ACK(7)


NAME
     ACK - Additional information on the Amsterdam Compiler Kit compilers

DESCRIPTION
     These are the details on the Amsterdam Compiler  Kit  compilers  for  the
     languages  C,  Modula-2, and Pascal.  The design decisions that were made
     where  the  respective  standards  allowed  or  mandated  this,  and  the
     extensions that were implemented.

ANSI C REPORT
     This section specifies the implementation-defined behavior of the  ANSI-C
     compiler as required by ANS X3.159-1989.

     ANS A.6.3.1

     o    Diagnostics are placed on the standard error output.  They have  the
          following specification:

                "<file>", line <nr>: [(<class>)] <diagnostic>

          There are  three  classes  of  diagnostics:  'error',  'strict'  and
          'warning'.   When  the  class is 'error', the class specification is
          absent.  The class 'strict' is used for violations of  the  standard
          which  are  not  severe  enough to stop compilation, for example the
          occurrence  of  non  white-space  after  an  '#endif'  preprocessing
          directive.   The  class  'warning'  is  used  for  legal but dubious
          constructions, for example the declaration of a structure-tag  in  a
          parameter type list.

     ANS A.6.3.2

     o    The function 'main' can have zero or two parameters.   When  it  has
          two  parameters,  the  first  parameter is an integer specifying the
          number of arguments on the command  line  (including  the  command).
          The  second  parameter  is  a pointer to an array of pointers to the
          arguments (as strings).

     o    Interactive devices are terminals.

     ANS A.6.3.3

     o    The number of significant characters is  64.   Corresponding  upper-
          case and lower-case letters are different.

     ANS A.6.3.4

     o    The  compiler  assumes  ASCII-characters  in  both  the  source  and
          execution character set.


     o    There are no multibyte characters.

     o    There are 8 bits in a character.

     o    Character constants  that  cannot  be  represented  in  8  bits  are
          truncated.

     o    Character constants that are more than 1 character  wide  will  have
          the first character specified in the least significant byte.

     o    The only supported locale is 'C'.

     o    A plain 'char' has the same range of values as 'signed char'.

     ANS A.6.3.5

     o    The i80x86 and 68000 both  have  a  two's  complement  binary-number
          system.  Shorts are 2 bytes; ints are 2 bytes under 16-bits Minix-PC
          and 68000 Minix, 4 bytes under  32-bits  Minix-PC;  longs  occupy  4
          bytes.

     o    Converting an integer to a shorter signed integer is implemented  by
          ignoring  the  high-order  byte(s)  of  the  former.   Converting  a
          unsigned integer to a signed integer of the same type is  only  done
          in   administration.    This  means  that  the  bit-pattern  remains
          unchanged.

     o    The result of bitwise operations on signed integers are what can  be
          expected on a two's complement machine.

     o    When either operand is negative, the result of the / operator is the
          largest  integer  less than or equal to the algebraic quotient.  The
          sign of the remainder  on  integer  division  is  the  sign  of  the
          enumerator.

     o    The right-shift of a negative value is negative.

     ANS A.6.3.6

     o    The compiler uses IEEE format  for  floating-point  numbers.   High-
          precision floating-point is used for constant folding.

     o    Truncation is done to the nearest floating-point number that can  be
          represented.

     ANS A.6.3.7

     o    The type of the sizeof-operator (also known as size_t) is  'unsigned
          int'.

     o    Casting an integer to a pointer or vice versa has no effect in  bit-
          pattern  when  the  sizes  are  equal.   Otherwise the value will be
          truncated or  zero-extended  (depending  on  the  direction  of  the
          conversion and the relative sizes).

     o    The type of a 'ptrdiff_t' is 'int' on Minix-PC, and  'long'  on  the
          68000 Minix versions.

     ANS A.6.3.8

     o    Since the front end has only limited control over the registers,  it
          can  only  make  it  more likely that variables that are declared as
          registers also end up  in  registers.   The  only  things  that  can
          possibly be put into registers are plain ints and pointers.

     ANS A.6.3.9

     o    When a member of a union object is accessed  using  a  member  of  a
          different  type,  the  resulting value will usually be garbage.  The
          compiler makes no effort to catch these errors.

     o    The  alignment  of  types  under  16-bit  Minix-PC  is  1  byte  for
          characters  and  2  bytes  for  all  other types.  Under other Minix
          versions 'int' and smaller types are aligned to a multiple of  their
          size,  bigger  scalar types are aligned like 'int'.  Arrays have the
          same alignment as their elements; structs  and  unions  are  aligned
          like their field with the worst alignment.

     o    A plain 'int' bit-field is taken as a 'signed int'.  This means that
          a field with a size 1 bit-field can only store the values 0 and -1.

     o    In bit-fields, high-order bits are allocated first.

     o    An enum has the same size as a plain 'int'.

     ANS A.6.3.10

     o    An access to a volatile object is either a load or  a  store.   Just
          mentioning  a  volatile  variable is not enough.  E.g. the statement
          'x;' where x is declared volatile, does not  constitute  an  access.
          When  a  volatile  object should be read, but its value ignored, 'if
          (x);' should do the trick.

     ANS A.6.3.11

     o    There is no fixed limit on the number of declarators that may modify
          an arithmetic, structure or union type, although specifying too many
          may cause the compiler to run out of memory.


     ANS A.6.3.12

     o    The maximum number of cases in a switch-statement is in the order of
          1e9, although the compiler may run out of memory somewhat earlier.

     ANS A.6.3.13

     o    Since  both  the  preprocessor  and  the  compiler   assume   ASCII-
          characters,  a  single character constant in a conditional-inclusion
          directive matches the same value in the execution character set.

     o    The  preprocessor  recognizes  -I...  command-line   options.    The
          directories   thus   specified  are  searched  first.   After  that,
          /usr/include is visited.

     o    Quoted names are first looked for in the directory in which the file
          which does the include resides.

     o    The characters in a h- or q- char-sequence  are  taken  to  be  UNIX
          paths.

     o    Neither the front-end nor the preprocessor know any pragmas.

     o    Since the compiler runs on Minix __DATE__ and __TIME__  will  always
          be defined.

     ANS A.6.3.14

     o    NULL is defined as ((void *)0).  This in  order  to  detect  dubious
          constructions like 'int x = NULL;'.

     o    The diagnostic printed by 'assert' is as follows:

                Assertion "<expr>" failed, file "<file>", line <line>

          where <expr> is the argument to the assert macro, printed as string.
          (the <file> and <line> should be clear)

     o    The sets for character test macros for the C locale are as follows:

                Name       Set
                isalnum    0-9A-Za-z
                isalpha    A-Za-z
                iscntrl    \000-\037\177
                islower    a-z
                isupper    A-Z
                isprint    \040-\176

          As an addition, there is an isascii macro,  which  tests  whether  a
          character  is an ASCII character.  Characters in the range from \000
          to \177 are ASCII characters.

     o    The behavior of ACK mathematical functions on  domain  error  is  as
          follows:

                Name    Returns
                asin    0.0
                acos    0.0
                atan2   0.0
                fmod    0.0
                log     -HUGE_VAL
                log10   -HUGE_VAL
                pow     0.0
                sqrt    0.0

          Minix-vmd uses the BSD4.4 C  library  and  the  Sun  FDLIBM  C  math
          library  instead  of the ACK library.  See math(3) for details about
          the math functions.  The Minix-vmd libraries offer at least the same
          functionality as the ACK library.

     o    Underflow range errors do not cause errno to be set.

