sum' - sum of values in list

Usage: (sum list ...)
Compute the sum of the list of expressions.

Arguments:
list - list of values. Strings are converted to numbers

3.5.60 `version-compare' - compare two version numbers

Usage: (version-compare op v1 v2)
Converts v1 and v2 strings into 64 bit values and returns the result of running 'op' on those values. It assumes that the version is a 1 to 4 part dot-separated series of numbers. Suffixes like, "5pre4" or "5-pre4" will be interpreted as two numbers. The first number ("5" in this case) will be decremented and the number after the "pre" will be added to 0xC000. (Unless your platform is unable to support 64 bit integer arithmetic. Then it will be added to 0xC0.) Consequently, these yield true:

(version-compare > "5.8.5"       "5.8.5-pre4")
(version-compare > "5.8.5-pre10" "5.8.5-pre4")

Arguments:
op - comparison operator
v1 - first version
v2 - compared-to version

3.6 AutoGen Native Macros

This section describes the various AutoGen natively defined macros. Unlike the Scheme functions, some of these macros are "block macros" with a scope that extends through a terminating macro. Block macros must not overlap. That is to say, a block macro started within the scope of an encompassing block macro must have its matching end macro appear before the encompassing block macro is either ended or subdivided.

The block macros are these:

CASE

This macro has scope through the ESAC macro. The scope is subdivided by SELECT macros. You must have at least one SELECT macro.

DEFINE

This macro has scope through the ENDDEF macro. The defined user macro can never be a block macro. This macro is extracted from the template before the template is processed.

FOR

This macro has scope through the ENDFOR macro.

IF

This macro has scope through the ENDIF macro. The scope may be subdivided by ELIF and ELSE macros. Obviously, there may be only one ELSE macro and it must be the last of these subdivisions.

INCLUDE

This macro has the scope of the included file. It is a block macro in the sense that the included file must not contain any incomplete block macros.

WHILE

This macro has scope through the ENDWHILE macro.

3.6.1 AutoGen Macro Syntax

The general syntax is:

[ { <native-macro-name> | <user-defined-name> } ] [ <arg> ... ]

The syntax for <arg> depends on the particular macro, but is generally a full expression (see section Macro Expression Syntax). Here are the exceptions to that general rule:

1. INVOKE macros, implicit or explicit, must be followed by a list of name/string value pairs. The string values are simple expressions, as described above.

   That is, the INVOKE syntax is one of these two:

   <user-macro-name> [ <name> [ = <expression> ] ... ] INVOKE <name-expression> [ <name> [ = <expression> ] ... ]

2. AutoGen FOR macros must be in one of three forms:

   FOR <name> [ <separator-string> ] FOR <name> (...Scheme expression list) FOR <name> IN <string-entry> [ ... ]

   where:

   `<name>'

   must be a simple name.

   `<separator-string>'

   is inserted between copies of the enclosed block. Do not try to use "IN" as your separator string. It won't work.

   `<string-entry>'

   is an entry in a list of strings. "<name>" is assigned each value from the "IN" list before expanding the FOR block.

   `(...Scheme expression list)'

   is expected to contain one or more of the for-from, for-to, for-by, and for-sep functions. (See section FOR - Emit a template block multiple times, and AutoGen Scheme Functions)

   The first two forms iterate over the FOR block if <name> is found in the AutoGen values. The last form will create the AutoGen value named <name>.

3. AutoGen DEFINE macros must be followed by a simple name. Anything after that is ignored. Consequently, that "comment space" may be used to document any named values the macro expects to have set up as arguments. See section DEFINE - Define a user AutoGen macro.
4. The AutoGen COMMENT, ELSE, ESAC and the END* macros take no arguments and ignore everything after the macro name (e.g. see COMMENT - A block of comment to be ignored)

3.6.2 CASE - Select one of several template blocks

The arguments are evaluated and converted to a string, if necessary. A simple name will be interpreted as an AutoGen value name and its value will be used by the SELECT macros (see the example below and the expression evaluation function, see section EXPR - Evaluate and emit an Expression). The scope of the macro is up to the matching ESAC macro. Within the scope of a CASE, this string is matched against case selection macros. There are sixteen match macros that are derived from four different ways matches may be performed, plus an "always true", "true if the AutoGen value was found", and "true if no AutoGen value was found" matches. The codes for the nineteen match macros are formed as follows:

1. Must the match start matching from the beginning of the string? If not, then the match macro code starts with an asterisk (*).
2. Must the match finish matching at the end of the string? If not, then the match macro code ends with an asterisk (*).
3. Is the match a pattern match or a string comparison? If a comparison, use an equal sign (=). If a pattern match, use a tilde (~).
4. Is the match case sensitive? If alphabetic case is important, double the tilde or equal sign.
5. Do you need a default match when none of the others match? Use a single asterisk (*).
6. Do you need to distinguish between an empty string value and a value that was not found? Use the non-existence test (!E) before testing a full match against an empty string (== ''). There is also an existence test (+E), more for symmetry than for practical use.

