BS(1) — HP-UX

NAME

bs − a compiler/interpreter for modest-sized programs

SYNOPSIS

bs [ file [ args ] ]

DESCRIPTION

Bs is a remote descendant of Basic and Snobol4 with some C language added. Bs is designed for programming tasks where program development time is as important as the resulting speed of execution. Formalities of data declaration and file/process manipulation are minimized. Line-at-a-time debugging, the trace and dump statements, and useful run-time error messages all simplify program testing. Furthermore, incomplete programs can be debugged; inner functions can be tested before outer functions have been written and vice versa.

If the command line file argument is provided, the file is used for input before the console is read. By default, statements read from file are compiled for later execution. Likewise, statements entered from the console are normally executed immediately (see compile and execute below). Unless the final operation is assignment, the result of an immediate expression statement is printed.

Bs programs are made up of input lines. If the last character on a line is a \, the line is continued. Bs accepts lines of the following form:

statement
label statement

A label is a name (see below) followed by a colon. A label and a variable can have the same name.

A bs statement is either an expression or a keyword followed by zero or more expressions. Some keywords (clear, compile, !, execute, include, ibase, obase, and run) are always executed as they are compiled.

Statement Syntax:

expression The expression is executed for its side effects (value, assignment, or function call). The details of expressions follow the description of statement types below.

break Break exits from the innermost for/while loop.

clear Clears the symbol table and compiled statements. Clear is executed immediately.

compile [ expression ]
Succeeding statements are compiled (overrides the immediate execution default). The optional expression is evaluated and used as a file name for further input. A clear is associated with this latter case. Compile is executed immediately.

continue Continue transfers to the loop-continuation of the current for/while loop.

dump [ name ] The name and current value of every non-local variable is printed. Optionally, only the named variable is reported. After an error or interrupt, the number of the last statement is displayed. The user-function trace is displayed after an error or stop that occurred in a function.

edit A call is made to the editor selected by the EDITOR environment variable if it is present, or ed(1) if EDITOR is undefined or null. If the file option is present on the command line, that file is passed to the editor as the file to edit. (Otherwise no file name is used.) Upon exiting the editor, a compile statement (and associated clear) is executed giving that file name as it’s argument.

exit [ expression ]
Return to system level. The expression is returned as process status.

execute Change to immediate execution mode (an interrupt has a similar effect). This statement does not cause stored statements to execute (see run below).

for name = expression expression statement

for name = expression expression
...

next

for expression , expression , expression statement

for expression , expression , expression
...

next The for statement repetitively executes a statement (first form) or a group of statements (second form) under control of a named variable. The variable takes on the value of the first expression, then is incremented by one on each loop, not to exceed the value of the second expression. The third and fourth forms require three expressions separated by commas. The first of these is the initialization, the second is the test (true to continue), and the third is the loop-continuation action (normally an increment).

fun f([a, ... ]) [v, ... ]
...

nuf Fun defines the function name, arguments, and local variables for a user-written function. Up to ten arguments and local variables are allowed. Such names cannot be arrays, nor can they be I/O associated. Function definitions may not be nested. Calling an undefined function is permissible, see function calls below.

freturn A way to signal the failure of a user-written function. See the interrogation operator (?) below. If interrogation is not present, freturn merely returns zero. When interrogation is active, freturn transfers to that expression (possibly by-passing intermediate function returns).

goto name Control is passed to the internally stored statement with the matching label.

ibase N Ibase sets the input base (radix) to N. The only supported values for N are the constants 8, 10 (the default), and 16. Hexadecimal values 10−15 are entered as a−f. A leading digit is required (i.e., f0a must be entered as 0f0a). Ibase (and obase, below) are executed immediately.

if expression statement

if expression
...

