Provided by: yodl_4.03.03-1_amd64
NAME
yodlbuiltins - Builtins for the Yodl converters
SYNOPSIS
This manual page lists the standard builtins of the Yodl package.
DESCRIPTION
The following list shows the builtins defined by the Yodl converters define and which can be used in Yodl documents. Refer to the Yodl user guide, distributed with the Yodl package, for a full description. The following list shows all builtins of the package in alphabetical order. `Yodl’s builtin commands’ As mentioned previously, YODL’s input consists of text and of commands. YODL supports a number of built-in commands which may either be used in a YODL document, or which can be used to create a macro package. Don’t despair if you find that the description of this section is too technical. Exactly for this reason, YODL supports the macro packages to make the life of a documentation writer easier. E.g., see chapter `[MACROPACKAGE]’ that describes a macro package for YODL. Most built-in functions and macros expand the information they receive the way they receive the information. I.e., the information itself is only evaluated by the time it is eventually inserted into an output medium (usually a file). However, some builtin functions evaluate their argument(s) once the argument is processed. They are: o The `ERROR’ built-in function (see section `[ERROR]’); o The `EVAL’ built-in function (see section `[EVAL]’); o The `FPUTS’ built-in function (see section `[FPUTS]’); o The `INTERNALINDEX’ built-in function (see section `[INTERNALINDEX]’); o The `PUSHSUBST’ built-in function (see section `[PUSHSUBST]’); o The `TYPEOUT’ built-in function (see section `[TYPEOUT]’); o The `UPPERCASE’ built-in function (see section `[UPPERCASE]’); o The `WARNING’ built-in function (see section `[WARNING]’); o The `XXSUBST’ internal use only built-in function; All other built-in functions will not evaluate their arguments. See the mentioned functions for details, and in particular `EVAL()’ for a description of this evaluation process. `ADDTOCOUNTER’ The `ADDTOCOUNTER’ function adds a given value to a counter. It expects two arguments: the counter name, and an additive expression defining the value to add. The counter must be previously created with `DEFINECOUNTER’. The additive expression may not contain blank spaces and may use + and - operators, its operands may either be integral numeric values or names of (defined) counters. The resulting value can be negative; in that case, a value is subtracted from the destination counter. For example, if `one’ and `two’ are counters, then ADDTOCOUNTER(one)(-two)\// subtracts two’s value from one ADDTOCOUNTER(one)(two+two)\// adds 2 x two’s value to one See further section `[COUNTERS]’. `ADDTOSYMBOL’ Since Yodl version 2.00 symbols can be manipulated. To add text to an existing symbol the builtin `ADDTOSYMBOL’ is available. It expects two parameter lists: the symbol’s name, and the text to add to the symbol. The symbol must have been created earlier using DEFINECOUNTER (see section `[DEFINECOUNTER]’). The macro’s second argument is not evaluated while `ADDTOSYMBOL’ is processed. Therefore, it is easy to add the text of another symbol or the expansion of a macro to a symbol value. E.g., ADDTOSYMBOL(one)(SYMBOLVALUE(two)XXnl()) This adds the text of symbol `two’, followed by a new line, to the contents of symbol `one’ only when symbol `one’ is evaluated, not when `ADDTOSYMBOL’ is evaluated. Example: ADDTOSYMBOL(LOCATION)(this is appended to LOCATION) `ATEXIT’ `ATEXIT’ expects one argument. The argument is appended to the output file. Note that this text is subject to character table translations etc.. An example using this function is the following. A document in the LaTeX typesetting language requires `\end{document}’ to occur at the end of the document. To automatically append this string to the output file, the following specification can be used: ATEXIT(NOEXPAND(\end{document})) Several `ATEXIT’ lists can be defined. They are appended to the output file in the reverse order of specification; i.e., the first `ATEXIT’ list is appended to the output file last. That means that in general the `ATEXIT’ text should be specified when a `matching’ starting command is sent to the output file; as in: COMMENT(Start the LaTeX document.) NOEXPAND(\begin{document}) COMMENT(Ensure its proper ending.) ATEXIT(NOEXPAND(\end{document})) `CHAR’ The command `CHAR’ takes one argument, a number or a character, and outputs its corresponding ASCII character to the final output file. This command is built for `emergency situations’, where you need to typeset a character despite the fact that it may be redefined in the current character table (for a discussion of character tables, see `[CHARTABLES]’). Also, the `CHAR’ function can be used to circumvent Yodl’s requirement that open- and close-parentheses must match. The following arguments may be specified with `CHAR’ (attempted in this order): o A decimal number indicating the number of the character in the ascii-table (for example `CHAR’`(41)’); o A plain, single character (for example `CHAR’`(#)’). So, when you’re sure that you want to send a printable character that is not a closing parenthesis to the output file, you can use the form `CHAR’`(c)’, `c’ being the character (as in, `CHAR’`(;)’). To send a non-printable character or a closing parenthesis to the output file, look up the ASCII number of the character, and supply that number as argument to the `CHAR’ command. Example: The following two statements send an `A’ to the output file. CHAR(65) CHAR(A) The following statement sends a closing parenthesis: CHAR(41) Another way to send a string to the output file without expansion by character tables or by macro interpretation, is by using the function `NOTRANS’ (see section `[NOTRANS]’). If you want to send a string to the output without macro interpretation, but with character table translation, use `NOEXPAND’ (see section `[NOEXPAND]’). `CHDIR’ The command `CHDIR’ takes one argument, a directory to change to. This command is implemented to simplify the working with `includefile’ (see `includefile’ in `yodlmacros(7)’). As a demonstration, consider the following fragment: includefile(subdir/onefile) includefile(subdir/anotherfile) includefile(subdir/yetanotherfile) This fragment can be changed to: CHDIR(subdir) includefile(onefile) includefile(anotherfile) includefile(yetanotherfile) CHDIR(..) The current directory, as given to `CHDIR’, only affects how `includefile’ searches for its files. Note that this example assumes that the current working directory is a member of Yodl’s include-path specification (cf., Yodl’s `--include’ option). `COMMENT’ The `COMMENT’ function defines one parameter list. The text that is passed as argument is treated as comment. I.e., it is ignored; it is not copied to the final output file. As an alternative to (short) `COMMENT’ the triplet `\’`//’ can be used. It starts `end of line’ comment, ignoring all characters on a line starting at `\’`//’ up to the first non-blank character encountered on the next line. If the next line’s first non-blank characters are `\’`//’, then that begins another end of line comment, which will therefore also be skipped. To actually write `\’`//’ or, using the current font: \// in a yodl-converted document, write, e.g., `tt(\)tt(//)’ or, using the current font: nop(/)// in a yodl-source file, and write `\CHAR’`(/)/’ in `verb’ sections. Example: Hello world\// producess Hello world, skipping the rest \// this line is completely ignored s\// at this point Hello worlds has been produced. `COUNTERVALUE’ `COUNTERVALUE’s’ argument expands to the value of a counter. Its single argument must contain the name of a counter. The counter must have been created earlier using the builtin `DEFINECOUNTER’. Example: The counter has value COUNTERVALUE(MYCOUNTER). See also section `[COUNTERS]’. `DECWSLEVEL’ `DECWSLEVEL’ requires one (empty) argument. It reduces the current white-space level. The white-space level typically is used in files that only define Yodl macros. When no output should be generated while processing these files, the white-space level can be used to check for this. If the white-space level exceeds zero, a warning is generated if the file produces non-whitespace output. The builtin function `DECWSLEVEL’ is used to reduce the whitespace level following a previous call of `INCWSLEVEL’. Once the white space level exceeds zero, no output will be generated. White space, therefore effectively is ignored. The white space level cannot be reduced to negative values. A warning is issued if that would have happened if it were allowed. Example: INCWSLEVEL() DEFINESYMBOL(....) DEFINEMACRO(...)