Contents
List of Tables
About this document
1 Compiler directives
1.1 Introduction
1.2 Local directives
1.2.1 $A or $ALIGN : Align Data
1.2.2 $A1, $A2,$A4 and $A8
1.2.3 $ASMMODE : Assembler mode (Intel 80x86 only)
1.2.4 $B or $BOOLEVAL : Complete boolean evaluation
1.2.5 $C or $ASSERTIONS : Assertion support
1.2.6 $BITPACKING : Enable bitsize packing
1.2.7 $CALLING : Specify calling convention
1.2.8 $CHECKPOINTER : Check pointer values
1.2.9 $CODEALIGN : Set the code alignment
1.2.10 $COPERATORS : Allow C like operators
1.2.11 $DEFINE or $DEFINEC : Define a symbol
1.2.12 $ELSE : Switch conditional compilation
1.2.13 $ELSEC : Switch conditional compilation
1.2.14 $ELSEIF or $ELIFC : Switch conditional compilation
1.2.15 $ENDC : End conditional compilation
1.2.16 $ENDIF : End conditional compilation
1.2.17 $ERROR or $ERRORC : Generate error message
1.2.18 $ENDREGION: End of a collapsible region
1.2.19 $EXCESSPRECISION: use excess precision with float
1.2.20 $EXTENDEDSYM: Ignored
1.2.21 $EXTERNALSYM: Ignored
1.2.22 $F : Far or near functions
1.2.23 $FATAL : Generate fatal error message
1.2.24 $FPUTYPE : Select coprocessor type
1.2.25 $GOTO : Support Goto and Label
1.2.26 $H or $LONGSTRINGS : Use AnsiStrings
1.2.27 $HINT : Generate hint message
1.2.28 $HINTS : Emit hints
1.2.29 $HPPEMIT: Ignored
1.2.30 $IF : Start conditional compilation
1.2.31 $IFC : Start conditional compilation
1.2.32 $IFDEF Name : Start conditional compilation
1.2.33 $IFEND : End conditional compilation
1.2.34 $IFNDEF : Start conditional compilation
1.2.35 $IFOPT : Start conditional compilation
1.2.36 $IMPLICITEXCEPTIONS : Implicit finalization code generation
1.2.37 $INFO : Generate info message
1.2.38 $INLINE : Allow inline code.
1.2.39 $INTERFACES : Specify Interface type.
1.2.40 $I or $IOCHECKS : Input/Output checking
1.2.41 $IEEEERRORS : Enable IEEE error checking for constants
1.2.42 $I or $INCLUDE : Include file
1.2.43 $I or $INCLUDE : Include compiler info
1.2.44 $J or $WRITEABLECONST : Allow assignments to typed consts
1.2.45 $L or $LINK : Link object file
1.2.46 $LIBEXPORT : Ignored
1.2.47 $LINKFRAMEWORK : Link to a framework
1.2.48 $LINKLIB : Link to a library
1.2.49 $M or $TYPEINFO : Generate type info
1.2.50 $MACRO : Allow use of macros.
1.2.51 $MAXFPUREGISTERS : Maximum number of FPU registers for variables
1.2.52 $MESSAGE : Generate user-defined message
1.2.53 $MINENUMSIZE : Specify minimum enumeration size
1.2.54 $MINFPCONSTPREC : Specify floating point constant precision
1.2.55 $MMX : Intel MMX support (Intel 80x86 only)
1.2.56 $NODEFINE : Ignored
1.2.57 $NOTE : Generate note message
1.2.58 $NOTES : Emit notes
1.2.59 $OBJECTCHECKS : Check Object
1.2.60 $OPTIMIZATION : Enable Optimizations
1.2.61 $PACKENUM or $Z : Minimum enumeration type size
1.2.62 $PACKRECORDS : Alignment of record elements
1.2.63 $PACKSET : Specify set size
1.2.64 $POP : Restore compiler settings
1.2.65 $PUSH : Save compiler settings
1.2.66 $Q or $OV or $OVERFLOWCHECKS: Overflow checking
1.2.67 $R or $RANGECHECKS : Range checking
1.2.68 $REGION : Mark start of collapsible region
1.2.69 $R or $RESOURCE : Include resource
1.2.70 $SATURATION : Saturation operations (Intel 80x86 only)
1.2.71 $SAFEFPUEXCEPTIONS Wait when storing FPU values on Intel x86
1.2.72 $SCOPEDENUMS Control use of scoped enumeration types
1.2.73 $SETC : Define and assign a value to a symbol
1.2.74 $S : Stack checking
1.2.75 $STOP : Generate fatal error message
1.2.76 $STRINGCHECKS : Ignored
1.2.77 $T or $TYPEDADDRESS : Typed address operator (@)
1.2.78 $UNDEF or $UNDEFC : Undefine a symbol
1.2.79 $V or $VARSTRINGCHECKS : Var-string checking
1.2.80 $W or $STACKFRAMES : Generate stackframes
1.2.81 $WAIT : Wait for enter key press
1.2.82 $WARN : Control emission of warnings
1.2.83 $WARNING : Generate warning message
1.2.84 $WARNINGS : Emit warnings
1.2.85 $Z1, $Z2 and $Z4
1.3 Global directives
1.3.1 $APPID : Specify application ID.
