Migrating an Application from OpenVMS VAX to OpenVMS Alpha

For detailed information about extensions and language features, see the Fortran langauge reference manual, which visually shows extensions of the FORTRAN-77 standard.

Note
The Digital Fortran for OpenVMS Alpha product supports most of the FORTRAN-77 language extensions supported by DEC Fortran for OpenVMS Alpha and the ISO/ANSI Fortran 90 standard. For information on compatibility, see the DEC Fortran 90 User Manual for OpenVMS Alpha Systems.

The remainder of this section summarizes language features specific to Digital Fortran 77 for OpenVMS VAX Systems and Digital Fortran for OpenVMS Alpha, language features that are shared but interpreted differently in each language, Digital Fortran for OpenVMS Alpha restrictions that do not apply to Digital Fortran 77 for OpenVMS VAX Systems, and data porting considerations.

11.4.1.1 Language Features Specific to Digital Fortran for OpenVMS Alpha

The following language features are available in Digital Fortran for OpenVMS Alpha but are not supported in Digital Fortran 77 for OpenVMS VAX Systems Version 6.4:

Quotation marks (") as delimiters for character constants. This can be disabled by specifying the /VMS qualifier.
Naturally aligned or packed boundaries for fields of records and items in COMMON blocks
The INTEGER*1, INTEGER*8, and LOGICAL*8 data types
Support for S_floating, T_floating, and X_floating IEEE data types as well as support for non-native unformatted data file formats, including big-endian numeric format. For a description of the native floating-point data types for Alpha systems, see the Alpha Architecture Reference Manual.
LIB$ESTABLISH and LIB$REVERT are provided as intrinsic functions for compatibility with Digital Fortran 77 for OpenVMS VAX Systems condition handling.
DEC Fortran converts declarations to LIB$ESTABLISH to DEC Fortran RTL specific entry points.
The alternate "Z" spelling for double-precision complex intrinsic functions. (For example, the square root double-precision intrinsic function can be spelled as CDSQRT or ZSQRT.)
The following intrinsic functions:
- IMAG
- AND
- OR
- XOR
- LSHIFT
- RSHIFT
Certain run-time errors are specific to Digital Fortran for OpenVMS Alpha.
Case-sensitive names
I/O unit numbers can be any nonnegative integer in Digital Fortran for OpenVMS Alpha. In Digital Fortran 77 for OpenVMS VAX Systems, the values for I/O unit numbers can range from 0 to 99.

Note
When you use the Digital Fortran 90 compiler, certain features associated with the ANSI/ISO Fortran 90 standard are not available in Digital Fortran 77.

For an explanation of Digital Fortran language features, see the Fortran language reference manual.

11.4.1.2 Language Features Specific to Digital Fortran 77 for OpenVMS VAX Systems

The following language features are available in Digital Fortran 77 for OpenVMS VAX Systems but are not supported in Digital Fortran for OpenVMS Alpha:

Automatic decomposition features of FORTRAN/PARALLEL=(AUTOMATIC)
Manual (directed) decomposition features of FORTRAN/PARALLEL=(MANUAL) using the CPAR$ directives, such as CPAR$ DO_PARALLEL
The following I/O and error subroutines for PDP-11 compatibility:

ASSIGN
CLOSE
ERRSET ERRTST
FDBSET
IRAD50 RAD50
R50ASC
USEREX

When porting existing programs, calls to ASSIGN, CLOSE, and FBDSET should be replaced with the appropriate OPEN statement. (You might consider converting DEFINE FILE statements at the same time, even though Digital Fortran for OpenVMS Alpha does support the DEFINE FILE statement.)
In place of ERRSET and ERRTST, OpenVMS condition handling might be used. Note that Digital Fortran for OpenVMS Alpha supports the ERRSNS subroutine.
Radix-50 constants in the form nRxxx
For existing programs being ported, radix-50 constants and the IRAD50, RAD50, and R50ASC routines should be replaced by data encoded in ASCII using CHARACTER declared data.