(...)(...) DECWSLEVEL() Without the `INCWSLEVEL’ and `DECWSLEVEL’, calls, the above definition would generate four empty lines to the output stream. The `INCWSLEVEL’ and `DECWSLEVEL’ calls may be nested. The best approach is to put an `INCWSLEVEL’ at the first line of a macro-defining Yodl-file, and a matching `DECWSLEVEL’ call at the very last line. `DEFINECHARTABLE’ `DEFINECHARTABLE’ is used to define a character translation table. The function expects two parameterlists, containing the name of the character table and character table translations on separate lines. These character table translations are of the form character = quoted-string Here, character is always a value within single quotes. It may be a single character, an octal character value or a hexadecimal character value. The single character may be prefixed by a \-character (e.g., `’\\’’). The octal character value must start with a backslash, followed by three octal digits (e.g., `’\045’’. The hexadecimal character value starts with `0x’, followed by two hexadecimal characters. E.g., `’0xbe’’. The double quoted string may contain anything (but the string must be on one line), possibly containing escape-sequences as well: in the double quoted string the standard C escape sequences `\a’ (alert), `\b’ (beep), `\f’ (formfeed), `\n’ (newline), `\r’ (carriage return), `\t’ (tab), and `\v’ (vertical tab) are recognized and automatically converted to their special meanings. Starting with Yodl 2.14.0 octal and hexadecimal constants may also be used. E.g., character `Y’ may also be specified using the octal value `\131’ or the hexadecimal value `\x59’. Any other character following a backslash character (`\’) defines itself: `\\’ represents a single backslash character. Example: DEFINECHARTABLE(demotable)( ’&’ = "&" ’\\’ = "\\backslash" ’\045’ = "oct(45)" ’0xa4’ = "hex(a4)" ) The builtin function `DEFINECHARTABLE’ does not activate the table. The table is merely defined. To activate the character translation table, use `USECHARTABLE’. The discussion of character tables is postponed to section `[CHARTABLES]’. `DEFINECOUNTER’ `DEFINECOUNTER’ creates a new counter. This builtin function expects two arguments: the name of the counter and an additive expression whose value is used to initialize the counter. The additive expression may not contain blank spaces and may use + and - operators, its operands may either be integral numeric values or names of (defined) counters. The resulting value can be negative; in that case, a value is subtracted from the destination counter. Examples: DEFINECOUNTER(year)(1950) DEFINECOUNTER(nTimes)(year+12)\// initializes nTimes to 1962 See also section `[COUNTERS]’ and the `USECOUNTER’ and `ADDTOCOUNTER’ builtin functions. `DEFINEMACRO’ `DEFINEMACRO’ is used to define new macros. This function expects three arguments: o An identifier, being the name of the macro to define. This identifier may only consist of uppercase or lowercase characters. Note that it can not contain numbers, nor underscore characters. o A number, stating the number of arguments that the macro will require once it’s used. The number must be in the range 0 to 61. o The text that the macro expands to, once used. This text may contain the strings `ARG’x, x being 1, 2, etc.. At these places the arguments to the macro are pasted in. The numbers that identify the arguments are 1 to 9, then A to Z and finally a to z. This gives a range of 61 expandable arguments, which is enough for all real-life applications. For example, the following fragment defines a macro `bookref’, which can be used to typeset a reference to a book. It requires three arguments; say, an author, a title and the name of a publisher: DEFINEMACRO(bookref)(3)( Author(s): ARG1 Book title: ARG2 Published by: ARG3 ) Such a macro could be used as follows: bookref(Sobotta/Becher) (Atlas der Anatomie des Menschen) (Urban und Schwarzenberg, Berlin, 1972) When called, it would produce the following output: Author(s): Sobotta/Becher Book title: Atlas der Anatomie des Menschen Published by: Urban und Schwarzenberg, Berlin, 1972 While applying a macro, the values of the three arguments are pasted to the places where `ARG1’, `ARG2’ etc. occur in the definition. Note the following when defining new macros: o The argument containing the name of the new macro, `(bookref)’ in the above example, must occur right after `DEFINEMACRO’. No spaces are allowed in between. Space characters and newlines may however occur following this first argument. This behavior of the `yodl’ program is similar to the usage of the defined macro: the author information must, enclosed in parentheses, follow right after the `bookref’ identifier. I implemented this feature to improve the distinguishing between macros and real text. E.g., a macro `me’ might be defined, but the text I like me (but so do you) still is simple text; the macro `me’ only is activated when a parenthesis immediately follows it. o Be careful when placing newlines or spaces in the definition of a new macro. E.g., the definition, as given: DEFINEMACRO(bookref)(3)( Author(s): ARG1 Book title: ARG2 Published by: ARG3 ) introduces extra newlines at the beginning and ending of the macro, which are copied to the output each time the macro is used. The extra newline occurs, of course, right before the sequence `Author(s):’ and following the evaluation of `ARG3’. A simple backslash character at the end of the `DEFINEMACRO’ line would prevent the insertion of extra newline characters: DEFINEMACRO(bookref)(3)(\ Author(s): ARG1 Book title: ARG2 Published by: ARG3 ) o Note that when a macro is used which requires no arguments at all, one empty argument still must be specified. E.g., my macro package (see chapter `[MACROPACKAGE]’) defines a macro `it’ that starts a bullet item in a list. The macro takes no arguments, but still must be typed as `it()’. This behavior is consistent: it helps distinguish which identifiers are macros and which are simple text. o Macro arguments may evaluate to text. When a \ is appended to the macro-argument, or in the default input handling within a non-zero white-space level (see section `[INCWSLEVEL]’) this may invalidate a subsequent macro call. E.g., the macro DEFINEMACRO(oops)(1)( ARG1 XXnl() ) when called as `oops(hello world)’, produces the output: hello worldXXnl() To prevent this gluing to arguments to subsequent macros, a single `+’ should be prepended to the macro call: DEFINEMACRO(oops)(1)( ARG1 +XXnl() ) See also section `[PLUSIDENT]’ obout the `+identifier’-sequence. o Note the preferred layout of macro definitions and macro calls. Adhere to this form, to prevent drowning in too many parentheses. In particular: o Put all elements of the macro definition on one line, except for the macro-expansion itself. Each expansion element should be on a line by itself. o When calling macros put the macro’s arguments underneath each other. If the macrolists themselves contain macro-calls, put each call again on a line of its own, indenting one tab-position beyond the location of the opening parenthesis of the argument. o No continnuation backslashes are required between arguments. o With complex calls, indent just the arguments, and put the parentheses in their required of logical locations. Example of a complex call: complex( first( ARG1 )( ARG2 +XXnl() ) ARG3 +nop() ARG4 +XXnl() ) o Macro expansion proceeds as follows: o The arguments are read from the input o The contents of the arguments then replace their `ARGx’ references in the macro’s definition (in some exceptional cases, clearly indicated as such when applicable, the arguments themselves are evaluated first, and then these evaluated arguments are used as replacements for their corresponding `ARGx’ references). o The now modified macro is read by Yodl’s lexical scanner. This may result in yet another macro expansion, which will then be evaluated recursively. o Eventually, all expansion is completed (well, should complete, since Yodl doesn’t test for eternal recursion) and scanning of the input continues beyond the original macro call. For example, assume we have the following two macros: DEFINEMACRO(First)(1)( Hello ARG1 +XXnl() ) DEFINEMACRO(Second)(1)( First(ARG1) First(ARG1) ) and the following call is issued: Second(Yodl) then the following happens: o `Second(Yodl)’ is read as encountered. o `ARG1’ in `Second’ is replaced by YODL, and the resulting macro body is sent to the lexical scanner for evaluation: It will see: First(Yodl)First(Yodl) o The first call to `First()’ is now evaluated. This puts (after replacing `ARG1’ by YODL) the following on the scanner’s input: Hello Yodl+XXnl()First(Yodl) o `Hello Yodl’ contains no macro call, so it is written to the output stream. Remains: +XXnl()First(Yodl) o Assume `XXnl()’ merely contains a newline (represented by `\n’, here), so `+XXnl()’ is now replaced by `\n’. This results in the following input for the lexical scanner: \nFirst(Yodl) o The `\n’ is now written to the output stream, and the scanner sees: First(Yodl) o The second call to `First()’ is now evaluated. This puts the following on the scanner’s input: Hello Yodl+XXnl() o `Hello Yodl’ is written to the output stream. Remains: +XXnl() o `+XXnl()’ is now replaced by `\n’. The lexical scanner sees: \n o The newline is printed and we’re done. `DEFINESYMBOL’ `DEFINESYMBOL’ expects two arguments. An identifier, which is the name of the symbol to define, and the textual value of the symbol. If the second argument is empty, the symbol is defined, but has an empty value. The earlier interpretation of a Yodl symbol as a logical flag can still be used, but allowing it to obtain textual values greatly simplifies various Yodl macros. Example: DEFINESYMBOL(Yodl)(Your own document language) DEFINESYMBOL(Options)() `DELETECHARTABLE’ `DELETECHARTABLE’ removes a definition of a character table that was defined by `DEFINECHARTABLE’. This function expects one argument: the name of the character table remove. It’s an error to attempt to delete a character table that is currently in use or to attempt to delete a non-existing character table. Example: DELETECHARTABLE(mytable) `DELETECOUNTER’ `DELETECOUNTER’ removes a definition of a counter that was defined by `DEFINECOUNTER’. This function expects one argument: the name of the counter to remove. If the counter does not exist, a warning is issued. It is not considered an error to try to delete a counter that has not been defined earlier. Example: DELETECOUNTER(mycounter) `DELETEMACRO’ `DELETEMACRO’ removes a definition of a macro that was defined by `DEFINEMACRO’. This function takes one argument: the macro name to remove. There is no error condition (except for syntax errors): when no macro with a matching name was previously defined, no action is taken. For example, the safe way to define a macro is by first undefining it. This ensures that possible previous definitions are removed first: Example: DELETEMACRO(mymacro) `DELETENOUSERMACRO’ `DELETENOUSERMACRO’ removes a `nousermacro’ definition. The function expects one argument: the name of the `nousermacro’ identifier to be removed from the nousermacro-set. There is no error condition (except for syntax errors): when the identifier wasn’t stored as a `nousermacro’ no action is taken. Example: DELETENOUSERMACRO(mymacro) `DELETESYMBOL’ `DELETESYMBOL’ removes the definition of a symbol variable. It expects one argument, holding the name of the variable to deleted. This macro has no error condition (except for syntax errors): the symbol in question may be previously defined, but that is not necessary. Example: DELETESYMBOL(Options) `ERROR’ The `ERROR’ function takes one argument: text to display to the standard error stream. The current input file and line number are also displayed. After displaying the text, the `yodl’ program aborts with an exit status of 1. The text passed to the function is expanded first. See the example. The `ERROR’ function is an example of a function that evaluates its argument itself. This command can be used, e.g., in a macro package when an incorrect macro is expanded. In my macro package (see chapter `[MACROPACKAGE]’) the `ERROR’ function is used when the sectioning command `chapter()’ is used in an `article’ document (in the package, `chapter’’s are only available in `book’s or `report’s). An analogous builtin function is `WARNING’, which also prints a message but does not exit (see section `[WARNING]’). Example: In the following call, `COUNTERVALUE(NTRIES)’ is replaced by its actual value: ERROR(Stopping after COUNTERVALUE(NTRIES) attempts) `EVAL’ The `EVAL’ function takes one argument: the text to be evaluated. This function allows you to perform an indirect evaluation of Yodl commands. Assume that there is a symbol `varnam’ containing the name of a counter variable, then the following displays the counter’s value, after having incremented it: EVAL(NOTRANS(USECOUNTER)(SYMBOLVALUE(varnam))) Here, `EVAL’ performs the following steps: o First, `NOTRANS(USECOUNTER)’ is evaluated, producing `USECOUNTER’. o Next, the open parenthesis is processed, producing the open parenthesis itself o Then, `SYMBOLVALUE(varnam)’ is evaluated, producing the name of a counter, e.g. ``counter’’. o The closing parentheis is processed, producing the closing parenthesis itself. o All this results in the text USECOUNTER(counter) o This text is presented to Yodl’s lexical scanner, resulting in incrementing the counter, and displaying its incremented value. b(Caveat): macro arguments themselves are usually not evaluated. So, a construction like USECOUNTER(EVAL(SYMBOLVALUE(varnam))) fails, as ``EVAL(SYMBOLVALUE(varnam))’’ is not a legal name for a counter. Here the `EVAL()’ call is used as an argument, and is therefore not expanded. The distinction is subtle, and is a consequence of the fact that builtin functions receive unprocessed arguments. Builtin functions may impose certain requirements on their arguments (like `USECOUNTER’ requiring the name of a counter) and these requirements are checked on the arguments as received. Summarizing: `EVAL’ acts as follows: o Its argument is presented to Yodl’s lexical scanner o The output produced by the processing of the argument is then inserted into the input stream in lieu of the original `EVAL’ call. Most built-in functions do not evaluate their arguments. In fact, only `ERROR, EVAL, FPUTS, INTERNALINDEX, PUSHSUBST, TYPEOUT, UPPERCASE, WARNING’ and the iinternally used `XXSUBST’ functions evaluate their arguments. Postponing evaluations allows you to write: DEFINESYMBOL(later)(SYMBOLVALUE(earlier)) Eventually, and not when `later’ is defined, a statement like SYMBOLVALUE(later) produces the value of `earlier’ at the moment `SYMBOLVALUE(later)’ is processed. This is, in all its complex consequences, what would be expected in most cases. It allows us to write general macros producing output that is only evaluated when the text of symbols and values of arguments become eventually, rather than when the macro is defined, available. Decisions like these invariably result in questions like `what if I have to define variables using values of other variables?’ In those cases `EVAL()’ must be used. The following example shows the definition of three symbols: `one’ receives an initial value, `two’ returns `one’’s actual value when `two’’s value is displayed, `three’, using `EVAL()’, stores `one’’s initial value. The example also shows yet another way to suppress macro calls, using the macro `nop()’ which is defined in the all standard conversion types: DEFINESYMBOL(one)(One’s first value) DEFINESYMBOL(two)(SYMBOLVALUE(one)) EVAL(DEFINESYMBOL+nop()(three)(SYMBOLVALUE(one))) SETSYMBOL(one)(One’s second value) "SYMBOLVALUE(two)" COMMENT(displays "One’s second value") "SYMBOLVALUE(three)" COMMENT(displays "One’s first value") `FILENAME’ The function `FILENAME()’ produces an absolute path to the currently processed Yodl file. This is not necessarily the canonical path name, as it may contain current- and parent-path directories. `FPUTS’ The function `FPUTS’ expects two arguments: the first argument is information to be appended to a file, whose name is given as the second argument. The first argument is processed by Yodl before it is appended to the requested filename, so it may contain macro calls. For example, the following statement appends a countervalue to the mentioned file: FPUTS(There have been COUNTERVALUE(attempts) attempts)(/tmp/logfile) The second argument (name of the file) is not evaluated, but is used as received. `IFBUILTIN’ The `IFBUILTIN’ function tests whether its first argument is the name of a builtin function. If so, the second argument is evaluated, else, the third argument is evaluated. All three arguments (the variable, the true-list and the false-list) must be present; though the true-list and/or the false-list may be empty. Example: IFBUILTIN(IFBUILTIN)(\ `BUILTIN’ is a builtin - function )(\ `BUILTIN’ is NOT a builtin - function ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFCHARTABLE’ The `IFCHARTABLE’ function tests whether its first argument is the name of a character table. The character table needs not be active. If the name is the name of a character table, the second argument is evaluated, else, the third argument is evaluated. All three arguments (the name, the true list and the false list) must be present; though the true list and/or the false list may be empty. Example: IFCHARTABLE(standard)(\ `standard’ is a character tablebuiltin - function )(\ `standard’ is NOT a character tablebuiltin - function ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFDEF’ The `IFDEF’ function tests for the definition status of the argument in its first argument. If it is a defined entity, the second argument is evaluated, else, the third argument is evaluated. All three arguments (the entity, the true list and the false list) must be present; though the true list and/or the false list may be empty. The true list is evaluated if the first argument is the name of: o a built-in function, or o a character table, or o a counter, or o a no-user-macro symbol, or o a symbol, or o a user-defined macro, or Example: IFDEF(someName)(\ `someName’ is a defined entity )(\ `someName is not defined. ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFEMPTY’ `IFEMPTY’ expects three arguments: a symbol, a true-list and a false-list. `IFEMPTY’ evaluates to the true-list if the symbol is an empty string; otherwise, it evaluates to the false-list. The function does not further evaluate its argument. Its use is primarily to test whether a macro has received an argument or not. If the intent is to check whether a symbol’s value is empty or not, IFSTREQUAL `[IFSTREQUAL]’ should be used, where the first argument is the name of a symbol, and the second argument is empty. Example: IFEMPTY(something)(\ `something’ is empty... )(\ `something’ is not an empty string ) In the same way, `IFEMPTY’ can be used to test whether an argument expands to a non-empty string. A more elaborate example follows below. Say you want to define a `bookref’ macro to typeset information about an author, a book title and about the publisher. The publisher information may be absent, the macro then typesets `unknown’: \ DEFINEMACRO(bookref)(3)(\ Author(s): ARG1 Title: ARG2 Published by: \ IFEMPTY(ARG3) (\ Unknown\ )(\ ARG3\ ) ) Using the macro, as in: \ bookref(Helmut Leonhardt) (Histologie, Zytologie und Microanatomie des Menschen) () would now result in the text `Unknown’ behind the `Published by:’ line. Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFEQUAL’ `IFEQUAL’ expects four argument lists. It tests whether its first argument is equal to its second argument. If so, the third argument is evaluated, else, the fourth argument is evaluated. All four argument lists must be present, though all can be empty. The first two arguments of `IFEQUAL’ should be integral numeric arguments. In order to determine whether the first two arguments are equal, their values are determined: o If the argument starts with an integral numerical value, that value is the value of the argument. o If the argument is the name of a counter, the counter’s value is the value of the argument o If the values of the first two arguments van be determined accordingly, their equality determines whether the true list (when the values are equal) or the false list (when the values are unequal) will be evaluated. o Otherwise, `IFEQUAL’ evaluates the false list. Example: IFEQUAL(0)()(\ 0 and an empty string are equal )(\ 0 and an empty string are not equal ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFGREATER’ `IFGREATER’ expects four argument lists. It tests whether its first argument is greater than its second argument. If so, the third parameter list is evaluated, otherwise its fourth argument is evaluated. All four argument lists must be present, though all can be empty. The first two arguments of `IFGREATER’ should be integral numeric arguments. In order to determine whether the first two arguments are equal, their values are determined: o If the argument starts with an integral numerical value, that value is the value of the argument. o If the argument is the name of a counter, the counter’s value is the value of the argument o If the values of the first two arguments van be determined accordingly, their order relation determines whether the true list (when the first value is greater than the second value) or the false list (when the first value is smaller or equal than the second value) is evaluated. o Otherwise, `IFGREATER’ evaluates the false list. Example: IFGREATER(counter)(5)(\ counter exceeds the value 5 )(\ counter does not exceeds the value 5, or counter is no Yodl-counter. ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFMACRO’ The `IFMACRO’ function tests whether its first argument is the name of a macro. If the name is the name of a macro, the second argument is evaluated, else, the third argument is evaluated. All three arguments (the name, the true list and the false list) must be present; though the true list and/or the false list may be empty. Example: IFMACRO(nested)(\ `nested’ is the name of a macro )(\ There is no macro named `nested’ ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFSMALLER’ `IFSMALLER’ expects four argument lists. It tests whether its first argument is smaller than its second argument. If so, the third parameter list is evaluated, otherwise its fourth argument is evaluated. All four argument lists must be present, though all can be empty. The first two arguments of `IFSMALLER’ should be integral numeric arguments. In order to determine whether the first two arguments are equal, their values are determined: o If the argument starts with an integral numerical value, that value is the value of the argument. o If the argument is the name of a counter, the counter’s value is the value of the argument o If the values of the first two arguments van be determined accordingly, their order relation determines whether the true list (when the first value is smaller than the second value) or the false list (when the first value is greater than or equal to the second value) is evaluated. o Otherwise, `IFSMALLER’ evaluates the false list. Example: IFSMALLER(counter)(5)(\ counter is smaller than the value 5, or counter is no Yodl-counter )(\ counter exceeds the value 5 ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. `IFSTREQUAL’ `IFSTREQUAL’ tests for the equality of two strings. It expects four arguments: two strings to match, a true list and a false list. The true list is only evaluated when the contents of the two string arguments exactly match. The first two arguments of `IFSTREQUAL’ are partially evaluated: o If the argument is the name of a symbol, the symbol’s value is the value of the argument o Otherwise, the argument itself is used. In the degenerate case where the string to be compared is actually the name of a `SYMBOL’, use a temporary `SYMBOL’ variable containing the name of that symbol, and compare it to whatever you want to compare it with. Alternatively, write a blank space behind the arguments, since the arguments are then interpreted `as is’. In practice, the need for these constructions seem to arise seldom, however. Example: IFSTREQUAL(MYSYMBOL)(Hello world)( The symbol `MYSYMBOL’ holds the value `Hello world’ )( The symbol `MYSYMBOL’ doesn’t hold the value `Hello world’ ) `IFSTRSUB’ `IFSTRSUB’ tests whether a string is a sub-string of another string. It acts similar to IFSTREQUAL, but it tests whether the second string is part of the first one. The first two arguments of `IFSTREQULA’ are partially evaluated: o If the argument is the name of a symbol, the symbol’s value is the value of the argument o Otherwise, the argument itself is used. In the degenerate case where the string to be compared is actually the name of a `SYMBOL’, use a temporary `SYMBOL’ variable containing the name of that symbol, and compare it to whatever you want to compare it with. Alternatively, write a blank space behind the arguments, since the arguments are then interpreted `as is’. In practice, the need for these constructions seem to arise seldom, however. Example: IFSTRSUB(haystack)(needle)( `needle’ was found in `haystack’ )( `needle’ was not found in `haystack’ ) Note that both `haystack’ and `needle’ may be the names of symbols. If they are, their contents are is compared, rather than the literal names `haystack’ and `needle’ `IFSYMBOL’ The `IFSYMBOL’ function tests whether its first argument is the name of a symbol. If it is the name of a symbol, the second argument is evaluated, otherwise the third argument is evaluated. All three arguments (the name, the true list and the false list) must be present; though the true list and/or the false list may be empty. Example: IFSYMBOL(nested)(\ `nested’ is the name of a symbol )(\ There is no symbol named `nested’ ) Please note the preferred layout: The first argument immediately follows the function name, then the