For example:

[+ CASE <full-expression> +]
[+ ~~*  "[Tt]est" +]reg exp must match at start, not at end
[+ ==   "TeSt"    +]a full-string, case sensitive compare
[+ =    "TEST"    +]a full-string, case insensitive compare
[+ !E             +]not exists - matches if no AutoGen value found
[+ ==   ""        +]expression yielded a zero-length string
[+ +E             +]exists - matches if there is any value result
[+ *              +]always match - no testing
[+ ESAC +]

<full-expression> (see section Macro Expression Syntax) may be any expression, including the use of apply-codes and value-names. If the expression yields a number, it is converted to a decimal string.

These case selection codes have also been implemented as Scheme expression functions using the same codes. They are documented in this texi doc as "string-*?" predicates (see section Common Scheme Functions).

3.6.3 COMMENT - A block of comment to be ignored

This function can be specified by the user, but there will never be a situation where it will be invoked at emit time. The macro is actually removed from the internal representation.

If the native macro name code is #, then the entire macro function is treated as a comment and ignored.

3.6.4 DEFINE - Define a user AutoGen macro

This function will define a new macro. You must provide a name for the macro. You do not specify any arguments, though the invocation may specify a set of name/value pairs that are to be active during the processing of the macro.

[+ define foo +]
... macro body with macro functions ...
[+ enddef +]
... [+ foo bar='raw text' baz=<<text expression>> +]

Once the macro has been defined, this new macro can be invoked by specifying the macro name as the first token after the start macro marker. Alternatively, you may make the invocation explicitly invoke a defined macro by specifying INVOKE (see section INVOKE - Invoke a User Defined Macro) in the macro invocation. If you do that, the macro name can be computed with an expression that gets evaluated every time the INVOKE macro is encountered.

Any remaining text in the macro invocation will be used to create new name/value pairs that only persist for the duration of the processing of the macro. The expressions are evaluated the same way basic expressions are evaluated. See section Macro Expression Syntax.

The resulting definitions are handled much like regular definitions, except:

1. The values may not be compound. That is, they may not contain nested name/value pairs.
2. The bindings go away when the macro is complete.
3. The name/value pairs are separated by whitespace instead of semi-colons.
4. Sequences of strings are not concatenated.

NB: The macro is extracted from the template as the template is scanned. You cannot conditionally define a macro by enclosing it in an IF/ENDIF (see section IF - Conditionally Emit a Template Block) macro pair. If you need to dynamically select the format of a DEFINEd macro, then put the flavors into separate template files that simply define macros. INCLUDE (see section INCLUDE - Read in and emit a template block) the appropriate template when you have computed which you need.

Due to this, it is acceptable and even a good idea to place all the DEFINE macros at the end of the template. That puts the main body of the template at the beginning of the file.

3.6.5 ELIF - Alternate Conditional Template Block

This macro must only appear after an IF function, and before any associated ELSE or ENDIF functions. It denotes the start of an alternate template block for the IF function. Its expression argument is evaluated as are the arguments to IF. For a complete description See section IF - Conditionally Emit a Template Block.

3.6.6 ELSE - Alternate Template Block

This macro must only appear after an IF function, and before the associated ENDIF function. It denotes the start of an alternate template block for the IF function. For a complete description See section IF - Conditionally Emit a Template Block.

3.6.7 ENDDEF - Ends a macro definition.

This macro ends the DEFINE function template block. For a complete description See section DEFINE - Define a user AutoGen macro.

3.6.8 ENDFOR - Terminates the FOR function template block

This macro ends the FOR function template block. For a complete description See section FOR - Emit a template block multiple times.

3.6.9 ENDIF - Terminate the IF Template Block

This macro ends the IF function template block. For a complete description See section IF - Conditionally Emit a Template Block.

3.6.10 ENDWHILE - Terminate the WHILE Template Block

This macro ends the WHILE function template block. For a complete description See section WHILE - Conditionally loop over a Template Block.

3.6.11 ESAC - Terminate the CASE Template Block

This macro ends the CASE function template block. For a complete description, See section CASE - Select one of several template blocks.

3.6.12 EXPR - Evaluate and emit an Expression

This macro does not have a name to cause it to be invoked explicitly, though if a macro starts with one of the apply codes or one of the simple expression markers, then an expression macro is inferred. The result of the expression evaluation (see section Macro Expression Syntax) is written to the current output.