Certain Digital Fortran 77 for OpenVMS VAX Systems features have restricted use or are not available in Digital Fortran for OpenVMS Alpha:

Numeric local variables are sometimes, but not always, initialized to a zero value, depending on the level of optimization used. To guarantee that a value will be initialized to zero under all circumstances, use an explicit assignment or DATA statement.
Character constants must be associated with character dummy arguments, not numeric dummy arguments. (Digital Fortran 77 for OpenVMS VAX Systems passed 'A' by reference if the dummy argument was numeric.) Consider using the /BY_REF_CALL qualifier for such arguments.

Saved dummy arrays do not work:

SUBROUTINE F_INIT (A, N) 
REAL A(N) 
RETURN 
ENTRY F_DO_IT (X, I) 
A (I) = X  ! No: A no longer visible 
RETURN 
END

Hollerith actual arguments must be associated with numeric dummy (formal) arguments, not character dummy arguments.

The following language features are available in Digital Fortran 77 for OpenVMS VAX Systems but are not supported in Digital Fortran for OpenVMS Alpha because of differences between the Alpha architecture and the VAX architecture:

Certain FORSYSDEF symbol definition modules may be specific to the VAX or Alpha architecture.
Precise exception-handling control
The handling of certain exceptions differs between OpenVMS VAX and OpenVMS Alpha systems. To request precise exception-handling control, use the /SYNCHRONOUS_EXCEPTIONS qualifier.
REAL*16 data uses the H_floating data format on VAX systems and X_floating on Alpha systems.
VAX support for D_floating
Because the Alpha instruction set does not support the D_floating REAL*8 format, D_floating data is converted to G_floating by software during computations and then converted back to D_floating format. Thus, there will be differences in D_floating arithmetic between VAX and Alpha systems.
For optimal performance on Alpha systems, consider using REAL*8 data in VAX G_floating or IEEE T_floating format, perhaps using the /FLOAT qualifier to specify the format. To create a Digital Fortran for OpenVMS Alpha application program to convert D_floating data to G_floating or T_floating format, use the file conversion methods described in the Fortran language reference manual.
Vectorization capabilities
Vectorization, including /VECTOR and its related qualifiers, and the CDEC$ INIT_DEP_FWD directive are not supported. The Alpha processor provides pipelining and other features that resemble vectorization capabilities.

11.4.1.3 Interpretation Differences

The following language features are interpreted differently between Digital Fortran 77 for OpenVMS VAX Systems and Digital Fortran for OpenVMS Alpha:

Random number generator (RAN)
The RAN function generates a different pattern of numbers in Digital Fortran for OpenVMS Alpha than in Digital Fortran 77 for OpenVMS VAX Systems for the same random seed. (The RAN and RANDU functions are provided for Digital Fortran 77 for OpenVMS VAX Systems compatibility.)
Hollerith constants in formatted I/O statements
Digital Fortran 77 for OpenVMS VAX Systems and Digital Fortran for OpenVMS Alpha behave differently if either of the following occurs:
- Two different I/O statements refer to the same CHARACTER PARAMETER constant as their format specifier. For example:
```
      CHARACTER*(*) FMT2 
      PARAMETER (FMT2='(10Habcdefghij)') 
      READ (5, FMT2) 
      WRITE (6, FMT2) 
```
- Two different I/O statements use the identical character constant as their format specifier. For example:
```
      READ (5, '(10Habcdefghij)') 
      WRITE (6, '(10Habcdefghij)') 
```
In Digital Fortran 77 for OpenVMS VAX Systems, the value obtained by the READ statement is the output of the WRITE statement (FMT2 is ignored). However, in Digital Fortran for OpenVMS Alpha, the output of the WRITE statement is "abcdefghij." (The value read by the READ statement has no effect on the value written by the WRITE statement.)

11.4.2 Command Line Qualifiers

While Digital Fortran for OpenVMS Alpha and Digital Fortran 77 for OpenVMS VAX Systems share most qualifiers, some qualifiers are specific to each platform. This section summarizes the differences between Digital Fortran for OpenVMS Alpha and Digital Fortran 77 for OpenVMS VAX Systems command line qualifiers.