1.3.2 $APPNAME : Specify application name.
1.3.3 $APPTYPE : Specify type of application.
1.3.4 $CODEPAGE : Set the source codepage
1.3.5 $COPYRIGHT specify copyright info
1.3.6 $D or $DEBUGINFO : Debugging symbols
1.3.7 $DESCRIPTION : Application description
1.3.8 $E : Emulation of coprocessor
1.3.9 $EXTENSION : Extension of generated binary.
1.3.10 $FRAMEWORKPATH : Specify framework path.
1.3.11 $G : Generate 80286 code
1.3.12 $IMAGEBASE : Specify DLL image base location.
1.3.13 $INCLUDEPATH : Specify include path.
1.3.14 $L or $LOCALSYMBOLS : Local symbol information
1.3.15 $LIBPREFIX : Specify library filename prefix.
1.3.16 $LIBRARYPATH : Specify library path.
1.3.17 $LIBSUFFIX : Set the library suffix
1.3.18 $MAXSTACKSIZE : Set maximum stack size
1.3.19 $M or $MEMORY : Memory sizes
1.3.20 $MINSTACKSIZE : Set minimum stack size
1.3.21 $MODE : Set compiler compatibility mode
1.3.22 $MODESWITCH : Select mode features
1.3.23 $N : Numeric processing
1.3.24 $O : Level 2 Optimizations
1.3.25 $OBJECTPATH : Specify object path.
1.3.26 $P or $OPENSTRINGS : Use open strings
1.3.27 $PASCALMAINNAME : Set entry point name
1.3.28 $PIC : Generate PIC code or not
1.3.29 $POINTERMATH : Allow use of pointer math
1.3.30 $PROFILE : Profiling
1.3.31 $SCREENNAME : Specify screen name
1.3.32 $SETPEFLAGS : Specify PE Executable flags
1.3.33 $SMARTLINK : Use smartlinking
1.3.34 $SYSCALLS : Select system calling convention on Amiga/MorphOS
1.3.35 $THREADNAME : Set thread name in Netware
1.3.36 $UNITPATH : Specify unit path.
1.3.37 $VARPROPSETTER : Enable use of var/out/constref parameters for property setters.
1.3.38 $VERSION : Specify DLL version.