     o    The function fmod returns 0.0 and sets errno to EDOM when the second
          argument is 0.0.

     o    The set of signals for  the  signal  function  is  as  described  by
          sigaction(2).

     o    A text-stream need not end in a new-line character.

     o    White space characters before a new-line appear when read in.

     o    There may be any number of null  characters  appended  to  a  binary
          stream.

     o    The file position indicator of an append mode  stream  is  initially
          positioned at the beginning of the file.

     o    A write on a text stream does not cause the associated  file  to  be
          truncated beyond that point.

     o    The buffering intended by the standard is fully supported.

     o    A zero-length file actually exists.

     o    A file name can consist of any character, except for  the  '\0'  and
          the '/'.



     o    A file can be open multiple times.

     o    When a remove is done on an open file, reading  and  writing  behave
          just  as  can  be  expected  from  a  non-removed  file.   When  the
          associated stream is closed, however, all written data will be lost.

     o    When a file exists prior to a call to rename, it is removed.

     o    The %p conversion in fprintf has the same effect as %#x on  Minix-PC
          and %#lx on the 68000 versions of Minix.

     o    The %p conversion in fscanf has the same effect as  %x  on  Minix-PC
          and %lx on the 68000 versions of Minix.

     o    A - character that is neither the first nor the  last  character  in
          the  scanlist  for  %[  conversion is taken to be a range indicator.
          When the first character has a higher ASCII-value than  the  second,
          the - will just be put into the scanlist.

     o    The value of errno when fgetpos or ftell failed is  that  of  lseek.
          This means:

                EBADF    -  when the stream is not valid
                ESPIPE   -  when fildes is associated with a pipe
                EINVAL   -  the resulting file pointer would be negative


     o    The messages generated by perror depend on the value of errno.   The
          mapping of errors to strings is done by strerror.

     o    When the requested size is zero, malloc, calloc and realloc return a
          null-pointer under Minix.  Under Minix-vmd a unique non-null pointer
          is returned.

     o    When abort is called, output buffers  will  be  flushed.   Temporary
          files  (made  with  the tmpfile function) will have disappeared when
          SIGABRT is not caught or ignored.

     o    The exit function returns the low-order eight bits of  its  argument
          to the environment.

     o    The  predefined  environment  names  are  controlled  by  the  user.
          Setting  environment  variables is done through the putenv function.
          This function accepts a pointer to char as its  argument.   To  set,
          for example, the environment variable TERM to a230 one writes

                static char terminal[] = "TERM=a230";
                putenv(terminal);

          The argument to putenv is stored in an internal table, so  malloc'ed
          strings cannot be freed until another call to putenv (which sets the
          same  environment variable) is made.  The argument to putenv must be
          writable, which means that officially,  the  argument  cannot  be  a
          string constant.  The function returns 1 if it fails, 0 otherwise.

     o    The argument to system is passed as argument to /bin/sh -c.

     o    The strings returned by strerror depend on errno.  They  are  listed
          in  intro(2).   Everything  else  causes strerror to return "unknown
          error" under Minix, or the  result  of  sprintf("Error  %d",  errno)
          under Minix-vmd.

     o    The local time zone is per default GMT.  This can be changed through
          the TZ environment variable, e.g. TZ=EST6.  See TZ(5).

     o    The clock  function  returns  the  number  of  ticks  since  process
          startup.

  References

     [1]  ANS X3.159-1989 American National Standard for Information Systems -
          Programming Language C

THE MINIX MODULA-2 COMPILER
     This section describes the implementation-specific features of the  Minix
     Modula-2  compiler.   It  is  not intended to teach Modula-2 programming.
     For a description of the Modula-2 language, the  reader  is  referred  to
     [1].

  The language implemented

     This paragraph discusses the deviations from  the  Modula-2  language  as
     described  in  the  'Report  on The Programming Language Modula-2', as it
     appeared in [1], from now on referred to  as  'the  Report'.   Also,  the
     Report  sometimes  leaves  room  for interpretation.  The section numbers
     mentioned are the section numbers of the Report.

  Syntax (section 2)

     The syntax recognized is that of the Report, with some extensions to also
     recognize  the  syntax  of an earlier definition, given in [2].  Only one
     compilation unit per file is accepted.

  Vocabulary and Representation (section 3)

     The input '10..' is parsed as two tokens: '10' and '..'.

     The empty string "" has type

           ARRAY [0 .. 0] OF CHAR

     and contains one character: 0C.

     When the text of a comment starts  with  a  '$',  it  may  be  a  pragma.
     Currently, the following pragmas exist:

     (*$F      (F stands for Foreign) *)
     (*$R[+|-] (Runtime checks, on or off, default on) *)
     (*$A[+|-] (Array bound checks, on or off, default off) *)
     (*$U      (Allow for underscores within identifiers) *)

     The Foreign pragma  is  only  meaningful  in  a  DEFINITION  MODULE,  and
     indicates  that this DEFINITION MODULE describes an interface to a module
     written in another language (for instance C or Pascal).   Runtime  checks
     that can be disabled are:  range checks, CARDINAL overflow checks, checks
     when assigning a CARDINAL to an INTEGER and vice versa, and  checks  that
     FOR-loop  control-variables  are  not  changed  in  the body of the loop.
     Array bound checks can be enabled, because many EM implementations do not
     implement  the  array  bound checking of the EM array instructions.  When
     enabled, the compiler generates a check before  generating  an  EM  array
     instruction.  Even when underscores are enabled, they still may not start
     an identifier.

     Constants of type LONGINT are  integers  with  a  suffix  letter  D  (for
     instance  1987D).   Constants  of  type LONGREAL have suffix D if a scale
     factor is missing, or have D in place of E  in  the  scale  factor  (f.i.
     1.0D,  0.314D1).   This  addition  was  made, because there was no way to
     indicate long constants, and  also  because  the  addition  was  made  in
     Wirth's newest Modula-2 compiler.

  Declarations and scope rules (section 4)

     Standard identifiers are  predeclared,  and  valid  in  all  parts  of  a
     program.   They are called pervasive.  Unfortunately, the Report does not
     state how this pervasiveness is accomplished.  However, page  87  of  [1]
     states:   'Standard  identifiers  are  automatically  imported  into  all
     modules'.  Our implementation therefore allows redeclarations of standard
     identifiers within procedures, but not within modules.

  Constant expressions (section 5)

     Each operand of a constant expression must be a constant:   a  string,  a
     number,  a  set,  an enumeration literal, a qualifier denoting a constant
     expression, a type transfer with a  constant  argument,  or  one  of  the
     standard procedures ABS, CAP, CHR, LONG, MAX, MIN, ODD, ORD, SIZE, SHORT,
     TSIZE, or VAL, with constant argument(s); TSIZE and SIZE may also have  a
     variable as argument.




  Type declarations (section 6)

     1. Basic types (section 6.1)

     The type CHAR includes the ASCII character set as a subset.  Values range
     from 0C to 377C, not from 0C to 177C.

     2. Enumerations (section 6.2)

     The maximum number of enumeration literals in any one enumeration type is
     MAX(INTEGER).

     3. Record types (section 6.5)

     The syntax of variant sections in [1] is different from the one  in  [2].
     Our implementation recognizes both, giving a warning for the older one.

     4. Set types (section 6.6)

     The only limitation imposed by the compiler is that the base type of  the
     set  must be a subrange type, an enumeration type, CHAR, or BOOLEAN.  So,
     the lower bound may be negative.  However, if a negative lower  bound  is
     used, the compiler gives a warning of the restricted class.