For complete details about the Digital Fortran for OpenVMS Alpha compilation command and options, see the DEC Fortran User Manual for OpenVMS AXP Systems. For complete details about the Digital Fortran 77 for OpenVMS VAX Systems compilation command and options, see the DEC Fortran User Manual for OpenVMS VAX Systems.

To initiate compilation on either VAX or Alpha systems, use the FORTRAN command. On Alpha systems, use the F90 command to initiate compilation using the Digital Fortran 90 compiler.

11.4.2.1 Qualifiers Specific to Digital Fortran for OpenVMS Alpha

Table 11-7 lists Digital Fortran for OpenVMS Alpha compiler qualifiers that have no equivalent Digital Fortran 77 for OpenVMS VAX Systems options and are not supported in Digital Fortran 77 for OpenVMS VAX Systems Version 6.4.

Table 11-7 Digital Fortran for OpenVMS Alpha Qualifiers Not in Digital Fortran 77 for OpenVMS VAX Systems
Qualifier Description

/BY_REF_CALL Allows character constant actual arguments to be associated with numeric dummy arguments (allowed by DEC Fortran for OpenVMS VAX Systems).

/CHECK=FP_EXCEPTIONS Controls whether messages about IEEE floating-point exceptional values are reported at run time.

/DOUBLE_SIZE Makes DOUBLE PRECISION declarations REAL*16 instead of REAL*8.

/FAST Sets several qualifiers that improve run-time performance.

/FLOAT Controls the format used for floating-point data (REAL or COMPLEX) in memory, including the selection of either VAX F_floating or IEEE S_floating for KIND=4 data and VAX G_floating, VAX D_floating, or IEEE T_floating for KIND=8 data. Digital Fortran 77 for OpenVMS VAX Systems provides the /[NO]G_FLOATING qualifier.

/GRANULARITY Controls the granularity of data access for shared data.

/IEEE_MODE Controls how floating-point exceptions are handled for IEEE data.

/INTEGER_SIZE Controls the size of INTEGER and LOGICAL declarations.

/NAMES Controls whether external names are converted to uppercase, lowercase, or as is.

/OPTIMIZE The /OPTIMIZE qualifier supports the INLINE keyword, the LOOPS keyword, the TUNE keyword, the UNROLL keyword, and software pipelining.

/REAL_SIZE Controls the size of REAL and COMPLEX declarations.

/ROUNDING_MODE Controls how floating-point calculations are rounded for IEEE data.

/SEPARATE_COMPILATION Controls whether the DEC Fortran compiler:

Places individual compilation units as separate modules in the object file like Digital Fortran 77 for OpenVMS VAX Systems (/SEPARATE_COMPILATION)
Groups compilation units as a single module in the object file (/NOSEPARATE_COMPILATION, the default), which allows more interprocedure optimizations.

/SYNTAX_ONLY Requests that only syntax checking occurs and no object file is created.

/WARNINGS Certain keywords are not available on Digital Fortran 77 for OpenVMS VAX Systems.

/VMS Requests that Digital Fortran use certain Digital Fortran 77 for OpenVMS VAX Systems conventions.