second argument (the true list) is indented, as is the false list. The layout closely follows the preferred layout of `if-else’ statements of many programming languages. C( FBB consider additive expressions ) `IFZERO’ `IFZERO’ expects three arguments. If the first argument is zero (0) the function expands to the true list (the second argument). Otherwise it expands to the false list (the third argument). The first argument of `IFZERO’ should be an integral numeric value. Its value is determined as follows: o If the argument starts with an integral numerical value, that value is the value of the argument. o If the argument is the name of a counter, the counter’s value is the value of the argument o Otherwise, the first arguments evaluates as 0, and the false list is used. Note that, starting with Yodl version 2.00 the first argument is not evaluated. So `COUNTERVALUE(somecounter)’ always evaluates as 0. If the value of a counter is required, simply provide its name as the first argument of the `IFZERO’ function. Example: DEFINEMACRO(environment)(2)(\ IFZERO(ARG2)(\ NOEXPAND(\end{ARG1})\ )(\ NOEXPAND(\begin{ARG1})\ )\ ) Such a macro may be used as follows: environment(center)(1) Now comes centered text. environment(center)(0) which would of course lead to `\begin’ and `\end{center}’. The numeric second argument is used here as a on/off switch. `INCLUDEFILE’ `INCLUDEFILE’ takes one argument, a filename. The file is processed by Yodl. If a file should be inserted without processing the builtin function NOEXPANDINCLUDE `[NOEXPANDINCLUDE]’ or NOEXPANDPATHINCLUDE `[NOEXPANDPATHINCLUDE]’ should be used. The `yodl’ program supplies, when necessary, an extension to the filename. The supplied extension is `.yo’, unless defined otherwise during the compilation of the program. Furthermore, Yodl tries to locate the file in the Yodl’s include path (which may be set using the `--include’ option). The actual value of the include path is shown in the usage information, displayed when Yodl is started without arguments. Example: INCLUDEFILE(latex) Here, Yodl attempts to include the file `latex’ or `latex.yo’ from the current include path. When the file is not found, Yodl aborts. `INCWSLEVEL’ `INCWSLEVEL’ requires one (empty) argument. It increments the current white-space level. The white-space level typically is used in files that only define Yodl macros. When no output should be generated while processing these files, the white-space level can be used to check for this. If the white-space level exceeds zero, a warning is generated if the file produces non-whitespace output. The builtin function `DECWSLEVEL’ is used to decrement the whitespace level following a previous `INCWSLEVEL’ call. Once the white space level exceeds zero, no output is generated. White space, therefore is effectively ignored. The white space level cannot be reduced to negative values. A warning is issued if that would have happened if it were allowed. Example: INCWSLEVEL() DEFINESYMBOL(....) DEFINEMACRO(...)(...)(...) DECWSLEVEL() Without the `INCWSLEVEL’ and `DECWSLEVEL’, calls, the above definition would generate four empty lines to the output stream. The `INCWSLEVEL’ and `DECWSLEVEL’ calls may be nested. The best approach is to put an `INCWSLEVEL’ at the first line of a macro-defining Yodl-file, and a matching `DECWSLEVEL’ call at the very last line. `INTERNALINDEX’ `INTERNALINDEX’ expects one argument list. The argument list is evaluated and written to the index file. The index file is defined since Yodl version 2.00, and contains the fixup information which was previously written to Yodl’s output as the `.YODLTAGSTART. ... .YODLTAGEND.’ sequence. The index file allows for greater processing speed, at the expense of an additional file. The associated `yodlpost’ postprocessing program reads and processes the index file, and modifies the corresponding yodl-output accordingly. The index file is not created when output is written to the standard output name, since Yodl is unable to request the system for the current file offset. The entries of the index file always fit on one line. `INTERNALINDEX’ changes newline characters in its argument into single blank spaces. Each line starts with the current offset of Yodl’s output file, thus indicating the exact location where a fixup is requested. An example of a produced fixup line could be 3004 ref MACROPACKAGE indicating that at offset 3004 in the produced output file a reference to the label `MACROPACKAGE’ is requested. Assuming a html conversion, The postprocessor thereupon writes something like <a href="outfile04.html#MACROPACKAGE">4.3.2.</a> into the actual output file while processing Yodl’s output up to offset location 3004. Consequently, producing Yodl-output normally consists of two steps: o First, Yodl itself is started, producing, e.g., `out.idx’ (the index file) and `out.yodl’ (Yodl’s raw output). o Then, Yodl’s post-processor processes `out.idx’ and `out.yodl’, producing one or more final output files, in which the elements of the index file have been properly handled. This may result in multiple output file, like `report.html, report01.html, report02.html’ etc. `NOEXPAND’ `NOEXPAND’ is used to send text to the final output file without being expanded by Yodl (the other methods are the `CHAR’ macro, see section `[CHAR]’, and the `NOTRANS’ macro, see section `[NOTRANS]’). `NOEXPAND’ takes one argument, the text in question. Whatever occurs in the argument is not subject to parsing or expansion by Yodl, but is simply copied to the output file (except for `CHAR’ and (iinternally used) `XXSUBST’ functions in the argument, which are expanded. If `CHAR’-expansion is not required either NOTRANS `[NOTRANS]’ can be used). Furthermore, the contents of the argument are also subject to character table translations, using the currently active table. This should come as no surprise. Ignoring character tables would make both the processing of `CHAR’ calls and the `NOTRANS’ function superfluous. So, the following situations are recognized: ────────────────────────────────────────────── support chartables and CHAR ────────────────────────────── Macro expansion yes no ────────────────────────────────────────────── Yes (standard) Push chartable (standard) Pop chartable No NOEXPAND NOTRANS ────────────────────────────────────────────── E.g., let’s assume that you need to write in your document the following text: INCLUDEFILE(something or the other) IFDEF(onething)( ... )( .... ) NOEXPAND(whatever) The way to accomplish this is by prefixing the text by `NOEXPAND’ followed by an open parenthesis, and by postfixing it by a closing parenthesis. Otherwise, the text would be expanded by Yodl while processing it (and would lead to syntax errors, since the text isn’t correct in the sense of the Yodl language). For this function, keep the following caveats in mind: o There is only one thing that a `NOEXPAND’ cannot protect from expansion: an `ARG’x in a macro definition. The argument specifier is always processed. E.g., after DEFINEMACRO(thatsit)(1)( That is --> NOEXPAND(ARG1) <-- it! ) thatsit(after all) the `ARG1’ inside the `NOEXPAND’ statement is replaced with `after all’. o The `NOEXPAND’ function must, as all functions, be followed by a argument. The parentheses of the list must therefore be `balanced’. For unbalanced lists, use `CHAR(40)’ to set an open parenthesis, or `CHAR(41)’ to typeset a closing parenthesis. `NOEXPANDINCLUDE’ `NOEXPANDINCLUDE’ takes one argument, a filename. The file is included. The filename is uses as specified. The include path is not used when locating this file. The argument to `NOEXPANDINCLUDE’ is partially evaluated: o If the argument is the name of a symbol, the symbol’s value is the value of the argument o Otherwise, the argument itself is used. The thus obtained file name is not further evaluated: in particular, it is not affected by available character translations. The contents of the file are included literally, not subject to macro expansion. Character translations are performed, though. If character translations are not appropriate, PUSHCHARTABLE can be used to suppress character table translations temporarily. The purpose of NOEXPANDINCLUDE is to include source code literally in the document, as in: NOEXPANDINCLUDE(literal.c) The function NOEXPANDPATHINCLUDE can be used to insert a file which is located in one of the directories specified in Yodl’s include path. `NOEXPANDPATHINCLUDE’ `NOEXPANDPATHINCLUDE’ takes one argument, a filename. The file is included. The file is searched for in the directories specified in Yodl’s includepath. The argument to `NOEXPANDPATHINCLUDE’ is partially evaluated: o If the argument