     The standard type BITSET is defined as

           TYPE BITSET = SET OF [0 .. 8*SIZE(INTEGER)-1];


  Expressions (section 8)

     1. Operators (section 8.2)

     1.1. Arithmetic operators (section 8.2.1)

     The Report does not specify the priority of the unary operators +  or  -:
     It does not specify whether

           - 1 + 1

     means

           - (1 + 1)

     or

           (-1) + 1

     The Minix Modula-2 compiler implements the second alternative.

  Statements (section 9)

     1. Assignments (section 9.1)

     The Report does not define the evaluation order in  an  assignment.   Our
     compiler certainly chooses an evaluation order, but it is explicitly left
     undefined.  Therefore, programs that depend  on  it  may  cease  to  work
     later.

     The types INTEGER and CARDINAL are  assignment-compatible  with  LONGINT,
     and REAL is assignment-compatible with LONGREAL.

     2. Case statements (section 9.5)

     The size of the type of the case-expression must be less than or equal to
     the word-size.

     The Report does not specify what  happens  if  the  value  of  the  case-
     expression  does  not occur as a label of any case, and there is no ELSE-
     part.  In our implementation, this results in a runtime error.

     3. For statements (section 9.8)

     The Report does not specify the legal types for a control variable.   Our
     implementation  allows  the basic types (except REAL), enumeration types,
     and subranges.  A runtime warning is generated  when  the  value  of  the
     control variable is changed by the statement sequence that forms the body
     of the loop, unless runtime checking is disabled.

     4. Return and exit statements (section 9.11)

     The  Report  does  not  specify  which  result-types  are   legal.    Our
     implementation allows any result type.

  Procedure declarations (section 10)

     Function procedures must exit through a RETURN statement,  or  a  runtime
     error occurs.

     1. Standard procedures (section 10.2)

     Our implementation supports NEW and DISPOSE for backwards  compatibility,
     but issues warnings for their use.

     Also, some new  standard  procedures  were  added,  similar  to  the  new
     standard procedures in Wirth's newest compiler:

     -    LONG converts an argument of type  INTEGER  or  REAL  to  the  types
          LONGINT or LONGREAL.

     -    SHORT performs the inverse transformation, without range checks.

     -    FLOATD is analogous to FLOAT, but yields a result of type LONGREAL.

     -    TRUNCD is analogous to TRUNC, but yields a result of type LONGINT.

  System-dependent facilities (section 12)

     The type BYTE is added to the SYSTEM module.  It occupies a storage  unit
     of  8  bits.  ARRAY OF BYTE has a similar effect to ARRAY OF WORD, but is
     safer.  In some obscure cases the ARRAY OF WORD mechanism does not  quite
     work properly.

     The procedure IOTRANSFER is not implemented.

  Backwards compatibility

     Besides recognizing the  language  as  described  in  [1],  the  compiler
     recognizes   most  of  the  language  described  in  [2],  for  backwards
     compatibility.   It  warns  the  user  for  old-fashioned   constructions
     (constructions  that  [1] does not allow).  If the -3 option is passed to
     m2, this backwards compatibility feature is disabled.

  Compile time errors

     The compile time error messages are intended to be self-explanatory,  and
     not   listed   here.    The  compiler  also  sometimes  issues  warnings,
     recognizable by a warning-classification between parentheses.  There  are
     3 classifications:

     (old-fashioned use)
          These warnings are given on constructions that are  not  allowed  by
          [1], but are allowed by [2].

     (strict)
          These warnings are given on constructions that are supported by  the
          Minix  Modula-2  compiler,  but  might  not  be supported by others.
          Examples:  functions  returning  structured  types,  SET  types   of
          subranges with negative lower bound.

     (warning)
          The other warnings, such as warnings about variables that are  never
          assigned, never used, etc.

  Runtime errors

     The Traps module enables the  user  to  install  his  own  runtime  error
     handler.   The  default  one  just  displays  what  happened  and  exits.
     Basically, a trap handler is just a procedure that takes  an  INTEGER  as
     parameter.   The  INTEGER is the trap number.  This INTEGER can be one of
     the EM trap numbers, listed in [3], or one of the numbers listed  in  the
     Traps definition module.

     The following runtime errors may occur:

     array bound error
          This error is detected if the -A option is given to m2.

     range bound error
          Range bound errors are always detected, unless  runtime  checks  are
          disabled.

     set bound error

     cardinal overflow
          This error is detected, unless runtime checks are disabled.

     cardinal underflow
          This error is detected, unless runtime checks are disabled.

     divide by 0

     divide by 0.0

     conversion error
          This error occurs when assigning a negative value of type INTEGER to
          a  variable  of type CARDINAL, or when assigning a value of CARDINAL
          that is > MAX(INTEGER), to  a  variable  of  type  INTEGER.   It  is
          detected, unless runtime checking is disabled.

     heap overflow
          This might happen when ALLOCATE fails.

     case error
          This error occurs when non of the cases  in  a  CASE  statement  are
          selected, and the CASE statement has no ELSE part.

     stack size of process too large
          This is most likely to happen if the reserved space for a  coroutine
          stack  is  too  small.   In this case, increase the size of the area
          given to NEWPROCESS.  It can also happen if the stack needed for the
          main  process  is too large and there are coroutines.  In this case,
          the only fix is to reduce the stack size needed by the main process,
          f.i. by avoiding local arrays.

     too many nested traps + handlers
          This error can only occur when the user has installed his  own  trap
          handler.  It means that during execution of the trap handler another
          trap has occurred, and that several times.  In some cases,  this  is
          an error because of overflow of some internal tables.

     no RETURN from function procedure
          This error occurs when a function procedure does not return properly
          ('falls' through).

     illegal instruction
          This error might occur when you use floating point operations on  an
          implementation that does not have floating point.

     In addition, some of the library modules may give  error  messages.   The
     Traps-module has a suitable mechanism for this.

  The procedure call interface

     Parameters  are  pushed  on  the  stack  in  reversed  order.   For   VAR
     parameters,  its  address is passed, for value parameters its value.  The
     only exception to this rule is with conformant  arrays.   For  conformant
     arrays,  the  address  is passed, and an array descriptor is passed.  The
     descriptor is an EM array descriptor.  It consists of three  fields:  the
     lower  bound  (always  0), upper bound - lower bound, and the size of the
     elements.  The descriptor is pushed first.  If the parameter is  a  value
     parameter,  the  called  routine  must  make sure that its value is never
     changed, for instance by making its own copy of the array.

     When the size of the return value of a function procedure is larger  than
     the  maximum  of  SIZE(LONGREAL)  and  twice the pointer-size, the caller
     reserves this space on the stack,  above  the  parameters.   Callee  then
     stores its result there, and returns no other value.

  The Modula-2 runtime library

     The definition modules of the modules available  in  the  Minix  Modula-2
     runtime library reside in the directory /usr/lib/ack/m2.

  References

     [1]  Niklaus Wirth, Programming in Modula-2,  third,  corrected  edition,
          Springer-Verlag, Berlin (1985)

     [2]  Niklaus Wirth,  Programming  in  Modula-2,  Stringer-Verlag,  Berlin
          (1983)

     [3]  A.S.Tanenbaum,  J.W.Stevenson,  Hans   van   Staveren,   E.G.Keizer,
          Description  of a machine architecture for use with block structured
          languages, Informatica rapport IR-81, Vrije Universiteit, Amsterdam

THE MINIX PASCAL COMPILER




     This section refers to the (1982) BSI standard  for  Pascal  [1].   Minix
     Pascal  complies  with the requirements of level 1 of BS 6192: 1982, with
     the exceptions as listed in this section.