**Table 11-7 Digital Fortran for OpenVMS Alpha Qualifiers Not in Digital Fortran 77 for OpenVMS VAX Systems**
Qualifier	Description
/BY_REF_CALL	Allows character constant actual arguments to be associated with numeric dummy arguments (allowed by DEC Fortran for OpenVMS VAX Systems).
/CHECK=FP_EXCEPTIONS	Controls whether messages about IEEE floating-point exceptional values are reported at run time.
/DOUBLE_SIZE	Makes DOUBLE PRECISION declarations REAL16 instead of REAL8.
/FAST	Sets several qualifiers that improve run-time performance.
/FLOAT	Controls the format used for floating-point data (REAL or COMPLEX) in memory, including the selection of either VAX F_floating or IEEE S_floating for KIND=4 data and VAX G_floating, VAX D_floating, or IEEE T_floating for KIND=8 data. Digital Fortran 77 for OpenVMS VAX Systems provides the /[NO]G_FLOATING qualifier.
/GRANULARITY	Controls the granularity of data access for shared data.
/IEEE_MODE	Controls how floating-point exceptions are handled for IEEE data.
/INTEGER_SIZE	Controls the size of INTEGER and LOGICAL declarations.
/NAMES	Controls whether external names are converted to uppercase, lowercase, or as is.
/OPTIMIZE	The /OPTIMIZE qualifier supports the INLINE keyword, the LOOPS keyword, the TUNE keyword, the UNROLL keyword, and software pipelining.
/REAL_SIZE	Controls the size of REAL and COMPLEX declarations.
/ROUNDING_MODE	Controls how floating-point calculations are rounded for IEEE data.
/SEPARATE_COMPILATION	Controls whether the DEC Fortran compiler: Places individual compilation units as separate modules in the object file like Digital Fortran 77 for OpenVMS VAX Systems (/SEPARATE_COMPILATION) Groups compilation units as a single module in the object file (/NOSEPARATE_COMPILATION, the default), which allows more interprocedure optimizations.
/SYNTAX_ONLY	Requests that only syntax checking occurs and no object file is created.
/WARNINGS	Certain keywords are not available on Digital Fortran 77 for OpenVMS VAX Systems.
/VMS	Requests that Digital Fortran use certain Digital Fortran 77 for OpenVMS VAX Systems conventions.

11.4.2.2 Qualifiers Specific to Digital Fortran 77 for OpenVMS VAX Systems

This section summarizes Digital Fortran 77 for OpenVMS VAX Systems compiler qualifiers that have no equivalent Digital Fortran for OpenVMS Alpha qualifiers.

Table 11-8 lists compilation qualifiers specific to Digital Fortran 77 for OpenVMS VAX Systems Version 6.4.

Table 11-8 Digital Fortran 77 for OpenVMS VAX Systems Qualifiers Not in Digital Fortran for OpenVMS Alpha
Digital Fortran for 77 OpenVMS VAX Systems Qualifier Description

/BLAS=(INLINE,MAPPED) Specifies whether Digital Fortran 77 for OpenVMS VAX Systems recognizes and inlines or maps the Basic Linear Algebra Subroutines (BLAS). Available only in Digital Fortran 77 for OpenVMS VAX Systems.

/CHECK=ASSERTIONS Enables or disables assertion checking. Available only in Digital Fortran 77 for OpenVMS VAX Systems.

/DESIGN=[NO]COMMENTS
/DESIGN=[NO]PLACEHOLDERS Analyzes program for design information.

/DIRECTIVES=DEPENDENCE Specifies whether specified compiler directives are used at compilation. Available only in Digital Fortran 77 for OpenVMS VAX Systems.

/PARALLEL=(MANUAL or AUTOMATIC) Supports parallel processing.

/SHOW=(DATA_DEPENDENCIES,DICTIONARY,LOOPS) Control whether the listing file includes:

Diagnostics about loops that are ineligible for dependence analysis and data dependencies that inhibit vectorization or autodecomposition (DATA_DEPENDENCIES)
Source lines from included Common Data Dictionary records (DICTIONARY)
Reports about loop structures after compilation (LOOPS)

The keywords DATA_DEPENDENCIES and LOOPS are available only in Digital Fortran 77 for OpenVMS VAX Systems.

/VECTOR Requests vector processing. Available only in Digital Fortran 77 for OpenVMS VAX Systems.

/WARNINGS=INLINE Controls whether the compiler prints informational diagnostic messages when it is unable to generate inline code for a reference to an intrinsic routine. Available only in Digital Fortran 77 for OpenVMS VAX Systems.