is the name of a symbol, the symbol’s value is the value of the argument o Otherwise, the argument itself is used. The thus obtained file name is not further evaluated: in particular, it is not affected by available character translations. Like the `NOEXPANDINCLUDE’ function, the contents of the file are included literally, not subject to macro expansion. Character translations are performed, though. If character translations are not appropriate, PUSHCHARTABLE `[PUSHCHARTABLE]’ can be used to suppress character table translations temporarily. The purpose of NOEXPANDPATHINCLUDE is to include source code as defined in a macro package literally into the document, as in: NOEXPANDPATHINCLUDE(rug-menubegin.xml) `NOTRANS’ `NOTRANS’ copies its one argument literally to the output file, without expanding macros in it and without translating the characters with the current translation table. The `NOTRANS’ function is typically used to send commands for the output format to the output file. For example, consider the following code fragment: COMMENT(--- Define character translations for \, { and } in LaTeX. ---) DEFINECHARTABLE(standard)( ’\\’ = "$\\backslash$" ’{’ = "\\verb+{+" ’}’ = "\\verb+}+" ) COMMENT(--- Activate the translation table. ---) USECHARTABLE(standard) COMMENT(--- Now two tests: ---) NOEXPAND(\input{epsf.tex}) NOTRANS(\input{epsf.tex}) `NOEXPAND’ sends $\backslash$input\verb+{+epsf.tex\verb+}+ since the characters in its argument are translated with the `standard’ translation table. In contrast, `NOTRANS’ sends `\input{epsf.tex}’. The argument of `NOTRANS’ must be balanced with respect to its parentheses. When using an unbalanced set of parentheses, use `CHAR(40)’ to send a literal (, or `CHAR(41)’ to send a `)’. While converting Yodl-documents to target document types Yodl frequently uses the (not further documented) builtin function `XXSUBST’. In the unlikely event that the text `XXSUBST(...)’ must be written in a document, the sequence XXSUBST+CHAR(40)...CHAR(41) can be used. The NOEXPAND description summarizes all combinations of character translations and/or macro expansion, and how they are handled and realized by Yodl. `NOUSERMACRO’ `NOUSERMACRO’ controls `yodl’’s warnings in the following way: When Yodl is started with the `-w’ flag on the command line, warnings are generated when Yodl encounters a possible macro name, i.e., a name that is followed by a parenthesized argument, while no macro by that name has been defined. Yodl then prints something like `cannot expand possible user macro’. Examples of such sequences are, `The necessary file(s) are in /usr/local/lib/yodl’, or `see the manual page for sed(1)’. The candidate macros are `file’ and `sed’; these names could just as well be `valid’ user macros followed by their argument. When a corresponding `NOUSERMACRO’ statement appears before `yodl’ encounters the candidate macros, no warning is generated. A fragment might therefore be: NOUSERMACRO(file sed) The necessary file(s) are in ... See the manual page for sed(1). The `NOUSERMACRO’ accepts one or more names in its argument, separated by white space, commas, colons, or semi-colons. `OUTBASE’ `OUTBASE’ inserts the current basename of the output file into the output file. The basename is the name of the file of which the directory components and extension were stripped. If the output file is the standard output file, `-’ is inserted. `OUTDIR’ `OUTDIR’ inserts the current path name of the output file into the output file. The path name is a, not necessarily absolute, designator of the directory in which the output file is located. If the output file is indicated as, e.g., `-o out’, then `OUTDIR’ simply inserts a dot. If the output file is the standard output file, a dot is inserted too. `OUTFILENAME’ `OUTFILENAME’ inserts the current filename of the output file into the output file. The filename is the name of the file of which the directory components were stripped. If the output file is the standard output file, `-’ is inserted. `PARAGRAPH’ `PARAGRAPH’ isn’t really a builtin function, but as Yodl handles paragraphs in a special way it is probably useful to describe paragraph handling here nonetheless. Starting with Yodl 2.00 `PARAGRAPH’ operates as follows: If the macro is not defined, new paragraphs, defined as series of consecutive empty lines written to the output stream, are not handled different from any other series of characters sent to the output stream. I.e., they are inserted into that stream. However, if the macro has been defined, Yodl calls it whenever a new paragraph (defined as a series of at least two blank lines) has been recognized. The empty lines that were actually recognized may be obtained inside the `PARAGRAPH’ macro from the `XXparagraph’ symbol, if this symbol has been be defined by that time. If defined, it contains the white space that caused Yodl to call the `PARAGRAPH’ macro. Note that, in order to inspect `XXparagraph’ it must have been defined first. Yodl itself does not define this symbol itself. The `PARAGRAPH’ macro should be defined as a macro not expecting arguments. The macro is thus given a chance to process the paragraph in a way that’s fitting for the particular conversion type. If the `PARAGRAPH’ macro produces series of empty lines itself, then those empty lines do not cause Yodl to activate `PARAGRAPH’. So, Yodl itself will not recursively call `PARAGRAPH’, although the macro could call itself recursively. Of course, such recursive activcation of `PARAGRAPH’ is then the sole responsibility of the macro’s author, and not Yodl’s. Some document languages do not need paragraph starts; e.g., LaTeX handles its own paragraphs. Other document languages do need it: typically, `PARAGRAPH’ is then defined in a macro file to trigger some special action. E.g., a HTML converter might define a paragraph as: DEFINEMACRO(PARAGRAPH)(0)( XXnl() NOTRANS(<p>) ) A system like `xml’ has more strict requirements. Paragraphs here must be opened and closed using pairs of `<p>’ and `</p>’ tags. In those cases an auxiliary counter can be used to indicate whether there is an open paragraph or not. The `PARAGRAPH’ macro could check for this as follows, assuming the availability of a counter `XXp’: DEFINEMACRO(PARAGRAPH)(0)( XXnl() IFZERO(XXp)( )( NOTRANS(</p>) ) NOTRANS(<p>) SETCOUNTER(XXp)(1) ) Note that the above fragment exemplifies an approach, not necessarily the implementation of the `PARAGRAPH’ macro for an xml-converter. `PIPETHROUGH’ The builtin function `PIPETHROUGH’ is, besides `SYSTEM’, the second function with which a Yodl document can affect its environment. `PIPETHROUGH’ can be very useful. It uses an external program to accomplish special features. The idea is that an external command is started, to which a block of text from within a Yodl document is `piped’. The output of that child program is piped back into the Yodl document; hence, a block of text is `piped through’ an external program. Whatever is received again in the Yodl run, is further processed. The `PIPETHROUGH’ function takes two arguments: o the command to run, and o the text to send to that command. Functionally, the occurrence of the `PIPETHROUGH’ function and of its two arguments is replaced by whatever the child program produces on its standard output. An example might be the inclusion of the current date, as in: The current date is: PIPETHROUGH(date)() In this example the command is `date’ and the text to send to that program is empty. The main purpose of this function is to provide a way by which external programs can be used to create, e.g., tables or figures for a given output format. Further releases of Yodl may contain such dedicated programs for the output formats. `POPCHARTABLE’ Character tables which are pushed onto the table stack using `PUSHCHARTABLE()’ are restored (popped) using `POPCHARTABLE()’. For a description of this mechanism please refer to section `[PUSHINGTABLES]’. `POPCOUNTER’ `POPCOUNTER’ is used to remove the topmost counter from the counter stack. The values of counters may be pushed on a stack using PUSHCOUNTER `[PUSHCOUNTER]’. To remove the topmost element of a counter’s stack `POPCOUNTER’ is available. `POPCOUNTER’ expects one argument: the name of the counter to pop. The previously pushed value then becomes the new value of the counter. A counter’s value may be popped after defining it, whereafter the stack is empty, but the counter will still be defined. In that case, using the counter’s value is considered an error. Examples: DEFINECOUNTER(YEAR)(1950) POPCOUNTER(YEAR) COMMENT(YEAR now has an undefined value) See also section `[COUNTERS]’. `POPMACRO’ `POPMACRO’ is used to remove the actual macro