     The  standard  requires   an   accompanying   document   describing   the
     implementation-defined   and   implementation-dependent   features,   the
     reaction on errors and the extensions to  standard  Pascal.   These  four
     items will be treated in the rest of this section.

  Implementation-defined features

     For each implementation-defined feature mentioned in the BSI standard  we
     give  the  section  number,  the  quotation  from  that  section  and the
     definition.  First we quote the definition of implementation-defined:

          Possibly  differing  between  processors,  but   defined   for   any
          particular processor.

     BS 6.1.7: Each string-character shall  denote  an  implementation-defined
     value of the required char-type.

           All 7-bit ASCII characters except linefeed LF (10) are allowed.

     BS 6.4.2.2: The values of type real shall  be  an  implementation-defined
     subset  of  the real numbers denoted as specified by 6.1.5 by the signed-
     real values.

           The set of real values range from a low of -1.7976931348623157e+308
           to a high of 1.7976931348623157e+308.

     BS 6.4.2.2:  The  type  char  shall  be  the  enumeration  of  a  set  of
     implementation-defined   characters,   some   possibly   without  graphic
     representations.

           The 7-bit ASCII character set is used, where LF  (10)  denotes  the
           end-of-line marker on text-files.

     BS 6.4.2.2: The ordinal numbers of the character values shall  be  values
     of integer-type, that are implementation-defined, and that are determined
     by mapping the character values on to  consecutive  non-negative  integer
     values starting at zero.

           The  normal  ASCII  ordering  is  used:  ord('0')=48,  ord('A')=65,
           ord('a')=97, etc.

     BS 6.6.5.2: The post-assertions imply  corresponding  activities  on  the
     external  entities, if any, to which the file-variables are bound.  These
     activities, and the point at which they are actually performed, shall  be
     implementation-defined.

           The reading and writing writing of objects on  files  is  buffered.
           This  means  that  when a program terminates abnormally, I/O may be
           unfinished.  Terminal I/O is unbuffered.  Files are closed whenever
           they are rewritten or reset, or on program termination.

     BS 6.7.2.2: The predefined constant maxint shall be of  integer-type  and
     shall   denote   an  implementation-defined  value,  that  satisfies  the
     following conditions:

     (a)  All integral values in the closed interval from -maxint  to  +maxint
          shall be values of the integer-type.

     (b)  Any monadic operation performed on an integer value in this interval
          shall be correctly performed according to the mathematical rules for
          integer arithmetic.

     (c)  Any dyadic integer operation on two  integer  values  in  this  same
          interval  shall be correctly performed according to the mathematical
          rules for integer arithmetic, provided that the result  is  also  in
          this interval.

     (d)  Any relational operation on two integer values in this same interval
          shall be correctly performed according to the mathematical rules for
          integer arithmetic.

          The representation of integers under 16-bit Minix-PC or under  68000
          Minix  is  a  16-bit  word  using  two's complement arithmetic.  The
          integers range from -32768 to +32767.  Under 32-bit Minix-PC  a  32-
          bit integer is used ranging from -2147483648 to +2147483647.

     BS 6.7.2.2: The result of the real  arithmetic  operators  and  functions
     shall  be  approximations to the corresponding mathematical results.  The
     accuracy of this approximation shall be implementation-defined

           The default size of reals is 8 bytes, the accuracy is 11  bits  for
           the exponent, and 53 bits for the mantissa.  This gives an accuracy
           of about 16 digits.  and exponents ranging from -307 to +307.

     BS 6.9.3.1: The default TotalWidth values for integer, Boolean  and  real
     types shall be implementation-defined.

           The defaults are:

                 integer   6   (16-bit)
                 integer   11  (32-bit)
                 Boolean   5
                 real      14



     BS 6.9.3.4.1: ExpDigits, the number of digits written in an exponent part
     of a real, shall be implementation-defined.

           ExpDigits is defined as 3.

     BS 6.9.3.4.1: The character written as part of the  representation  of  a
     real   to   indicate   the  beginning  of  the  exponent  part  shall  be
     implementation-defined, either 'E' or 'e'.

           The exponent part starts with 'e'.

     BS 6.9.3.5: The case of the characters written as representation  of  the
     Boolean values shall be implementation-defined.

           The representations of true and false are 'true' and 'false'.

     BS 6.9.5: The effect caused by the standard procedure page on a text file
     shall be implementation-defined.

           The ASCII character form feed FF (12) is written.

     BS 6.10: The binding of the variables denoted by  the  program-parameters
     to  entities  external  to the program shall be implementation-defined if
     the variable is of a file-type.

           The program parameters must be files  and  all,  except  input  and
           output, must be declared as such in the program block.

     The program parameters input and output, if  specified,  will  correspond
     with the UNIX streams 'standard input' and 'standard output'.

     The other program parameters will  be  mapped  to  the  argument  strings
     provided  by  the  caller  of  this  program.   The  argument strings are
     supposed to be path names of the files to  be  opened  or  created.   The
     order  of  the  program  parameters  determines  the  mapping:  the first
     parameter is mapped onto the first argument string, etc.  Note that input
     and output are ignored in this mapping.

     The mapping is recalculated each time a program parameter is  opened  for
     reading or writing by a call to the standard procedures reset or rewrite.
     This gives the programmer the  opportunity  to  manipulate  the  list  of
     string  arguments  using  the external procedures argc, argv and argshift
     available in the Pascal library.

     BS 6.10: The effect of an  explicit  use  of  reset  or  rewrite  on  the
     standard text files input or output shall be implementation-defined.

           The procedures reset and rewrite are no-ops if applied to input  or
           output.

  Implementation-dependent features

     For each implementation-dependent feature mentioned in the BSI  standard,
     we  give  the section number, the quotation from that section and the way
     this feature is treated by the Minix Pascal system.  First we  quote  the
     definition of 'implementation-dependent':

          Possibly differing between processors and  not  necessarily  defined
          for any particular processor.

     BS 6.7.2.1: The order of evaluation of the operands of a dyadic  operator
     shall be implementation-dependent.

           Operands are always evaluated, so the program part

                 if (p<>nil) and (p^.value<>0) then

           is probably incorrect.

     The left-hand operand of a dyadic operator  is  almost  always  evaluated
     before  the  right-hand  side.   Some  peculiar evaluations exist for the
     following cases:

     1.   The modulo operation is performed by a library routine to check  for
          negative values of the right operand.

     2.   The expression

                set1 <= set2

          where set1 and set2 are compatible set types  is  evaluated  in  the
          following steps:

                -   evaluate set2;
                -   evaluate set1;
                -   compute set2+set1;
                -   test set2 and set2+set1 for equality.


     3.   The expression

                set1 >= set2

          where set1 and set2 are compatible set types  is  evaluated  in  the
          following steps:

                -   evaluate set1;
                -   evaluate set2;
                -   compute set1+set2;
                -   test set1 and set1+set2 for equality.

     BS 6.7.3: The order of evaluation, accessing and binding of  the  actual-
     parameters for functions shall be implementation-dependent.

           The order of evaluation is from right to left.

     BS 6.8.2.2: The decision as to the order of accessing  the  variable  and
     evaluating   the   expression   in   an  assignment-statement,  shall  be
     implementation-dependent.

           The expression is evaluated first.

     BS 6.8.2.3: The order of evaluation and binding of the  actual-parameters
     for procedures shall be implementation-dependent.