**Table 11-8 Digital Fortran 77 for OpenVMS VAX Systems Qualifiers Not in Digital Fortran for OpenVMS Alpha**
Digital Fortran for 77 OpenVMS VAX Systems Qualifier	Description
/BLAS=(INLINE,MAPPED)	Specifies whether Digital Fortran 77 for OpenVMS VAX Systems recognizes and inlines or maps the Basic Linear Algebra Subroutines (BLAS). Available only in Digital Fortran 77 for OpenVMS VAX Systems.
/CHECK=ASSERTIONS	Enables or disables assertion checking. Available only in Digital Fortran 77 for OpenVMS VAX Systems.
/DESIGN=[NO]COMMENTS /DESIGN=[NO]PLACEHOLDERS	Analyzes program for design information.
/DIRECTIVES=DEPENDENCE	Specifies whether specified compiler directives are used at compilation. Available only in Digital Fortran 77 for OpenVMS VAX Systems.
/PARALLEL=(MANUAL or AUTOMATIC)	Supports parallel processing.
/SHOW=(DATA_DEPENDENCIES,DICTIONARY,LOOPS)	Control whether the listing file includes: Diagnostics about loops that are ineligible for dependence analysis and data dependencies that inhibit vectorization or autodecomposition (DATA_DEPENDENCIES) Source lines from included Common Data Dictionary records (DICTIONARY) Reports about loop structures after compilation (LOOPS) The keywords DATA_DEPENDENCIES and LOOPS are available only in Digital Fortran 77 for OpenVMS VAX Systems.
/VECTOR	Requests vector processing. Available only in Digital Fortran 77 for OpenVMS VAX Systems.
/WARNINGS=INLINE	Controls whether the compiler prints informational diagnostic messages when it is unable to generate inline code for a reference to an intrinsic routine. Available only in Digital Fortran 77 for OpenVMS VAX Systems.

All CPAR$ directives and certain CDEC$ directives associated with directed (manual) decomposition and their associated qualifiers or keywords are specific to Digital Fortran 77 for OpenVMS VAX Systems, as described in the DEC Fortran Language Reference ManualDEC Fortran Language Reference Manual.

For details about the Digital Fortran 77 for OpenVMS VAX Systems compilation commands and options, see the DEC Fortran User Manual for OpenVMS VAX Systems.

11.4.3 Interoperability with Translated Shared Images

Using Digital Fortran for OpenVMS Alpha, you can create images that can interoperate with translated images at image activation (run time).

To allow the use of translated shared images:

On the FORTRAN or F90 command line, specify the /TIE qualifier.
On the LINK command line, specify the /NONATIVE_ONLY qualifier.

The created executable image contains code that allows the resulting executable image to interoperate with shared images, including allowing the Digital Fortran 77 for OpenVMS VAX Systems RTL (FORRTL) to work with the Digital Fortran for OpenVMS Alpha RTL (DEC$FORTRTL). The native (Digital Fortran for OpenVMS Alpha RTL) and translated (Digial Fortran 77 for OpenVMS VAX Systems RTL) programs can perform I/O to the same unit number, as long as the RTL that opens the file also closes it.

Programs should use the intrinsic names (without the prefix) rather than calling routines by their complete (fac$xxxx) name. One allowable exception to using fac$xxxx names is that translated image programs declare the FOR$RAB system function as EXTERNAL. Native Alpha programs should use FOR$RAB as an intrinsic function.

11.4.4 Porting Digital Fortran 77 for OpenVMS VAX Systems Data

Record types are identical for Digial Fortran 77 for OpenVMS VAX Systems and Digital Fortran for OpenVMS Alpha. If needed, transport the data using the EXCHANGE command with the /NETWORK and /TRANSFER=BLOCK qualifiers. To convert the file to Stream_LF format during the copy operation, use /TRANSFER=(BLOCK,RECORD_SEPARATOR=LF) instead of /TRANSFER=BLOCK, or specify the /FDL qualifier to the EXCHANGE command to change the record type or other file characteristics.

If you need to convert unformatted floating-point data, keep in mind that Digital Fortran 77 for OpenVMS VAX programs (VAX hardware) store REAL*4 or COMPLEX*8 data in F_floating format, REAL*8, REAL*16, or COMPLEX*16 data in either D_floating or G_floating format, and REAL*16 data in H_floating format. Digital Fortran for OpenVMS Alpha programs (running on Alpha hardware) store REAL*4, REAL*8, REAL*16, COMPLEX*8, and COMPLEX*16 data in one of the formats shown in Table 11-9.