1.3.39 $WEAKPACKAGEUNIT : ignored
1.3.40 $X or $EXTENDEDSYNTAX : Extended syntax
1.3.41 $Y or $REFERENCEINFO : Insert Browser information
2 Using conditionals, messages and macros
2.1 Conditionals
2.1.1 Predefined symbols
2.2 Macros
2.3 Compile time variables
2.4 Compile time expressions
2.4.1 Definition
2.4.2 Usage
2.5 Messages
3 Using Assembly language
3.1 Using assembler in the sources
3.2 Intel 80x86 Inline assembler
3.2.1 Intel syntax
3.2.2 AT&T Syntax
3.3 Motorola 680x0 Inline assembler
3.4 Signaling changed registers
4 Generated code
4.1 Units
4.2 Programs
5 Intel MMX support
5.1 What is it about?
5.2 Saturation support
5.3 Restrictions of MMX support
5.4 Supported MMX operations
5.5 Optimizing MMX support
6 Code issues
6.1 Register Conventions
6.1.1 accumulator register
6.1.2 accumulator 64-bit register
6.1.3 float result register
6.1.4 self register
6.1.5 frame pointer register
6.1.6 stack pointer register
6.1.7 scratch registers
6.1.8 Processor mapping of registers
6.2 Name mangling
6.2.1 Mangled names for data blocks
6.2.2 Mangled names for code blocks
6.2.3 Modifying the mangled names
6.3 Calling mechanism
6.4 Nested procedure and functions
6.5 Constructor and Destructor calls
6.5.1 objects
6.5.2 classes
6.6 Entry and exit code
6.6.1 Intel 80x86 standard routine prolog / epilog
6.6.2 Motorola 680x0 standard routine prolog / epilog
6.7 Parameter passing
6.7.1 Parameter alignment
6.8 Stack limitations
7 Linking issues
7.1 Using external code and variables
7.1.1 Declaring external functions or procedures
7.1.2 Declaring external variables
7.1.3 Declaring the calling convention modifier
7.1.4 Declaring the external object code
7.2 Making libraries
7.2.1 Exporting functions
7.2.2 Exporting variables
7.2.3 Compiling libraries
7.2.4 Unit searching strategy
7.3 Using smart linking
8 Memory issues
8.1 The memory model.
8.2 Data formats
8.2.1 Integer types
8.2.2 Char types
8.2.3 Boolean types
8.2.4 Enumeration types
8.2.5 Floating point types
8.2.6 Pointer types
8.2.7 String types
8.2.8 Set types
8.2.9 Static array types
8.2.10 Dynamic array types
8.2.11 Record types
8.2.12 Object types
8.2.13 Class types
8.2.14 File types
8.2.15 Procedural types
8.3 Data alignment
8.3.1 Typed constants and variable alignment
8.3.2 Structured types alignment
8.4 The heap
8.4.1 Heap allocation strategy
8.4.2 The heap grows
8.4.3 Debugging the heap
8.4.4 Writing your own memory manager
8.5 Using dos memory under the Go32 extender
8.6 When porting Turbo Pascal code
8.7 Memavail and Maxavail
9 Resource strings
9.1 Introduction
9.2 The resource string file
9.3 Updating the string tables
9.4 GNU gettext
9.5 Caveat
10 Thread programming
10.1 Introduction
10.2 Programming threads
10.3 Critical sections
10.4 The Thread Manager
11 Optimizations
11.1 Non processor specific
11.1.1 Constant folding
11.1.2 Constant merging
11.1.3 Short cut evaluation
11.1.4 Constant set inlining
11.1.5 Small sets
11.1.6 Range checking
11.1.7 And instead of modulo
11.1.8 Shifts instead of multiply or divide
11.1.9 Automatic alignment
11.1.10 Smart linking
11.1.11 Inline routines
11.1.12 Stack frame omission
11.1.13 Register variables
11.2 Processor specific
11.2.1 Intel 80x86 specific
11.2.2 Motorola 680x0 specific
11.3 Optimization switches
11.4 Tips to get faster code
11.5 Tips to get smaller code
11.6 Whole Program Optimization
11.6.1 Overview
11.7 General principles
11.7.1 How to use
11.7.2 Available WPO optimizations
11.7.3 format of the WPO file
12 Programming shared libraries
12.1 Introduction
12.2 Creating a library
12.3 Using a library in a pascal program
12.4 Using a pascal library from a C program
12.5 Some Windows issues
13 Using Windows resources
13.1 The resource directive $R
13.2 Creating resources
13.3 Using string tables.
13.4 Inserting version information
13.5 Inserting an application icon
13.6 Using a Pascal preprocessor
A Anatomy of a unit file
A.1 Basics
A.2 reading ppufiles
A.3 The Header
A.4 The sections
A.5 Creating ppufiles
B Compiler and RTL source tree structure
B.1 The compiler source tree
B.2 The RTL source tree
C Compiler limits
D Compiler modes
D.1 FPC mode
D.2 TP mode
D.3 Delphi mode
D.4 DelphiUnicode mode
D.5 OBJFPC mode
D.6 MACPAS mode
D.7 ISO mode
D.8 Extended Pascal mode
E Using fpcmake
E.1 Introduction
E.2 Functionality
E.3 Usage
E.4 Format of the configuration file
E.4.1 clean
E.4.2 compiler
E.4.3 Default
E.4.4 Dist
E.4.5 Install
E.4.6 Package
E.4.7 Prerules
E.4.8 Requires
E.4.9 Rules
E.4.10 Target
E.5 Programs needed to use the generated makefile
E.6 Variables that affect the generated makefile
E.6.1 Directory variables
E.6.2 Compiler command line variables
E.7 Variables set by fpcmake
E.7.1 Directory variables
E.7.2 Target variables
E.7.3 Compiler command line variables
E.7.4 Program names
E.7.5 File extensions
E.7.6 Target files
E.8 Rules and targets created by fpcmake
E.8.1 Pattern rules
E.8.2 Build rules
E.8.3 Cleaning rules
E.8.4 Archiving rules
E.8.5 Installation rules
E.8.6 Informative rules
F Compiling the compiler
F.1 Introduction
F.2 Before starting
F.3 Compiling using make
F.3.1 Basic compilation
F.3.2 Compiling for another target
F.3.3 Installing the new compiler
F.3.4 Speeding up compilation
F.3.5 Adding debug information
F.3.6 Various other options
F.4 Compiling by hand
F.4.1 Compiling the RTL
F.4.2 Compiling the compiler
G Compiler defines during compilation