           The same as for functions.

     BS 6.9.5: The effect  of  inspecting  a  text  file  to  which  the  page
     procedure was applied during generation is implementation-dependent.

           The formfeed character written by page is  treated  like  a  normal
           character, with ordinal value 12.

     BS 6.10: The binding of the variables denoted by  the  program-parameters
     to  entities  external  to  the program shall be implementation-dependent
     unless the variable is of a file-type.

           Only variables of a file-type are allowed as program parameters.

  Error handling

     There are three classes of errors to  be  distinguished.   In  the  first
     class are the error messages generated by the compiler.  The second class
     consists of  the  occasional  errors  generated  by  the  other  programs
     involved  in  the compilation process.  Errors of the third class are the
     errors as defined in the standard by:

          An error is a violation by a program of  the  requirements  of  this
          standard that a processor is permitted to leave undetected.

     Compiler errors

     Error are written on the standard error output.  Each line has the form:

           <file>, line <number>: <description>

     Every time the compiler detects an error that does not have influence  on
     the  code  produced by the compiler or on the syntax decisions, a warning
     messages is given.  If only warnings are generated, compilation  proceeds
     and probably results in a correctly compiled program.

     Sometimes the compiler produces several errors for the same  line.   They
     are  only  shown  up to a maximum of 5 errors per line.  Warning are also
     shown up to a maximum of 5 per line.

     Extensive treatment of these errors is outside the scope of this manual.

     Runtime errors

     Errors detected at run time cause an error message to be generated on the
     diagnostic  output stream (UNIX file descriptor 2).  The message consists
     of the name of the program followed by a message  describing  the  error,
     possibly  followed  by  the  source line number.  Unless the -n option is
     turned on, the compiler generates code to keep track of which source line
     causes which instructions to be generated.

     For each error mentioned in the standard we give the section number,  the
     quotation  from  that  section and the way it is processed by the Pascal-
     compiler or runtime system.

     For detected errors the corresponding message and trap number are  given.
     Trap  numbers are useful for exception-handling routines.  Normally, each
     error causes the  program  to  terminate.   By  using  exception-handling
     routines  one  can ignore errors or perform alternate actions.  Only some
     of the errors can be  ignored  by  restarting  the  failing  instruction.
     These  errors  are  marked  as non-fatal, all others as fatal.  A list of
     errors with trap number between 0 and 63 (EM errors) can be found in [2].
     Errors  with  trap  number  between 64 and 127 (Pascal errors) are listed
     below.

     BS 6.4.6: It shall be an error if a value of type T2 must be  assignment-
     compatible with type T1, while T1 and T2 are compatible ordinal-types and
     the value of type T2 is not in the closed interval specified by T1.

           The compiler distinguishes between array-index expressions and  the
           other places where assignment-compatibility is required.

     Array subscripting errors are only detected when the 'A' option is  used.
     In  the other cases, a range bound error occurs when the value of type T2
     is not in the closed interval specified by T1, unless  range  checks  are
     disabled.

     BS 6.4.6: It shall be an error if a value of type T2 must be  assignment-
     compatible with type T1, while T1 and T2 are compatible set-types and any
     member of the value of type T2 is not in the closed interval specified by
     the base-type of the type T1.

           This error is not detected.



     BS 6.5.3.3: It shall be an error if a component of a  variant-part  of  a
     variant,  where  the  selector  of  the  variant-part  is not a field, is
     accessed unless the variant is active for the entirety of each  reference
     and access to each component of the variant.

           This error is not detected.

     BS 6.5.4: It shall be an error if the pointer-variable of an  identified-
     variable either denotes a nil-value or is undefined.

           This error is not detected.

     BS 6.5.4: It shall be an error to  remove  the  identifying-value  of  an
     identified  variable  from  its  pointer-type  when  a  reference  to the
     variable exists.

           When the identified variable is an element of the  record-variable-
           list  of  a  with-statement,  a  warning  is given at compile-time.
           Otherwise, this error is not detected.

     BS 6.5.5: It shall be an error to alter the value of  a  file-variable  f
     when a reference to the buffer-variable f^ exists.

           When f is altered when it is an element of the record-variable-list
           of a with-statement, a warning is given.  When a buffer-variable is
           used as a variable-parameter, an error is given.  This is  done  at
           compile-time.

     BS 6.6.5.2: It shall be an error if the  stated  pre-assertion  does  not
     hold  immediately  prior  to  any  use  of  the  file handling procedures
     rewrite, put, reset and get.

           For each  of  these  four  operations  the  pre-assertions  can  be
           reformulated as:

                 rewrite(f):  no pre-assertion.
                 put(f):      f is opened for writing and f^ is not undefined.
                 reset(f):    f exists.
                 get(f):      f is opened for reading and eof(f) is false.

           The following errors are detected for these operations:

           rewrite(f):
              more args expected, trap 64, fatal:
                 f is a program-parameter and the corresponding file  name  is
                 not supplied by the caller of the program.
              rewrite error, trap 101, fatal:
                 the caller of the program lacks the necessary  access  rights
                 to  create  the  file  in the file system or operating system
                 problems like table overflow prevent creation of the file.

           put(f):
              file not yet open, trap 72, fatal:
                 reset or rewrite are never applied to the file.   The  checks
                 performed by the run time system are not foolproof.
              not writable, trap 96, fatal:
                 f is opened for reading.
              write error, trap 104, fatal:
                 probably caused by file system problems.  For  instance,  the
                 file  storage  is  exhausted.   Because  I/O  is  buffered to
                 improve performance, it might happen that this  error  occurs
                 if  the  file  is closed.  Files are closed whenever they are
                 rewritten or reset, or on program termination.

           reset(f):
              more args expected, trap 64, fatal:
                 same as for rewrite(f).
              reset error, trap 100, fatal:
                 f does not exist,  or  the  caller  has  insufficient  access
                 rights, or operating system tables are exhausted.

           get(f):
              file not yet open, trap 72, fatal:
                 as for put(f).
              not readable, trap 97, fatal:
                 f is opened for writing.
              end of file, trap 98, fatal:
                 eof(f) is true just before the call to get(f).
              read error, trap 103, fatal:
                 unlikely to happen.  Probably caused by hardware problems  or
                 by  errors  elsewhere in your program that destroyed the file
                 information maintained by the run time system.
              truncated, trap 99, fatal:
                 the file is not properly formed by an integer number of  file
                 elements.   For  instance,  the  size of a file of integer is
                 odd.
              non-ASCII char read, trap 106, non-fatal:
                 the character value of the next character-type  file  element
                 is out of range (0..127).  Only for text files.

     BS 6.6.5.3: It shall be an error if a variant of  a  variant-part  within
     the  new  variable becomes active and a different variant of the variant-
     part is one of the specified variants.

           This error is not detected.

     BS 6.6.5.3: It shall be an error to use  dispose(q)  if  the  identifying
     variable has been allocated using the form new(p,c1,...,cn).



           This error is not detected.  However, this  error  can  cause  more
           memory to be freed then was allocated.  Dispose causes a fatal trap
           73 when memory already on the free list is freed again.

     BS 6.6.5.3: It shall be an  error  to  use  dispose(q,k1,...,km)  if  the
     identifying  variable  has been allocated using the form new(p,c1,...,cn)
     and m is not equal to n.

           This error is not detected.  However, this  error  can  cause  more
           memory to be freed then was allocated.  Dispose causes a fatal trap
           73 when memory already on the free list is freed again.