Table 11-9 Floating-Point Data on VAX and Alpha Systems
Data Declaration VAX Formats Alpha Formats

REAL*4 and COMPLEX*8 VAX F_floating format IEEE S_floating or VAX F_floating format

REAL*8 and COMPLEX*16 VAX D_floating or G_floating format IEEE T_floating, VAX D_floating¹, or VAX G_floating format

REAL*16 VAX H_floating X_floating. Requires conversion, perhaps using the /CONVERT qualifier or associated OPTION statement, logical name, or OPEN statement /CONVERT keyword. You can also use the RTL routine CVT$CONVERT_FLOAT.

**Table 11-9 Floating-Point Data on VAX and Alpha Systems**
Data Declaration	VAX Formats	Alpha Formats
REAL4 and COMPLEX8	VAX F_floating format	IEEE S_floating or VAX F_floating format
REAL8 and COMPLEX16	VAX D_floating or G_floating format	IEEE T_floating, VAX D_floating¹, or VAX G_floating format
REAL*16	VAX H_floating	X_floating. Requires conversion, perhaps using the /CONVERT qualifier or associated OPTION statement, logical name, or OPEN statement /CONVERT keyword. You can also use the RTL routine CVT$CONVERT_FLOAT.

¹On Alpha systems, the use of VAX D_floating format involving many computations is not recommended. Consider converting D_floating format to IEEE T_floating (or VAX G_floating) format in a conversion program that uses the Digital Fortran for OpenVMS Alpha conversion routines.

11.5 Compatibility of DEC Pascal for OpenVMS Alpha Systems with VAX Pascal

This section compares DEC Pascal to other Digital Pascal compilers and lists the differences between DEC Pascal on VAX and Alpha systems. For a complete description of these features, see the DEC Pascal Language Reference Manual.

11.5.1 New Features of DEC Pascal

Table 11-10 lists features not previously supplied in VAX Pascal.

Table 11-10 New Features of DEC Pascal
Feature Description

Support for OpenVMS systems Including all the data types available on the OpenVMS platforms.

Redefinable values for predeclared constants Values for MAXINT, MAXUNSIGNED, MAXREAL, MINREAL, EPSREAL are defined by the platform and the compiler switches for specifying the integer size and floating-point format.

An optional quoted parameter to the COMMON, EXTERNAL, GLOBAL, PSECT, WEAK_EXTERNAL, and WEAK_GLOBAL attributes Allows you to pass an unmodified identifier to the linker.

Double-quoted strings DEC Pascal now accepts the double-quote characters as string and character delimiters.

Embedded string values Inside of double-quoted strings, DEC Pascal now supports constant characters specified with a backslash as in the C programming language, such as ""\n"" for the linefeed character.

Additional data types and values DEC Pascal now supports these data types: ALFA, CARDINAL, CARDINAL16, CARDINAL32, INTEGER16, INTEGER32, INTEGER64, INTSET, POINTER, UNIV_PTR, UNSIGNED16, UNSIGNED32, and UNSIGNED64.

Assignment of UNSIGNED values to INTEGER variables DEC Pascal now allows UNSIGNED values to be assignment-compatible with INTEGER variables and array indices.

Assignment of string values into unpacked arrays of characters DEC Pascal now allows ARRAY of CHAR variables to be treated as fixed-length character strings.

Additional statements DEC Pascal now supports these statements: BREAK, CONTINUE, EXIT, NEXT, and RETURN.

Additional predeclared routines DEC Pascal now supports these functions and procedures: ADDR, ARGC, ARGV, ASSERT, BITAND, BITNOT, BITOR, BITXOR, HBOUND, LBOUND, FIRST, FIRSTOF, LAST, LASTOF, IN_RANGE, LSHIFT, RSHIFT, LSHFT, RSHFT, MESSAGE, NULL, RANDOM, SEED, REMOVE, SIZEOF, SYSCLOCK, and WALLCLOCK.