3.6.13 FOR - Emit a template block multiple times

This macro has a slight variation on the standard syntax:

FOR <value-name> [ <separator-string> ]

FOR <value-name> (...Scheme expression list)

FOR <value-name> IN "string" [ ... ]

Other than for the last form, the first macro argument must be the name of an AutoGen value. If there is no value associated with the name, the FOR template block is skipped entirely. The scope of the FOR macro extends to the corresponding ENDFOR macro. The last form will create an array of string values named <value-name> that only exists within the context of this FOR loop. With this form, in order to use a separator-string, you must code it into the end of the template block using the (last-for?) predicate function (see section `last-for?' - detect last iteration).

If there are any arguments after the value-name, the initial characters are used to determine the form. If the first character is either a semi-colon (;) or an opening parenthesis ((), then it is presumed to be a Scheme expression containing the FOR macro specific functions for-from, for-by, for-to, and/or for-sep. See section AutoGen Scheme Functions. If it consists of an 'i' an 'n' and separated by white space from more text, then the FOR x IN form is processed. Otherwise, the remaining text is presumed to be a string for inserting between each iteration of the loop. This string will be emitted one time less than the number of iterations of the loop. That is, it is emitted after each loop, excepting for the last iteration.

If the from/by/to functions are invoked, they will specify which copies of the named value are to be processed. If there is no copy of the named value associated with a particular index, the FOR template block will be instantiated anyway. The template must use methods for detecting missing definitions and emitting default text. In this fashion, you can insert entries from a sparse or non-zero based array into a dense, zero based array.

NB: the for-from, for-to, for-by and for-sep functions are disabled outside of the context of the FOR macro. Likewise, the first-for, last-for and for-index functions are disabled outside of the range of a FOR block.

Also: the <value-name> must be a single level name, not a compound name (see section Naming a value).

[+FOR var (for-from 0) (for-to <number>) (for-sep ",") +]
... text with various substitutions ...[+
ENDFOR var+]

this will repeat the ... text with various substitutions ... <number>+1 times. Each repetition, except for the last, will have a comma , after it.

[+FOR var ",\n" +]
... text with various substitutions ...[+
ENDFOR var +]

This will do the same thing, but only for the index values of var that have actually been defined.

3.6.14 IF - Conditionally Emit a Template Block

Conditional block. Its arguments are evaluated (see section EXPR - Evaluate and emit an Expression) and if the result is non-zero or a string with one or more bytes, then the condition is true and the text from that point until a matched ELIF, ELSE or ENDIF is emitted. ELIF introduces a conditional alternative if the IF clause evaluated FALSE and ELSE introduces an unconditional alternative.

[+IF <full-expression> +]
emit things that are for the true condition[+

ELIF <full-expression-2> +]
emit things that are true maybe[+

ELSE "This may be a comment" +]
emit this if all but else fails[+

ENDIF "This may *also* be a comment" +]

<full-expression> may be any expression described in the EXPR expression function, including the use of apply-codes and value-names. If the expression yields an empty string, it is interpreted as false.

3.6.15 INCLUDE - Read in and emit a template block

The entire contents of the named file is inserted at this point. The contents of the file are processed for macro expansion. The arguments are eval-ed, so you may compute the name of the file to be included. The included file must not contain any incomplete function blocks. Function blocks are template text beginning with any of the macro functions `CASE', `DEFINE', `FOR', `IF' and `WHILE'; extending through their respective terminating macro functions.

3.6.16 INVOKE - Invoke a User Defined Macro

User defined macros may be invoked explicitly or implicitly. If you invoke one implicitly, the macro must begin with the name of the defined macro. Consequently, this may not be a computed value. If you explicitly invoke a user defined macro, the macro begins with the macro name INVOKE followed by a basic expression that must yield a known user defined macro. A macro name _must_ be found, or AutoGen will issue a diagnostic and exit.

Arguments are passed to the invoked macro by name. The text following the macro name must consist of a series of names each of which is followed by an equal sign (=) and a basic expression that yields a string.

The string values may contain template macros that are parsed the first time the macro is processed and evaluated again every time the macro is evaluated.

3.6.17 SELECT - Selection block for CASE function

This macro selects a block of text by matching an expression against the sample text expression evaluated in the CASE macro. See section CASE - Select one of several template blocks.

You do not specify a SELECT macro with the word "select". Instead, you must use one of the 19 match operators described in the CASE macro description.

3.6.18 UNKNOWN - Either a user macro or a value name.

The macro text has started with a name not known to AutoGen. If, at run time, it turns out to be the name of a defined macro, then that macro is invoked. If it is not, then it is a conditional expression that is evaluated only if the name is defined at the time the macro is invoked.

You may not specify UNKNOWN explicitly.

3.6.19 WHILE - Conditionally loop over a Template Block

Conditionally repeated block. Its arguments are evaluated (see section EXPR - Evaluate and emit an Expression) and as long as the result is non-zero or a string with one or more bytes, then the condition is true and the text from that point until a matched ENDWHILE is emitted.