     BS 6.6.5.3: It shall be an error if the variants  of  a  variable  to  be
     disposed  are  different  from  those  specified by the case-constants to
     dispose.

           This error is not detected.

     BS 6.6.5.3: It shall be an error if the value of the pointer parameter of
     dispose has nil-value or is undefined.

           This error is detected  for  nil-value  (dispose  error,  trap  73,
           fatal).

     BS 6.6.5.3: It shall be an error if a variable created using  the  second
     form of new is accessed by the identified variable of the variable-access
     of a factor, of an assignment-statement, or of an actual-parameter.

           This error is not detected.

     BS 6.6.6.2: It shall be an error if the value of sqr(x) does not exist.

           This error is detected for real-type arguments (real overflow, trap
           4, non-fatal).

     BS 6.6.6.2: It shall be an error if x in ln(x) is smaller than  or  equal
     to 0.

           This error is detected (error in ln, trap 66, non-fatal)

     BS 6.6.6.2: It shall be an error if x in sqrt(x) is smaller than 0.

           This error is detected (error in sqrt, trap 67, non-fatal)

           In addition to these errors, overflow in the expression  exp(x)  is
           detected  (error in exp, trap 65, non-fatal; real overflow, trap 4,
           non-fatal)



     BS 6.6.6.3: It shall be an error if the integer value  of  trunc(x)  does
     not exist.

           This error is detected (conversion error, trap 10, non-fatal).

     BS 6.6.6.3: It shall be an error if the integer value  of  round(x)  does
     not exist.

           This error is detected (conversion error, trap 10, non-fatal).

     BS 6.6.6.4: It shall be an error if the integer value of ord(x) does  not
     exist.

           This error can not occur, because the compiler will not allow  such
           ordinal types.

     BS 6.6.6.4: It shall be an error if the character value  of  chr(x)  does
     not exist.

           This error is detected (range bound error, trap 1, non-fatal).

     BS 6.6.6.4: It shall be an error if the value of succ(x) does not exist.

           Same comments as for chr(x).

     BS 6.6.6.4: It shall be an error if the value of pred(x) does not exist.

           Same comments as for chr(x).

     BS 6.6.6.5: It shall be an error if f in eof(f) is undefined.

           This error is detected (file not yet open, trap 72, fatal).

     BS 6.6.6.5: It shall be an error if f in  eoln(f)  is  undefined,  or  if
     eof(f) is true at that time.

           The following errors may occur:

           file not yet open, trap 72, fatal;
           not readable, trap 97, fatal;
           end of file, trap 98, fatal.

     BS 6.7.1: It shall be an error if a variable-access used as an operand in
     an expression is undefined at the time of its use.

           The compiler performs some limited checks to see if identifiers are
           used  before  they  are  set.  Since it can not always be sure (one
           could, for instance, jump  out  of  a  loop),  only  a  warning  is
           generated.   When  an expression contains a function-call, an error
           occurs if the function is not assigned at run-time.

     BS 6.7.2.2: A term of the form x/y shall be an error if y is zero.

           This error is detected (divide by 0.0, trap 7, non-fatal).

     BS 6.7.2.2: It shall be an error if j is zero in 'i div j'.

           This error is detected (divide by 0, trap 6, non-fatal).

     BS 6.7.2.2: It shall be an error if j is zero or negative in i MOD j.

           This error is detected (only positive j in 'i mod j', trap 71, non-
           fatal).

     BS 6.7.2.2: It shall be an error  if  the  result  of  any  operation  on
     integer operands is not performed according to the mathematical rules for
     integer arithmetic.

           This implementation does not detect integer overflow.

     BS 6.8.3.5: It shall be an error if none of the case-constants  is  equal
     to the value of the case-index upon entry to the case-statement.

           This error is detected (case error, trap 20, fatal).

     BS 6.9.1: It shall be an error if the sequence of characters read looking
     for an integer does not form a signed-integer as specified in 6.1.5.

           This error is detected (digit expected, trap 105, non-fatal).

     BS 6.9.1: It shall be an error if the sequence of characters read looking
     for a real does not form a signed-number as specified in 6.1.5.

           This error is detected (digit expected, trap 105, non-fatal).

     BS 6.9.1: When read is applied to f, it shall be an error if the  buffer-
     variable f^ is undefined or the pre-assertions for get do not hold.

           This error is detected (see get(f)).

     BS 6.9.3: When write is applied to a text file f, it shall be an error if
     f is undefined or f is opened for reading.

           This error is detected (see put(f)).  Furthermore,  this  error  is
           also detected when f is not a text file.

     BS 6.9.3.1: The values of TotalWidth or FracDigits shall be greater  than
     or equal to one; it shall be an error if either value is less then one.



           When either value is less than zero, an error (illegal field width,
           trap  75,  non-fatal) occurs.  Zero values are allowed, in order to
           maintain some compatibility with the old Minix Pascal compiler.

     BS 6.9.5: It  shall  be  an  error  if  the  pre-assertion  required  for
     writeln(f) doe not hold prior to the invocation of page(f);

           This error is detected (see put(f)).

  Extensions to the standard

     1. External routines

     Except for the required directive 'forward'  the  Minix  Pascal  compiler
     recognizes  the  directive  'extern'.   This directive tells the compiler
     that the procedure block of this procedure will not  be  present  in  the
     current  program.   The  code  for  the  body  of  this procedure must be
     included at a later stage of the compilation process.

     This  feature  allows  one  to  build  libraries  containing  often  used
     routines.   These routines do not have to be included in all the programs
     using them.  Maintenance is much simpler if there  is  only  one  library
     module to be changed instead of many Pascal programs.

     Another advantage is that these library  modules  may  be  written  in  a
     different language, for instance C.

     The use of  external  routines,  however,  is  dangerous.   The  compiler
     normally  checks  for  the  correct  number and type of parameters when a
     procedure is called and for the result type of functions.  If an external
     routine  is  called these checks are not sufficient, because the compiler
     can not check whether the procedure heading of the  external  routine  as
     given  in  the  Pascal program matches the actual routine implementation.
     It should be the loader's task  to  check  this.   However,  the  current
     loaders are not that smart.

     For those who wish the use the interface between C and Pascal we give  an
     incomplete list of corresponding formal parameters in C and Pascal.


           Pascal                  C
           a:integer               int a
           a:char                  int a
           a:boolean               int a
           a:real                  double a
           a:^type                 type *a
           var a:type              type *a
           procedure a(pars)       struct {
                                       void (*a)() ;
                                       char *static_link ;

                                   }
           function a(pars):type   struct {
                                       type (*a)() ;
                                       char *static_link ;
                                   }

     The Pascal runtime system  uses  the  following  algorithm  when  calling
     function/procedures passed as parameters.

           if (static_link) {
                   (*a)(static_link, pars);
           } else {
                   (*a)(pars);
           }


     2. Separate compilation.

     The  compiler  is  able  to  (separately)   compile   a   collection   of
     declarations,  procedures  and  functions to form a library.  The library
     may be linked with the main program, compiled later.  The syntax of these
     modules is

           module = [constant-definition-part]
                    [type-definition-part]
                    [var-declaration-part]
                    [procedure-and-function-declaration-part]

     The compiler accepts a program or a module:

           unit = program | module

     All variables declared outside a module must be imported  by  parameters,
     even  the  files  input  and  output.  Access to a variable declared in a
     module is only possible using the procedures and  functions  declared  in
     that  same  module.   By  giving  the  correct procedure/function heading
     followed by the directive 'extern' you may use procedures  and  functions
     declared in other units.