definition, restoring a previously pushed definition. The values of macros may be pushed on a stack using PUSHMACRO. To remove the topmost element of a macro’s stack `POPMACRO’ is available. `POPMACRO’ expects one argument: the name of the macro to pop. The previously pushed value then becomes the new value of the macro. A macro’s value may be popped after defining it, after which its stack is empty. In that case, using the macro (although the macro’s name is still defined) is considered an error. Example: DEFINEMACRO(Hello)(1)(Hello, ARG1, this is a macro definition) Hello(Karel) PUSHMACRO(Hello)(1)(Hello, ARG1, this is the new definition) Hello(Karel) POPMACRO(Hello) Hello(Karel) COMMENT(The third activation of Hello() produces the same output as the first activation) `POPSUBST’ `POPSUBST’ is used to revert to a previous level of interpretation of `SUBST’ definitions. Refer to the descriptions of the `PUSHSUBST’ and `SUBST’ builtin commands below for details. There is no limit to the number of times `POPSUBST’ can be called. Once the ``PUSHSUBST’ stack’ is empty `SUBST’ definitions are automatically interpreted (so no stack-underflow error is ever encountered). `POPSYMBOL’ `POPSYMBOL’ is used to remove the topmost symbol from the symbol stack. The values of symbols may be pushed on a stack using PUSHSYMBOL `[PUSHSYMBOL]’. To remove the topmost element of a symbol’s stack `POPSYMBOL’ is available. `POPSYMBOL’ expects one argument: the name of the symbol to pop. The previously pushed value then becomes the new value of the symbol. A symbol’s value may be popped after defining it, after which its stack is empty. In that case, using the symbol (although the symbol’s name is still defined) is considered an error. Example: DEFINESYMBOL(YEAR)(This happened in 1950) POPSYMBOL(YEAR) COMMENT(YEAR now has an undefined value) `POPWSLEVEL’ `POPWSLEVEL’ is used to remove the topmost wslevel from the wslevel stack. The values of wslevels may be pushed on a stack using PUSHWSLEVEL `[PUSHWSLEVEL]’. See also section DECWSLEVEL `[DECWSLEVEL]’ To remove the topmost element of a wslevel’s stack `POPWSLEVEL’ is available. `POPWSLEVEL’ expects one argument: the name of the wslevel to pop. The previously pushed value then becomes the new value of the wslevel. A wslevel’s value may be popped after defining it, emptying the stack, but the wslevel will still be defined. In that case, using the wslevel’s value is considered an error. Example: COMMENT(Assume WS level is zero) PUSHWSLEVEL(1) COMMENT(WS level now equals 1) POPWSLEVEL() COMMENT(WS level now equals 0 again) `PUSHCHARTABLE’ Once a character table has been defined, it can be pushed onto a stack using `PUSHCHARTABLE’. The pushed chartable may be popped later. `PUSHCHARTABLE’ is described in more detail in section `[PUSHINGTABLES]’. `PUSHCOUNTER’ `PUSHCOUNTER’ is used to start another lifetime for a counter, pushing its current value on a stack. A stack is available for each individual counter. `PUSHCOUNTER’ expects two arguments: the name of the counter to push and an additive expression whose value becomes the counter’s new value (after pushing the current value) The additive expression may not contain blank spaces and may use + and - operators, its operands may either be integral numeric values or names of (defined) counters. The resulting value can be negative; in that case, a value is subtracted from the destination counter. When the second argument is empty, then the new value will be zero. Specify the name of the counter twice to merely push its value, without modifying its current value. Examples: DEFINECOUNTER(YEAR)(1950) PUSHCOUNTER(YEAR)(1962) COMMENT(YEAR now has the value 1962, and a pushed value of 1950) See also section `[COUNTERS]’. `PUSHMACRO’ `PUSHMACRO’ is used to start another lifetime for a macro, pushing its current definition on a stack. A stack is available for each individual macro. `PUSHMACRO’ expects three arguments: the name of the macro to push, the number of its arguments after pushing (which may be different from the number of arguments interpreted by the pushed macro) and its new definition. So, `PUSHMACRO’ is used exactly like DEFINEMACRO, but redefines a current macro (or define a new macro if no macro was defined by the name specified as its first argument. Example: DEFINEMACRO(Hello)(1)(Hello, ARG1, this is a macro definition) Hello(Karel) PUSHMACRO(Hello)(1)(Hello, ARG1, this is the new definition) Hello(Karel) POPMACRO(Hello) Hello(Karel) COMMENT(The third activation of Hello() produces the same output as the first activation) `PUSHSUBST’ `PUSHSUBST’ can be used to (temporarily) suppress the interpretation of `SUBST’ definitions (the `SUBST’ built-in command is covered below, refer to its description for an example). `PUSHSUBST’ expects one argument: an integral number which is either 0 or non-zero (commonly: 1). After calling `PUSHSUBST’`(0)’ `SUBST’ definitions are not interpreted anymore; use `POPSUBST’`()’ to revert to the previous type of interpretation. Alternatively, `PUSHSUBST’`(0)’ can be used to stack another level of `SUBST’ interpretations on top of the last-used one. On a 64-bit computer the `PUSHSUBST’ stack can hold slightly more than 60 `SUBST’ interpretation levels. When more levels are pushed, the oldest levels are silently forgotten. Calling `POPSUBST’ once the `PUSHSUBST’ stack is empty results in activating the `SUBST’ interpretations (and so a stack-underflow error will not be encountered). `PUSHSYMBOL’ `PUSHSYMBOL’ is used to start another lifetime for a symbol, pushing its current value on a stack. A stack is available for each individual symbol. `PUSHSYMBOL’ expects two arguments: the name of the symbol to push and its new text after pushing. When the second argument is an empty argument, the new text will be empty. The new text may be specified as a literal text, or as the name of an existing symbol. Specify the name of the symbol twice to merely push its value, without modifying its current value. Examples: DEFINESYMBOL(YEAR)(This happened in 1950) PUSHSYMBOL(YEAR)(This happened in 1962) COMMENT(YEAR now has the value `This happened in 1962’ and a pushed value of `This happened in 1950’) `PUSHWSLEVEL’ `PUSHWSLEVEL’ is used to start another lifetime of the white-space level pushing the level’s current value on a stack. See also section INCWSLEVEL `[INCWSLEVEL]’ `PUSHWSLEVEL’ expects one argument, the new value of the white-space level. This value may be specified as a numerical value or as the name of a counter. The argument may be empty, in which case the new value will be zero. Example: COMMENT(Assume WS level is zero) PUSHWSLEVEL(1) COMMENT(WS level now equals 1) POPWSLEVEL() COMMENT(WS level now equals 0 again) `RENAMEMACRO’ `RENAMEMACRO’ takes two arguments: the name of a built-in macro (such as `INCLUDEFILE’) and its new name. E.g., after RENAMEMACRO(INCLUDEFILE)(include) a file must be included by `include(file)’. `INCLUDEFILE’ can no longer be used for this: following the `RENAMEMACRO’ action, the old name can no longer be used; it becomes an undefined symbol. If you want to make an alias for a built-in command, do it with `DEFINEMACRO’. E.g., after: DEFINEMACRO(include)(1)(INCLUDEFILE(ARG1)) both `INCLUDEFILE’ and `include’ can be used to include a file. `SETCOUNTER’ `SETCOUNTER’ expects two arguments: the name of a counter, and an additive expression defining the value to assign. The counter must be previously created with `DEFINECOUNTER’. The additive expression may not contain blank spaces and may use + and - operators, its operands may either be integral numeric values or names of (defined) counters. The resulting value can be negative; in that case, a negative value is assigned to the destination counter. For example, if `one’ and `two’ are counters, then SETTOCOUNTER(one)(-two)\// assigns -two’s value to one SETTOCOUNTER(one)(two+two)\// assigns 2 x two’s value to one See also section `[COUNTERS]’. `SETSYMBOL’ `SETSYMBOL’ expects two arguments: the name of a symbol, and the text to assign to the named symbol. The symbol must previously have been defined by `DEFINESYMBOL’. `SUBST’ `SUBST’ is a general-purpose substitution mechanism for strings appearing in the input. `SUBST’ takes two arguments: a search string and a substitution string. E.g., after SUBST(VERSION)(1.00) YODL transforms all occurrences of `VERSION’ in its input into `1.00’. `SUBST’ is also useful in situations where multi-character sequences should be converted to accented characters. E.g., a LaTeX converter might define: SUBST(’e)(+NOTRANS(\’{e})) Each `’e’ in the input will subsequently be converted to `e’. `SUBST’ may be used in combination with the command