Optional second parameter to RESET, REWRITE, and EXTEND DEC Pascal now accepts a second parameter that is a literal string expression for the file name to be associated with the file variable.

Compiler command switches DEC Pascal now includes switches that allow you to specify the storage and alignment allocation for data types. You can also specify the level of optimization with a switch. On Alpha systems, an option controls the default meaning of the REAL and DOUBLE data types. Arguments to the usage switch enable messages relating to alignment, alignment compatibility on different platforms, and features that are not available on a specified platform.

**Table 11-10 New Features of DEC Pascal**
Feature	Description
Support for OpenVMS systems	Including all the data types available on the OpenVMS platforms.
Redefinable values for predeclared constants	Values for MAXINT, MAXUNSIGNED, MAXREAL, MINREAL, EPSREAL are defined by the platform and the compiler switches for specifying the integer size and floating-point format.
An optional quoted parameter to the COMMON, EXTERNAL, GLOBAL, PSECT, WEAK_EXTERNAL, and WEAK_GLOBAL attributes	Allows you to pass an unmodified identifier to the linker.
Double-quoted strings	DEC Pascal now accepts the double-quote characters as string and character delimiters.
Embedded string values	Inside of double-quoted strings, DEC Pascal now supports constant characters specified with a backslash as in the C programming language, such as ""\n"" for the linefeed character.
Additional data types and values	DEC Pascal now supports these data types: ALFA, CARDINAL, CARDINAL16, CARDINAL32, INTEGER16, INTEGER32, INTEGER64, INTSET, POINTER, UNIV_PTR, UNSIGNED16, UNSIGNED32, and UNSIGNED64.
Assignment of UNSIGNED values to INTEGER variables	DEC Pascal now allows UNSIGNED values to be assignment-compatible with INTEGER variables and array indices.
Assignment of string values into unpacked arrays of characters	DEC Pascal now allows ARRAY of CHAR variables to be treated as fixed-length character strings.
Additional statements	DEC Pascal now supports these statements: BREAK, CONTINUE, EXIT, NEXT, and RETURN.
Additional predeclared routines	DEC Pascal now supports these functions and procedures: ADDR, ARGC, ARGV, ASSERT, BITAND, BITNOT, BITOR, BITXOR, HBOUND, LBOUND, FIRST, FIRSTOF, LAST, LASTOF, IN_RANGE, LSHIFT, RSHIFT, LSHFT, RSHFT, MESSAGE, NULL, RANDOM, SEED, REMOVE, SIZEOF, SYSCLOCK, and WALLCLOCK.
Optional second parameter to RESET, REWRITE, and EXTEND	DEC Pascal now accepts a second parameter that is a literal string expression for the file name to be associated with the file variable.
Compiler command switches	DEC Pascal now includes switches that allow you to specify the storage and alignment allocation for data types. You can also specify the level of optimization with a switch. On Alpha systems, an option controls the default meaning of the REAL and DOUBLE data types. Arguments to the usage switch enable messages relating to alignment, alignment compatibility on different platforms, and features that are not available on a specified platform.

11.5.2 Establishing Dynamic Condition Handlers

DEC Pascal provides the built-in routines, ESTABLISH and REVERT, to use in place of LIB$ESTABLISH. If you declare and try to use LIB$ESTABLISH, you will get a compile-time warning.

11.5.3 Modifying Default Alignment Rules for Record Fields

DEC Pascal allows you to override field alignment and position with the POS, ALIGNED, and DATA attributes and the data compiler switch.

11.5.4 Recommended Use of Predeclared Identifiers

Although for backward compatibility DEC Pascal compiles programs that include the predeclared identifiers listed in Table 11-11, Digital recommends that you use the listed replacements.

