ack/doc/m2ref.doc

.TL
The ACK Modula-2 Compiler
.AU
Ceriel J.H. Jacobs
.AI
Department of Mathematics and Computer Science
Vrije Universiteit
Amsterdam
The Netherlands
.AB no
.AE
.NH
Introduction
.PP
This document describes the implementation-specific features of the
ACK 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].
.PP
The ACK Modula-2 compiler is currently available for use with the VAX,
Motorola MC68020, Motorola MC68000,
PDP-11, and Intel 8086 code-generators.
For the 8086, MC68000, and MC68020,
floating point emulation is used. This is made available with the \fI-fp\fP
option, which must be passed to \fIack\fP[4,5].
.NH
The language implemented
.PP
This section 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.
.PP
Basically, the compiler recognizes the language as described in [1], and
most of [2], for backwards compatibility. It warns the user for old-fashioned
constructions (constructions that [1] does not allow).
If the \fI-Rm2-3\fP option (see [6]) is passed to \fIack\fP, this backwards
compatibility feature is disabled. Also, it may not be present on some
smaller machines, like the PDP-11.
.NH 2
Syntax (section 2)
.PP
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.
.NH 2
Vocabulary and Representation (section 3)
.PP
The input "\f510..\fP" is parsed as two tokens: "\f510\fP" and "\f5..\fP".
.PP
The empty string \f5""\fP has type
.DS
.ft 5
ARRAY [0 .. 0] OF CHAR
.ft P
.DE
and contains one character: \f50C\fP.
.PP
When the text of a comment starts with a '\f5$\fP', it may be a pragma.
Currently, two pragmas exist:
.DS
.ft 5
(*$F (F stands for Foreign) *)
(*$R[+|-] (Runtime checks, on or off) *)
.ft P
.DE
The Foreign pragma is only meaningful in a \f5DEFINITION MODULE\fP,
and indicates that this
\f5DEFINITION MODULE\fP describes an interface to a module written in another
language (for instance C, Pascal, or EM).
Runtime checks are: range-checks, checks when assigning CARDINALS to INTEGERS
and vice versa, and checks that FOR-loop control-variables are not changed
in the body of the loop.
.PP
Constants of type \f5LONGINT\fP are integers with a suffix letter \f5D\fP
(for instance \f51987D\fP).
Constants of type \f5LONGREAL\fP have suffix \f5D\fP if a scale factor is missing,
or have \f5D\fP in place of \f5E\fP in the scale factor (f.i. \f51.0D\fP,
\f50.314D1\fP).
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.
.NH 2
Declarations and scope rules (section 4)
.PP
Standard identifiers are considered to be predeclared, and valid in all
parts of a program. They are called \fIpervasive\fP.
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.
.NH 2
Constant expressions (section 5)
.PP
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 typetransfer with a constant argument, or
one of the standard procedures 
\f5ABS\fP, \f5CAP\fP, \f5CHR\fP, \f5LONG\fP, \f5MAX\fP, \f5MIN\fP, \f5ODD\fP, \f5ORD\fP,
\f5SIZE\fP, \f5SHORT\fP, \f5TSIZE\fP, or \f5VAL\fP, with constant argument(s);
\f5TSIZE\fP and \f5SIZE\fP may also have a variable as argument.
.PP
Floating point expressions are never evaluated compile time, because
the compiler basically functions as a cross-compiler, and thus cannot
use the floating point instructions of the machine on which it runs.
Also, \f5MAX(REAL)\fP and \f5MIN(REAL)\fP are not allowed.
.NH 2
Type declarations (section 6)
.NH 3
Basic types (section 6.1)
.PP
The type \f5CHAR\fP includes the ASCII character set as a subset. Values range from
\f50C\fP to \f5377C\fP, not from \f50C\fP to \f5177C\fP.
.NH 3
Enumerations (section 6.2)
.PP
The maximum number of enumeration literals in any one enumeration type
is \f5MAX(INTEGER)\fP.
.NH 3
Record types (section 6.5)
.PP
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.
However, see section 2.
.NH 3
Set types (section 6.6)
.PP
The only limitation imposed by the compiler is that the base type of the
set must be a subrange type, an enumeration type, \f5CHAR\fP, or
\f5BOOLEAN\fP. So, the lower bound
does not have to be positive. However, if a negative lower bound is used,
the compiler gives a warning of the \fIrestricted\fP class (see the manual
page of the compiler).
.PP
The standard type \f5BITSET\fP is defined as
.DS
.ft 5
TYPE BITSET = SET OF [0 .. 8*SIZE(INTEGER)-1];
.ft P
.DE
.NH 2
Expressions (section 8)
.NH 3
Operators (section 8.2)
.NH 4
Arithmetic operators (section 8.2.1)
.PP
The Report does not specify the priority of the unary operators \f5+\fP or \f5-\fP:
It does not specify whether
.DS
.ft 5
- 1 + 1
.ft P
.DE
means
.DS
.ft 5
- (1 + 1)
.ft P
.DE
or
.DS
.ft 5
(-1) + 1
.ft P
.DE
I have seen some compilers that implement the first alternative, and others
that implement the second. Our compiler implements the second, which is
suggested by the fact that their priority is not specified, which might
indicate that it is the same as that of their binary counterparts.
And then the rule about left to right decides for the second.
On the other hand, one might argue that, since the grammar only allows
for one unary operator in a simple expression, it must apply to the
whole simple expression, not just the first term.
.NH 2
Statements (section 9)
.NH 3
Assignments (section 9.1)
.PP
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.