     3. Assertions.

     When the s-option is off, Minix Pascal compiler recognizes an  additional
     statement,  the assertion.  Assertions can be used as an aid in debugging
     and documentation.  The syntax is:

           assertion = 'assert' Boolean-expression

     An assertion is a simple-statement, so

           simple-statement = [assignment-statement |
                               procedure-statement |
                               goto-statement |
                               assertion
                              ]

     An assertion causes an error if the Boolean-expression is false.  That is
     its  only  purpose.  It does not change any of the variables, at least it
     should not.  Therefore, do not use functions  with  side-effects  in  the
     Boolean-expression.   If  the  a-option is turned on, then assertions are
     skipped by the compiler. 'assert' is not a word-symbol (keyword) and  may
     be  used as identifier.  However, assignment to a variable and calling of
     a procedure with that name will be impossible.  If the s-option is turned
     on,  the  compiler  will  not  know  a  thing  about assertions, so using
     assertions will then give a parse error.

     4. Additional procedures.

     Three additional standard procedures are available:

     halt:
          a call of this procedure is equivalent to jumping to the end of your
          program.  It is always the last statement executed.  The exit status
          of the program may be supplied as optional  argument.   If  not,  it
          will be zero.

     release:

     mark:
          for most applications it is sufficient to use  the  heap  as  second
          stack.   Mark  and  release  are  suited  for this type of use, more
          suited than dispose.  mark(p), with p of type  pointer,  stores  the
          current  value  of  the  heap  pointer  in  p.  release(p),  with  p
          initialized by a call of mark(p), restores the heap pointer  to  its
          old  value.   All  the heap objects, created by calls of new between
          the call of mark and the call of release, are removed and the  space
          they  used  can be reallocated.  Never use mark and release together
          with dispose!

     5. UNIX interfacing.

     If the c-option is turned on, then some special features are available to
     simplify  an  interface  with  the  UNIX  environment.  First of all, the
     compiler allows you to use a different type of string  constants.   These
     string  constants  are delimited by double quotes ('"').  To put a double
     quote into these strings, you must repeat  the  double  quote,  like  the
     single  quote in normal string constants.  These special string constants
     are terminated by a zero byte (chr(0)).  The type of these constants is a
     pointer  to  a packed array of characters, with lower bound 1 and unknown
     upper bound.
     Secondly, the compiler predefines a new type identifier 'string' denoting
     this just described string type.

     The only thing you can do with these features is declaration of constants
     and  variables  of type 'string'.  String objects may not be allocated on
     the heap and string pointers  may  not  be  de-referenced.   Still  these
     strings  are  very  useful  in  combination  with external routines.  The
     procedure write  is  extended  to  print  these  zero-terminated  strings
     correctly.

     6. Double length (32 bit) integers.

     If the d-option is turned on, then the additional type 'long' is known to
     the   compiler.    Long  variables  have  integer  values  in  the  range
     -2147483648 .. +2147483647.  Long constants can not be  declared.   Longs
     can  not  be  used  as  control-variables.   It  is  not  allowed to form
     subranges of type long.  All operations  allowed  on  integers  are  also
     allowed  on longs and are indicated by the same operators: '+', '-', '*',
     '/', 'div', 'mod'.  The procedures read and write have been  extended  to
     handle  long  arguments  correctly.   It is possible to read longs from a
     file of integers and vice-versa, but only if longs and integers have  the
     same  size.   The default width for longs is 11.  The standard procedures
     'abs' and 'sqr' have been extended to work on long arguments.  Conversion
     from  integer to long, long to real, real to long and long to integer are
     automatic, like the conversion from integer to real.   These  conversions
     may cause a

          conversion error, trap 10, non-fatal

     7. Underscore as letter.

     The character '_' may be used in forming identifiers, if  the  u-  or  U-
     option  is  turned  on.   It  is  forbidden  to  start  identifiers  with
     underscores, since this may cause name-clashes with run-time routines.

     8. Zero field width in write.

     Zero TotalWidth arguments are allowed.  In this case, no  characters  are
     written  for  character,  string  or  Boolean  type  arguments.   A  zero
     FracDigits argument for fixed-point representation of  reals  causes  the
     fraction and the character '.' to be suppressed.

     9. Pre-processing.

     If the very first character of a file containing a Pascal program is  the
     sharp  ('#', ASCII 23(hex)) the file is preprocessed in the same way as C
     programs.  Lines beginning with a '#' are taken as  preprocessor  command
     lines  and  not  fed  to  the  Pascal compiler proper.  C style comments,
     /*......*/, are removed by the C preprocessor,  thus  C  comments  inside
     Pascal   programs  are  also  removed  when  they  are  fed  through  the
     preprocessor.

  Deviations from the standard

     Minix Pascal deviates from the standard in the following ways:

     1.   Standard procedures and functions are not allowed as  parameters  in
          Minix  Pascal.   You can obtain the same result with negligible loss
          of performance by declaring some user routines like:

                function sine(x:real):real;
                begin
                    sine:=sin(x)
                end;


     2.   The  standard  procedures  read,  readln,  write  and  writeln   are
          implemented as word-symbols, and can therefore not be redeclared.

  Compiler options

     Some options of the compiler may be controlled by using '{$....}'.   Each
     option  consists  of  a lower case letter followed by +, - or an unsigned
     number.  Options are separated by commas.  The following options exist:

     a+/- This option switches assertions on and off.  If this option  is  on,
          then code is included to test these assertions at run time.  Default
          +.

     c+/- This option, if on,  allows  you  to  use  C-type  string  constants
          surrounded  by  double  quotes.   Moreover,  a  new  type identifier
          'string' is predefined.  Default -.

     d+/- This option, if on, allows you to  use  variables  of  type  'long'.
          Default -.

     i<num>
          With this flag the setsize for a set of integers can be manipulated.
          The number must be the number of bits per set.  The default value is
          16.

     l+/- If + then code is inserted to keep track of the source line  number.
          When  this flag is switched on and off, an incorrect line number may
          appear if the error occurs in a part of your program for which  this
          flag is off.  Default +.

     r+/- If + then code is inserted to check subrange variables against lower
          and upper subrange limits.  Default +.

     s+/- If + then the compiler will hunt for places in  your  program  where
          non-standard  features  are  used,  and for each place found it will
          generate a warning.  Default -.

     t+/- If + then each time a procedure is entered, the routine  'procentry'
          is called, and each time a procedure exits, the procedure 'procexit'
          is called.  Both 'procentry'  and  'procexit'  have  a  'string'  as
          parameter.   This  means  that  when a user specifies his or her own
          procedures, the c-option  must  be  used.   Default  procedures  are
          present in the run time library.  Default -.

     u+/- If + then the character '_' is treated like a letter, so that it may
          be  used in identifiers.  Procedure and function identifiers are not
          allowed to start with an underscore because they  may  collide  with
          library routine names.  Default -.

     Some of these flags (c, d, i, s, u, C and U) are only effective when they
     appear  before  the  'program' symbol.  The others may be switched on and
     off.

     A very  powerful  debugging  tool  is  the  knowledge  that  inaccessible
     statements and useless tests are removed by the optimizer.  For instance,
     a statement like:

           if debug then
               writeln('initialization done');

     is completely removed by the optimizer if debug is a constant with  value
     false.  The first line is removed if debug is a constant with value true.
     Of course, if debug is a variable nothing can be removed.