Table 11-11 Recommended Use of Predeclared Identifiers
Identifier Recommended Usage

ADDR Use the ADDRESS function

ALFA Equivalent to TYPE ALFA = PACKED ARRAY [1..10]OF CHAR

BITAND Equivalent to the UAND statement

BITNOT Equivalent to the UNOT statement

BITOR Equivalent to the UOR statement

BITXOR Equivalent to the UXOR statement

EXIT Equivalent to the BREAK statement

FIRST, FIRSTOF Equivalent to the LOWER function

HBOUND Equivalent to the UPPER function

IN_RANGE Useful only when subrange checking is disabled. IN_RANGE(X) is equivalent to (X=>LOWER(X))AND(X <=UPPER(X)).

INTSET Equivalent to TYPE INTSET = SET OF 0 .. 255;

LAST, LASTOF Equivalent to the UPPER function

LBOUND Equivalent to the LOWER function

LSHFT Equivalent to the LSHIFT function

MESSAGE Equivalent to WRITELN(ERR,expression)

NEXT Equivalent to the CONTINUE statement

NULL Equivalent to the empty statement

REMOVE Equivalent to the DELETE_FILE procedure

RSHFT Equivalent to the RSHIFT function

SIZEOF Equivalent to the SIZE function

STLIMIT Compiles but does not return an error

UNIV_PTR Equivalent to TYPE UNIV_PTR = POINTER;

**Table 11-11 Recommended Use of Predeclared Identifiers**
Identifier	Recommended Usage
ADDR	Use the ADDRESS function
ALFA	Equivalent to TYPE ALFA = PACKED ARRAY [1..10]OF CHAR
BITAND	Equivalent to the UAND statement
BITNOT	Equivalent to the UNOT statement
BITOR	Equivalent to the UOR statement
BITXOR	Equivalent to the UXOR statement
EXIT	Equivalent to the BREAK statement
FIRST, FIRSTOF	Equivalent to the LOWER function
HBOUND	Equivalent to the UPPER function
IN_RANGE	Useful only when subrange checking is disabled. IN_RANGE(X) is equivalent to (X=>LOWER(X))AND(X <=UPPER(X)).
INTSET	Equivalent to TYPE INTSET = SET OF 0 .. 255;
LAST, LASTOF	Equivalent to the UPPER function
LBOUND	Equivalent to the LOWER function
LSHFT	Equivalent to the LSHIFT function
MESSAGE	Equivalent to WRITELN(ERR,expression)
NEXT	Equivalent to the CONTINUE statement
NULL	Equivalent to the empty statement
REMOVE	Equivalent to the DELETE_FILE procedure
RSHFT	Equivalent to the RSHIFT function
SIZEOF	Equivalent to the SIZE function
STLIMIT	Compiles but does not return an error
UNIV_PTR	Equivalent to TYPE UNIV_PTR = POINTER;

11.5.5 Platform-Dependent Features

DEC Pascal can use an environment file only on the same platform (the combination of operating system and hardware) on which it was compiled.

In addition, the following lists features of DEC Pascal supplied only on VAX systems:

QUADRUPLE data type
H_floating-point data type
VAX Pascal Version 1.0 dynamic arrays
MFPR and MTPR predeclared routines
[OVERLAID] attribute
Table of contents in listing
Optimize attribute on routines

The following lists features of DEC Pascal supplied only on Alpha systems:

Abbreviations when reading enumerated data types
Indexed file organization
Relative file organization

11.5.6 Obsolete Features

This section describes features that are supported, but not recommended, by Digital. They are provided only for compatibility with other Digital Pascal compilers.

11.5.6.1 /OLD_VERSION Qualifier

The /OLD_VERSION qualifier directed the compiler to resolve differences between VAX Pascal Version 1.0 and subsequent versions by using the VAX Pascal Version 1.0 definition of the language. The qualifier is provided so that existing programs continue to work.

11.5.6.2 /G_FLOATING Qualifier

The /G_FLOATING qualifier directs the compiler to use the G_floating representation and instructions for values of type DOUBLE. The [[NO]G_FLOATING] attribute can be specified on both OpenVMS VAX and OpenVMS Alpha systems.

  6459P014.HTM
  OSSG Documentation
  22-NOV-1996 13:07:26.23

Legal