.PP
The types \f5INTEGER\fP and \f5CARDINAL\fP are assignment-compatible with
\f5LONGINT\fP, and \f5REAL\fP is assignment-compatible with \f5LONGREAL\fP.
.NH 3
Case statements (section 9.5)
.PP
The size of the type of the case-expression must be less than or equal to
the word-size.
.PP
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 \f5ELSE\fP-part.
In our implementation, this results in a runtime error.
.NH 3
For statements (section 9.8)
.PP
The Report does not specify the legal types for a control variable.
Our implementation allows the basic types (except \f5REAL\fP),
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.
.NH 3
Return and exit statements (section 9.11)
.PP
The Report does not specify which result-types are legal.
Our implementation allows any result type.
.NH 2
Procedure declarations (section 10)
.PP
Function procedures must exit through a RETURN statement, or a runtime error
occurs.
.NH 3
Standard procedures (section 10.2)
.PP
Our implementation supports \f5NEW\fP and \f5DISPOSE\fP
for backwards compatibility,
but issues warnings for their use. However, see section 2.
.PP
Also, some new standard procedures were added, similar to the new standard
procedures in Wirth's newest compiler:
.IP \-
\f5LONG\fP converts an argument of type \f5INTEGER\fP or \f5REAL\fP to the types \f5LONGINT\fP or
\f5LONGREAL\fP.
.IP \-
\f5SHORT\fP performs the inverse transformation, without range-checks.
.IP \-
\f5FLOATD\fP is analogous to \f5FLOAT\fP, but yields a result of type
\f5LONGREAL\fP.
.IP \-
\f5TRUNCD\fP is analogous to \f5TRUNC\fP, but yields a result of type
\f5LONGINT\fP.
.NH 2
System-dependent facilities (section 12)
.PP
The type \f5BYTE\fP is added to the \f5SYSTEM\fP module.
It occupies a storage unit of 8 bits.
\f5ARRAY OF BYTE\fP has a similar effect to \f5ARRAY OF WORD\fP, but is
safer. In some obscure cases the \f5ARRAY OF WORD\fP mechanism does not quite
work properly.
.PP
The procedure \f5IOTRANSFER\fP is not implemented.
.NH 1
Compile time errors
.PP
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.
Currently, there are 3 classifications:
.IP "(old-fashioned use)"
.br
These warnings are given on constructions that are not allowed by [1], but are
allowed by [2].
.IP (strict)
.br
These warnings are given on constructions that are supported by the 
ACK Modula-2 compiler, but might not be supported by others.
Examples: functions returning structured types, SET types of subranges with
negative lower bound.
.IP (warning)
.br
The other warnings, such as warnings about variables that are never assigned,
never used, etc.
.NH 1
Runtime errors
.PP
The ACK Modula-2 compiler produces code for an EM machine as defined in [3].
Therefore, it depends on the implementation
of the EM machine for detection some of the runtime errors that could occur.
.PP
The \fITraps\fP module enables the use 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
\fITraps\fP definition module.
.PP
The following runtime errors may occur:
.IP "array bound error"
.br
The detection of this error depends on the EM implementation.
.IP "range bound error"
.br
Range bound errors are always detected, unless runtime checks are disabled.
.IP "set bound error"
.br
The detection of this error depends on the EM implementation.
The current implementations detect this error.
.IP "integer overflow"
.br
The detection of this error depends on the EM implementation.
.IP "cardinal overflow"
.br
This error is detected, unless runtime checks are disabled.
This message is also given on cardinal underflow.
.IP "real overflow"
.br
The detection of this error depends on the EM implementation.
.IP "real underflow"
.br
The detection of this error depends on the EM implementation.
.IP "divide by 0"
.br
The detection of this error depends on the EM implementation.
.IP "divide by 0.0"
.br
The detection of this error depends on the EM implementation.
.IP "undefined integer"
.br
The detection of this error depends on the EM implementation.
.IP "undefined real"
.br
The detection of this error depends on the EM implementation.
.IP "conversion error"
.br
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.
.IP "stack overflow"
.br
The detection of this error depends on the EM implementation.
.IP "heap overflow"
.br
The detection of this error depends on the EM implementation.
Might happen when ALLOCATE fails.
.IP "case error"
.br
This error occurs when non of the cases in a CASE statement are selected,
and the CASE statement has no ELSE part.
The detection of this error depends on the EM implementation.
All current EM implementations detect this error.
.IP "stack size of process too large"
.br
The current implementation limits the stack size of processes to 1024 bytes.
.IP "too many nested traps + handlers"
.br
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.
.IP "no RETURN from procedure function"
.br
This error occurs when a procedure function does not return properly
("falls" through).
.IP "illegal instruction"
.br
This error might occur when you use floating point operations on an
implementation that does not have floating point.
.NH 1
Calling the compiler
.PP
See [4,5,6] for a detailed explanation.
.NH 1
References
.IP [1]
Niklaus Wirth,
.I
Programming in Modula-2, third, corrected edition,
.R
Springer-Verlag, Berlin (1985)
.IP [2]
Niklaus Wirth,
.I
Programming in Modula-2,
.R
Stringer-Verlag, Berlin (1983)
.IP [3]
A.S.Tanenbaum, J.W.Stevenson, Hans van Staveren, E.G.Keizer,
.I
Description of a machine architecture for use with block structured languages,
.R
Informatica rapport IR-81, Vrije Universiteit, Amsterdam
.IP [4]
UNIX manual \fIack\fP(1)
.IP [5]
UNIX manual \fImodula-2\fP(1)
.IP [6]
UNIX manual \fIem_m2\fP(6)