  Library routines

     The following  library  of  external  routines  for  Pascal  programs  is
     available:

     const       bufsize = ?;
     type        br1 =  1..bufsize;
                 br2 =  0..bufsize;
                 br3 = -1..bufsize;
                 ok = -1..0;
                 buf = packed array[br1] of char;
                 alfa = packed array[1..8] of char;
                 string = ^packed array[1..?] of char;
                 filetype = file of ?;
                 long = ?;

     {all routines must be declared extern}

     function    argc:integer;
     function    argv(i:integer):string;
     function    environ(i:integer):string;
     procedure   argshift;

     procedure   buff(var f:filetype);
     procedure   nobuff(var f:filetype);
     procedure   notext(var f:text);
     procedure   diag(var f:text);
     procedure   pcreat(var f:text; s:string);
     procedure   popen(var f:text; s:string);
     procedure   pclose(var f:filetype);

     procedure   trap(err:integer);
     procedure   encaps(procedure p; procedure q(n:integer));

     function    perrno:integer;
     function    uread(fd:integer; var b:buf; len:br1):br3;
     function    uwrite(fd:integer; var b:buf; len:br1):br3;

     function    strbuf(var b:buf):string;
     function    strtobuf(s:string; var b:buf; len:br1):br2;
     function    strlen(s:string):integer;
     function    strfetch(s:string; i:integer):char;
     procedure   strstore(s:string; i:integer; c:char);

     function    clock:integer;

     This library contains  some  often  used  external  routines  for  Pascal
     programs.  The routines can be divided into several categories:

   Argument control:

     argc      Gives the number of arguments  provided  when  the  program  is
               called.

     argv      Selects the specified  argument  from  the  argument  list  and
               returns  a  pointer to it.  This pointer is nil if the index is
               out of bounds (<0 or >=argc).

     environ   Returns a  pointer  to  the  i-th  environment  string  (i>=0).
               Returns  nil  if  i  is  beyond the end of the environment list
               (UNIX version 7).

     argshift  Effectively deletes the first argument from the argument  list.
               Its  function is equivalent to shift in the UNIX shell: argv[2]
               becomes argv[1], argv[3] becomes argv[2], etc.  It is a  useful
               procedure  to  skip  optional  flag  arguments.   Note that the
               matching of arguments and files is done at the time a  file  is
               opened by a call to reset or rewrite.

   Additional file handling routines:



     buff      Turn on buffering of a file.   Not  very  useful,  because  all
               files  are  buffered  except  standard output to a terminal and
               diagnostic output.  Input files are always buffered.

     nobuff    Turn off buffering of an output file.  It  causes  the  current
               contents of the buffer to be flushed.

     notext    Only useful for input files.  End of line  characters  are  not
               replaced  by a space and character codes out of the ASCII range
               (0..127) do not cause an error message.

     diag      Initialize a file for output on the  diagnostic  output  stream
               (fd=2).  Output is not buffered.

     pcreat    The same as rewrite(f), except that you must provide  the  file
               name  yourself.   The  name must be zero terminated.  Only text
               files are allowed.

     popen     The same as reset(f), except that you  must  provide  the  file
               name  yourself.   The  name must be zero terminated.  Only text
               files are allowed.

     pclose    Gives you the opportunity to close files hidden in  records  or
               arrays.  All other files are closed automatically.

   String handling:

     strbuf    Type conversion from character array to string.  It is your own
               responsibility that the string is zero terminated.

     strtobuf  Copy string into buffer until the string terminating zero  byte
               is  found  or  until  the buffer if full, whatever comes first.
               The zero byte is also copied.  The number of copied characters,
               excluding  the  zero  byte,  is  returned.  So if the result is
               equal to the buffer length, then the end of buffer  is  reached
               before the end of string.

     strlen    Returns the string length excluding the terminating zero byte.

     strfetch  Fetches the i-th character from a string.  There  is  no  check
               against the string length.

     strstore  Stores a character in a string.   There  is  no  check  against
               string length, so this is a dangerous procedure.

   Trap handling:

     These routines  allow  you  to  handle  almost  all  the  possible  error
     situations yourself.  You may define your own trap handler, replacing the
     default handler that produces an error message and quits.  You  may  also
     generate traps yourself.

     trap      Trap generates the trap passed as argument (0..252).  The  trap
               numbers 128..252 may be used freely.  The others are reserved.

     encaps    Encapsulate the execution of p with the trap handler q.  Encaps
               replaces  the  previous trap handler by q, calls p and restores
               the previous handler when p returns.  If, during the  execution
               of  p,  a trap occurs, then q is called with the trap number as
               parameter.  For the duration of q the previous trap handler  is
               restored,  so that you may handle only some of the errors in q.
               All the other errors must then be raised again  by  a  call  to
               trap.
               Encapsulations may be nested: you may encapsulate  a  procedure
               while executing an encapsulated routine.
               Jumping out of an encapsulated procedure  (non-local  goto)  is
               dangerous,  because the previous trap handler must be restored.
               Therefore, you may only jump out of procedure p from  inside  q
               and  you  may  only jump out of one level of encapsulation.  If
               you want to exit several levels of  encapsulation,  use  traps.
               See pc_prlib(7) for lists of trap numbers for EM machine errors
               and Pascal run time system errors.  Note that p  may  not  have
               parameters.

   UNIX system calls:

     uread     Equal to the read system call.  Its normal name is  blocked  by
               the standard Pascal routine read.

     uwrite    As above but for write(2).

     perrno    Because external data references are not  possible  in  Pascal,
               this  routine returns the global variable errno, indicating the
               result of the last system call.

   Miscellaneous:

     clock     Return the number of ticks of user and system time consumed  by
               the program.

     The following program presents an example of how these  routines  can  be
     used.  This program is equivalent to the UNIX command cat(1).

     {$c+}
     program cat(input,inp,output);
     var         inp:text;
                 s:string;

     function argc:integer; extern;
     function argv(i:integer):string; extern;
     procedure argshift; extern;
     function strlen(s:string):integer; extern;
     function strfetch(s:string; i:integer):char; extern;

     procedure copy(var fi:text);
     var c:char;
     begin reset(fi);
       while not eof(fi) do
       begin
         while not eoln(fi) do
         begin
           read(fi,c);
           write(c)
         end;
         readln(fi);
         writeln
       end
     end;

     begin  {main}
       if argc = 1 then
                 copy(input)
       else
         repeat
           s := argv(1);
           if (strlen(s) = 1) and (strfetch(s,1) = '-')
           then copy(input)
           else copy(inp);
           argshift;
         until argc <= 1;
     end.

     Another example gives some idea of the way to manage trap handling:

     program bigreal(output);
     const EFOVFL=4;
     var trapped:boolean;

     procedure encaps(procedure p; procedure q(n:integer)); extern;
     procedure trap(n:integer); extern;

     procedure traphandler(n:integer);
     begin if n=EFOVFL then trapped:=true else trap(n) end;

     procedure work;
     var i,j:real;
     begin trapped:=false; i:=1;
       while not trapped do
         begin j:=i; i:=i*2 end;
       writeln('bigreal = ',j);

     end;

     begin
       encaps(work,traphandler);
     end.

     Two routines may cause fatal error messages to be generated.  These are:

     pcreat    Rewrite error (trap 77) if the file cannot be created.

     popen     Reset error (trap 76) if the file cannot be opened for reading

  References

     [1]       BSI standard BS 6192: 1982 (ISO 7185).

     [2]       A.S.Tanenbaum, J.W.Stevenson, Hans  van  Staveren,  E.G.Keizer,
               "Description  of  a  machine  architecture  for  use with block
               structured languages", Informatica rapport IR-81.