line flag `-P’, as in a invocation yodl2html -P’SUBST(VERSION)(1.00)’ myfile.yo Another useful substitution might be: SUBST(_OP_)(CHAR(40)) SUBST(_CP_)(CHAR(41)) which defines an opening parenthesis (`_OP_’) and a closing parenthesis (`_CP_’) as mapped to the `CHAR’ function. The strings `_OP_’ and `_CP_’ might then be used to produce unbalanced arguments. Note that: o The first argument of the `SUBST’ command, the search string, is taken literally. Yodl does not expand it; the string must be literally matched in the input. o The second argument, the replacement, is further processed by Yodl. Protect this text by `NOTRANS’ or `NOEXPAND’ where appropriate. Substitutions occur extremely early while YODL processes its input files. In order to process its input files, YODL takes the following steps: 1. It requests input from its lexical scanner (so-called tokens) 2. Its parser processes the tokens produced by the lexical scanner 3. Its parser may send text to an output `object’, which eventually appears in the output file generated by YODL. YODL performs all macro substitutions in step 2, and all character table conversions in step 3. However, the lexical scanner has access to the `SUBST’ definitions: as soon as its lexical analyzer detects a series of characters matching the defining sequence of a `SUBST’ definition, it replaces that defining sequence by its definition. That definition is then again read by the lexical scanner. Of course, this definition may, in turn, contain defining sequences of other `SUBST’ definitions: these are then replaced by their definitions as well. This implies: o Circular definitions may cause the lexical scanner to get stuck in a replacement loop. It is the responsibility of the author defining `SUBST’ definitions to make sure that this doesn’t happen. o Neither the parser, nor the output object ever sees the `SUBST’ defining character sequences: they only see their definitions. In some cases substitutions must be suppressed. Consider double quoted text strings that are frequently used in programming languages. E.g., `"hello world"’. The text inside the string should not be converted by Yodl, but unless substitutions can be suppressed the string "\"evil code" appears as "evil code" To suppress the interpretation of `SUBST’ definitions `PUSHSUBST’, introduced earlier, can be used. The predefined macro `verb’ suppresses the interpretation of `SUBST’ definitions by starting with `PUSHSUBST’`(0)’ and ending with `POPSUBST’`()’. `SYMBOLVALUE’ `SYMBOLVALUE’ expands to the value of a symbol. Its single argument must be the name of a symbol. The symbol must have been created earlier using `DEFINESYMBOL’. Example: The symbol has value SYMBOLVALUE(MYSYMBOL). `SYSTEM’ `SYSTEM’ takes one argument: a command to execute. The command is run via the standard C function `system’. `SYSTEM’ can be useful in many ways. E.g., you might want to log when someone processes your document, as in: SYSTEM(echo Document processed! | mail myself@my.host) Note that `SYSTEM’ merely performs an system-related task. It’s a process that is separated from the YODL process itself. One of the consequences of this is that any output generated by `SYSTEM’ not normally appears into YODL’s output file. If the output of a subprocess should be inserted into YODL’s output file, either use PIPETHROUGH `[PIPETHROUGH]’, or insert a temporary file as shown in the following example: SYSTEM(date > datefile) The current date is: INCLUDEFILE(datefile) SYSTEM(rm datefile) `TYPEOUT’ `TYPEOUT’ requires one argument. The text of the list is sent to the standard error stream, followed by a newline. This feature can be handy to show, e.g., messages such as version numbers in macro package files. Example: The following macro includes a file and writes to the screen that this file is currently processed. DEFINEMACRO(includefile)(1)( TYPEOUT(About to process document: ARG1) INCLUDEFILE(ARG1) ) `UPPERCASE’ `UPPERCASE’ converts a string or a part of it to upper case. It has two arguments: o The string to convert; o A length, indicating how many characters (starting from the beginning of the string) should be converted. The length indicator can be smaller than one or larger than the length of the string; in that case, the whole string is convertered. Example: UPPERCASE(hello world)(1) UPPERCASE(hello world)(5) UPPERCASE(hello world)(0) This code sample expands to: Hello world HELLO world HELLO WORLD `USECHARTABLE’ `USECHARTABLE’ takes one argument: the name of a translation table to activate. The table must previously have been defined using `DEFINECHARTABLE’. See section `[CHARTABLES]’ for a description of character translation tables. Alternatively, the name may be empty in which case the default character mapping is restored. `USECOUNTER’ `USECOUNTER’ is a combination of `ADDTOCOUNTER’ and `COUNTERVALUE’. It expects one argument: the name of an defined counter (see DEFINECOUNTER `[DEFINECOUNTER]’). The counter is first incremented by 1. Then the function expands to the counter’s value. See also section `[COUNTERS]’. `VERBOSITY’ `VERBOSITY’ expects two arguments, and may be used to change the verbosity level inside YODL files. The function may be used profitably for debugging purposes, to debug the expansion of a macro or the processing of a YODL input file. The first argument indicates the processing mode of the second argument, and it may be: o Empty, in which case the message-level is set to the value specified in the second argument; o `+’, in which case the value specified in the second argument augments the current message level; o `-’, in which case the value specified in the second argument augments is removed from the current message level The second argument specifies one or more, separated by blanks, message level names or it may be set to a hexadecimal value (starting with `0x’), using hexadecimal values to represent message levels. Also, `NONE’ may be used, to specify no message level, or `ALL’ can be used to specify all message levels. The following message levels are defined: o ALERT (0x40). When an alert-error occurs, Yodl terminates. Here Yodl requests something of the system (like a `get_cwd()’), but the system fails. o CRITICAL (0x20). When a critical error occurs, Yodl terminates. The message itself can be suppressed, but exiting can’t. A critical condition is, e.g., the omission of an open parenthesis at a location where a parenthesized argument should appear, or a non-existing file in an `INCLUDEFILE’ specification (as this file should be parsed). A non-existing file with a `NOEXPANDINCLUDE’ specification is a plain (non-critical) error. o DEBUG (0x01). Probably too much info, like getting information about each character that was read by Yodl. o ERROR (0x10). An error (like doubly defined symbols). Error messages will not stop the parsing of the input (up to a maximum number of errors), but no output is generated. o INFO (0x02). Not as detailed as `debug’, but still very much info, like information about media switches. o NOTICE (0x04). Information about, e.g., calls to the builtin function calls. o WARNING (0x08). Something you should know about, but probably not affecting Yodl’s proper functioning There also exists a level EMERG (0x80) which cannot be suppressed. The value `0x00’ represents `NONE’, the value `0xff’ represents `ALL’. When specifying multiple message levels using the hexadecimal form, their hexadecimal values should be binary-or-ed: adding them is ok, as long as you don’t specify `ALL’: VERBOSITY()(0x06) COMMENT(this specifies `INFO’ and `NOTICE’) When specifying message levels by their names, the names may be truncated at a unique point. However, the message level names are interpreted case sensitively, so `INF’ for `INFO’ is recognized as such, but `info’ for `INFO’ isn’t. The following examples all specify verbosity levels INFO and NOTICE: VERBOSITY()(I N) VERBOSITY()(N I) VERBOSITY()(NOT IN) VERBOSITY()(INFO NOTICE) `WARNING’ `WARNING’ takes one argument: text to display as a warning. The `yodl’ program makes sure that before showing the text, the current file and line number are printed. Other than this, `WARNING’ works just as `TYPEOUT’ (see section `[TYPEOUT]’). Note that an analogous function `ERROR’ exists, which prints a message and then terminates the program (see section `[ERROR]’).
FILES
The files in tmp/wip/macros define the converter’s macro packages. The scripts yodl2tex, yodl2html, yodl2man etc. perform the conversions.
SEE ALSO
yodl(1), yodlconverters(1), yodlletter(7), yodlmacros(7), yodlmanpage(7), yodlpost(1), yodlstriproff(1), yodltables(7), yodlverbinsert(1).
BUGS
--
AUTHOR
Frank B. Brokken (f.b.brokken@rug.nl),