[else ...]

fi The statement (first form) or group of statements (second form) is executed if the expression evaluates to non-zero. The strings 0 and "" (null) evaluate as zero. In the second form, an optional else allows for a group of statements to be executed when the first group is not. The only statement permitted on the same line with an else is an if; only other fi’s can be on the same line with a fi. The concatenation of else and if into an elif is supported. Only a single fi is required to close an if ... elif ... [ else ... ] sequence.

include expression
The expression must evaluate to a file name. The file must contain bs source statements. Such statements become part of the program being compiled. Include statements may not be nested.

obase N Obase sets the output base to N (see ibase above).

onintr label

onintr The onintr command provides program control of interrupts. In the first form, control will pass to the label given, just as if a goto had been executed at the time onintr was executed. The effect of the statement is cleared after each interrupt. In the second form, an interrupt will cause bs to terminate.

return [expression]
The expression is evaluated and the result is passed back as the value of a function call. If no expression is given, zero is returned.

run The random number generator is reset. Control is passed to the first internal statement. If the run statement is contained in a file, it should be the last statement.

stop Execution of internal statements is stopped. Bs reverts to immediate mode.

trace [ expression ]
The trace statement controls function tracing. If the expression is null (or evaluates to zero), tracing is turned off. Otherwise, a record of user-function calls/returns will be printed. Each return decrements the trace expression value.

while expression statement

while expression
...

next While is similar to for except that only the conditional expression for loop-continuation is given.

! shell command
An immediate escape to the Shell.

# ... This statement is ignored. It is used to interject commentary in a program.

Expression Syntax:

name A name is used to specify a variable. Names are composed of a letter (uppercase or lowercase) optionally followed by letters and digits. Only the first six characters of a name are significant. Except for names declared in fun statements, all names are global to the program. Names can take on numeric (double float) values, string values, or can be associated with input/output (see the built-in function open() below).

name ( [expression [ , expression] ... ] )
Functions can be called by a name followed by the arguments in parentheses separated by commas. Except for built-in functions (listed below), the name must be defined with a fun statement. Arguments to functions are passed by value. If the function is undefined, the call history to the call of that function is printed, and a request for a return value (as an expression) is made. The result of that expression is taken to be the result of the undefined function. This permits debugging programs where not all the functions are yet defined. The value is read from the current input file.

name [ expression [ , expression ] ... ]
This syntax is used to reference either arrays or tables (see built-in table functions below). For arrays, each expression is truncated to an integer and used as a specifier for the name. The resulting array reference is syntactically identical to a name; a[1,2] is the same as a[1][2]. The truncated expressions are restricted to values between 0 and 32 767.

number A number is used to represent a constant value. A number is written in Fortran style, and contains digits, an optional decimal point, and possibly a scale factor consisting of an e followed by a possibly signed exponent.

string Character strings are delimited by " characters. The \ escape character allows the double quote (\"), new-line (\n), carriage return (\r), backspace (\b), and tab (\t) characters to appear in a string. Otherwise, \ stands for itself.

( expression ) Parentheses are used to alter the normal order of evaluation.

( expression , expression [ , expression ... ] ) [ expression ]
The bracketed expression is used as a subscript to select a comma-separated expression from the parenthesized list. List elements are numbered from the left, starting at zero.
The expression:

( False, True )[ a == b ]

has the value True if the comparison is true.

? expression The interrogation operator tests for the success of the expression rather than its value. At the moment, it is useful for testing end-of-file (see examples in the Programming Tips section below), the result of the eval built-in function, and for checking the return from user-written functions (see freturn). An interrogation “trap” (end-of-file, etc.) causes an immediate transfer to the most recent interrogation, possibly skipping assignment statements or intervening function levels.

− expression The result is the negation of the expression.

++ name Increments the value of the variable (or array reference). The result is the new value.

−− name Decrements the value of the variable. The result is the new value.

! expression The logical negation of the expression. Watch out for the shell escape command.

expression operator expression Common functions of two arguments are abbreviated by the two arguments separated by an operator denoting the function. Except for the assignment, concatenation, and relational operators, both operands are converted to numeric form before the function is applied.

Binary Operators (in increasing precedence):

= = is the assignment operator. The left operand must be a name or an array element. The result is the right operand. Assignment binds right to left, all other operators bind left to right.

_ _ (underscore) is the concatenation operator.