[+WHILE <full-expression> +]
emit things that are for the true condition[+

ENDWHILE +]

<full-expression> may be any expression described in the EXPR expression function, including the use of apply-codes and value-names. If the expression yields an empty string, it is interpreted as false.

3.7 Redirecting Output

AutoGen provides a means for redirecting the template output to different files or, in `M4' parlance, to various diversions. It is accomplished by providing a set of Scheme functions named out-* (see section AutoGen Scheme Functions).

`out-push-new (see section `out-push-new' - purge and create output file)'

This allows you to logically "push" output files onto a stack. If you supply a string name, then a file by that name is created to hold the output. If you do not supply a name, then the text is written to a scratch pad and retrieved by passing a "#t" argument to the out-pop (see section `out-pop' - close current output file) function.

`out-pop (see section `out-pop' - close current output file)'

This function closes the current output file and resumes output to the next one in the stack. At least one output must have been pushed onto the output stack with the out-push-new (see section `out-push-new' - purge and create output file) function. If "#t" is passed in as an argument, then the entire contents of the diversion (or file) is returned.

`out-suspend (see section `out-suspend' - suspend current output file)'

This function does not close the current output, but instead sets it aside for resumption by the given name with out-resume. The current output must have been pushed on the output queue with out-push-new (see section `out-push-new' - purge and create output file).

`out-resume (see section `out-resume' - resume suspended output file)'

This will put a named file descriptor back onto the top of stack so that it becomes the current output again.

`out-switch (see section `out-switch' - close and create new output)'

This closes the current output and creates a new file, purging any preexisting one. This is a shortcut for "pop" followed by "push", but this can also be done at the base level.

`out-move (see section `out-move' - change name of output file)'

Renames the current output file without closing it.

There are also several functions for determining the output status. See section AutoGen Scheme Functions.

4. Augmenting AutoGen Features

AutoGen was designed to be simple to enhance. You can do it by providing shell commands, Guile/Scheme macros or callout functions that can be invoked as a Guile macro. Here is how you do these.

4.1 Shell Output Commands

Shell commands are run inside of a server process. This means that, unlike `make', context is kept from one command to the next. Consequently, you can define a shell function in one place inside of your template and invoke it in another. You may also store values in shell variables for later reference. If you load functions from a file containing shell functions, they will remain until AutoGen exits.

If your shell script should determine that AutoGen should stop processing, the recommended method for stopping AutoGen is:

die "some error text"

That is a shell function added by AutoGen. It will send a SIGTERM to autogen and exit from the "persistent" shell.

4.2 Guile Macros

Guile also maintains context from one command to the next. This means you may define functions and variables in one place and reference them elsewhere. You also may load Guile macro definitions from a Scheme file by using the --load-scheme command line option (see section load-scheme option (-S)). Beware, however, that the AutoGen specific scheme functions have not been loaded at this time, so though you may define functions that reference them, do not invoke the AutoGen functions at this time.

If your Scheme script should determine that AutoGen should stop processing, the recommended method for stopping AutoGen is:

(error "some error text")

4.3 Guile Callout Functions

Callout functions must be registered with Guile to work. This can be accomplished either by putting your routines into a shared library that contains a void scm_init( void ) routine that registers these routines, or by building them into AutoGen.

To build them into AutoGen, you must place your routines in the source directory and name the files `exp*.c'. You also must have a stylized comment that `getdefs' can find that conforms to the following:

/*=gfunc <function-name>
 *
 *  what:    <short one-liner>
 *  general_use:
 *  string:  <invocation-name-string>
 *  exparg:  <name>, <description> [, ['optional'] [, 'list']]
 *  doc:     A long description telling people how to use
 *           this function.
=*/
SCM
ag_scm_<function-name>( SCM arg_name[, ...] )
{ <code> }

`gfunc'

You must have this exactly thus.

`<function-name>'

This must follow C syntax for variable names

`<short one-liner>'

This should be about a half a line long. It is used as a subsection title in this document.

`general_use:'

You must supply this unless you are an AutoGen maintainer and are writing a function that queries or modifies the state of AutoGen.

`<invocation-name-string>'

Normally, the function-name string will be transformed into a reasonable invocation name. However, that is not always true. If the result does not suit your needs, then supply an alternate string.

`exparg:'

You must supply one for each argument to your function. All optional arguments must be last. The last of the optional arguments may be a list, if you choose.

`doc:'

Please say something meaningful.

`[, ...]'

Do not actually specify an ANSI ellipsis here. You must provide for all the arguments you specified with exparg.

See the Guile documentation for more details. More information is also available in a large comment at the beginning of the `agen5/snarf.tpl' template file.