& │ & (logical and) has result zero if either of its arguments are zero. It has result one if both of its arguments are non-zero; │ (logical or) has result zero if both of its arguments are zero. It has result one if either of its arguments is non-zero. Both operators treat a null string as a zero.

< <= > >= == !=
The relational operators (< less than, <= less than or equal, > greater than, >= greater than or equal, == equal to, != not equal to) return one if their arguments are in the specified relation. They return zero otherwise. Relational operators at the same level extend as follows: a>b>c is the same as a>b & b>c. A string comparison is made if both operands are strings.

+ − Add and subtract.

∗ / % Multiply, divide, and remainder.

^ Exponentiation.

Built-in Functions:

Dealing with arguments

arg(i) is the value of the i-th actual parameter on the current level of function call. At level zero, arg returns the i-th command-line argument (arg(0) returns bs).

narg() returns the number of arguments passed. At level zero, the command argument count is returned.

Mathematical

abs(x) is the absolute value of x.

atan(x) is the arctangent of x. Its value is between −π/2 and π/2.

ceil(x) returns the smallest integer not less than x.

cos(x) is the cosine of x (radians).

exp(x) is the exponential function of x.

floor(x) returns the largest integer not greater than x.

log(x) is the natural logarithm of x.

rand() is a uniformly distributed random number between zero and one.

sin(x) is the sine of x (radians).

sqrt(x) is the square root of x.

String operations

size(s) the size (length in bytes) of s is returned.

format(f, a) returns the formatted value of a. F is assumed to be a format specification in the style of printf(3S). Only the %...f, %...e, and %...s types are safe. Since it is not always possible to know whether a is a number or a string when the format call is coded, coercing a to the type required by f by either adding zero (for e or f format) or concatenating (_) the null string (for s format) should be considered.

index(x, y) returns the number of the first position in x that any of the characters from y matches. No match yields zero.

trans(s, f, t) Translates characters of the source s from matching characters in f to a character in the same position in t. Source characters that do not appear in f are copied to the result. If the string f is longer than t, source characters that match in the excess portion of f do not appear in the result.

substr(s, start, width)
returns the sub-string of s defined by the starting position and width.

match(string, pattern)

mstring(n) The pattern is similar to the regular expression syntax of the ed(1) command. The characters ., [, ], ^ (inside brackets), ∗ and $ are special. The mstring function returns the n-th (1 <= n <= 10) substring of the subject that occurred between pairs of the pattern symbols $ and $ for the most recent call to match. To succeed, patterns must match the beginning of the string (as if all patterns began with ^). The function returns the number of characters matched. For example:

match("a123ab123", ".∗$[a−z]$") == 6
mstring(1) == "b"

File handling

open(name, file, function)

close(name) The name argument must be a bs variable name (passed as a string). For the open, the file argument may be 1) a 0 (zero), 1, or 2 representing standard input, output, or error output, respectively; 2) a string representing a file name; or 3) a string beginning with an ! representing a command to be executed (via sh −c). The function argument must be either r (read), w (write), W (write without new-line), or a (append). After a close, the name reverts to being an ordinary variable. If name was a pipe, a wait(2) is executed before the close completes. The bs exit command does not do such a wait. The initial associations are:

open("get", 0, "r")
open("put", 1, "w")
open("puterr", 2, "w")

Examples are given in the following section.

access(s, m) executes access(2).

ftype(s) returns a single character file type indication: f for regular file, p for FIFO (i.e., named pipe), d for directory, b for block special, or c for character special.

Tables

table(name, size)
A table in bs is an associatively accessed, single-dimension array. “Subscripts” (called keys) are strings (numbers are converted). The name argument must be a bs variable name (passed as a string). The size argument sets the minimum number of elements to be allocated. Bs prints an error message and stops on table overflow. The result of table is name.

item(name, i)

key() The item function accesses table elements sequentially (in normal use, there is no orderly progression of key values). Where the item function accesses values, the key function accesses the “subscript” of the previous item call. It fails (or in the absence of an interrogate operator, returns null) if there was no valid subscript for the previous item call. The name argument should not be quoted. Since exact table sizes are not defined, the interrogation operator should be used to detect end-of-table; for example:

table("t", 100)
...
# If word contains "party", the following expression adds one
# to the count of that word:
++t[word]
...
# To print out the the key/value pairs:
for i = 0, ?(s = item(t, i)), ++i if key() put = key()_":"_s

If the interrogation operator is not used, the result of item is null if there are no further elements in the table. Null is, however, a legal “subscript”.

iskey(name, word)
The iskey function tests whether the key word exists in the table name and returns one for true, zero for false.

Odds and ends

eval(s) The string argument is evaluated as a bs expression. The function is handy for converting numeric strings to numeric internal form. Eval can also be used as a crude form of indirection, as in:

name = "xyz" eval("++"_ name)

which increments the variable xyz. In addition, eval preceded by the interrogation operator permits the user to control bs error conditions. For example:

?eval("open(\"X\", \"XXX\", \"r\")")

returns the value zero if there is no file named “XXX” (instead of halting the user’s program). The following executes a goto to the label L (if it exists):

label="L"
if !(?eval("goto "_ label)) puterr = "no label"

plot(request, args)
If the tplot command is available, the plot function produces output on devices recognized by tplot. The requests are as follows:

Call Function

plot(0, term)causes further plot output to be piped into tplot with an argument of −Tterm. Term may be up to 40 characters in length.

plot(1)“erases” the plotter.

plot(2, string)labels the current point with string.

plot(3, x1, y1, x2, y2)draws the line between (x1,y1) and (x2,y2).

plot(4, x, y, r)draws a circle with center (x,y) and radius r.

plot(5, x1, y1, x2, y2, x3, y3)draws an arc (counterclockwise) with center (x1,y1) and endpoints (x2,y2) and (x3,y3).

plot(6)is not implemented.

plot(7, x, y)makes the current point (x,y).

plot(8, x, y)draws a line from the current point to (x,y).

plot(9, x, y)draws a point at (x,y).

plot(10, string)sets the line mode to string.

plot(11, x1, y1, x2, y2)makes (x1,y1) the lower left corner of the plotting area and (x2,y2) the upper right corner of the plotting area.

plot(12, x1, y1, x2, y2)causes subsequent x (y) coordinates to be multiplied by x1 (y1) and then added to x2 (y2) before they are plotted. The initial scaling is plot(12, 1.0, 1.0, 0.0, 0.0).

Some requests do not apply to all plotters. All requests except zero and twelve are implemented by piping characters to tplot.

Each statement executed from the keyboard re-invokes tplot, making the results unpredictable if a complete picture is not done in a single operation. Plotting should thus be done either in a function or a complete program, so all the output can be directed to tplot in a single stream.

last() in immediate mode, last returns the most recently computed value.

EXAMPLES

Using bs as a calculator:

$ bs
# Distance (inches) light travels in a nanosecond.
186000 ∗ 5280 ∗ 12 / 1e9
11.78496
...

# Compound interest (6% for 5 years on $1,000).
int = .06 / 4
bal = 1000
for i = 1 5∗4 bal = bal + bal∗int
bal − 1000
346.855007
...
exit

The outline of a typical bs program:

# initialize things:
var1 = 1
open("read", "infile", "r")
...
# compute:
while ?(str = read)
...
next
# clean up:
close("read")
...
# last statement executed (exit or stop):
exit
# last input line:
run

Input/Output examples:

# Copy "oldfile" to "newfile".
open("read", "oldfile", "r")
open("write", "newfile", "w")
...
while ?(write = read)
...
# close "read" and "write":
close("read")
close("write")

# Pipe between commands.
open("ls", "!ls ∗", "r")
open("pr", "!pr −2 −h ′List′", "w")
while ?(pr = ls) ...
...
# be sure to close (wait for) these:
close("ls")
close("pr")

WARNINGS

The graphics mode is nearly useless unless the tplot command is provided on your system.

Bs is not extremely tolerant of some errors. Mistyping a fun declaration is painful, as a new definition cannot be made without doing a clear. Starting using the edit command is the best solution in this case.

Museum

Related Articles