CD Actual 1

home *** CD-ROM | disk | FTP | other *** search

/ CD Actual 1 / PC Actual CD 01.iso / share / dos / progr / micro_c.arj / MC.DOC < prev next >

Wrap

Text File | 1993-10-05 | 171.6 KB | 3,602 lines

=========================================================== DDDDD DDDDD SSSS D Dunfield D D S D D D Development SSSS D D D D Systems DDDDD DDDDD SSSS =========================================================== MM MM IIIIIII CCCC RRRRRR OOO CCCC M M M M I C C R R O O C C M M M I C R R O O C M M I C RRRRRR O O ----- C M M I C R R O O C M M I C C R R O O C C M M IIIIIII CCCC R R OOO CCCC =========================================================== A compact 'C' compiler for Small Systems. Technical Manual Release 3.01 Revised 18-Sep-93 Copyright 1988-1993 Dave Dunfield All rights reserved MICRO-C Page: 1 1. INTRODUCTION DDS MICRO-C is a compact, portable compiler for the 'C' language which is suitable for implementation on small 8 or 16 bit computer systems. It may be used as a resident or cross compiler, and is capable of generating ROMable code. It is distributed in source form, and thus provides a learning tool for those wishing to examine the internals of a working compiler. The main goal in the development of MICRO-C, has been to provide a reasonably powerful language which will be portable with little difficulty to many target systems. The 'C' language was chosen because it was designed to be portable, and provides ample programming power. MICRO-C provides an alternative to interpreted BASIC or assembly language programming which are often the only languages available on small 8 bit systems. With today's focus on larger computers and workstations, software support for small systems and micro-controller sized devices is getting both hard to find, and expensive. MICRO-C helps fill that gap, because all files necessary to port and maintain the compiler are included on the distribution disks, allowing it to be moved to ANY system without dependance on a software vendor. Included in this package is a complete IBM/PC implementation of MICRO-C, with a comprehensive library, which you may use as an example installation. This implementation of the compiler requires the Microsoft MASM (or compatible) assembler. The MICRO-C "package" (software and documentation) is copyrighted, and may not be re-distributed in any form without my written permission. If you have received the "DEMO" archive, and wish to obtain the entire MICRO-C package complete with source code, fill out and send the order form contained in the "READ.ME" file, along with the required payment to: Dunfield Development Systems P.O. Box 31044 Nepean, Ontario (Canada) K2B 8S8 Please make your cheque or money order payable to "Dunfield Development Systems". MICRO-C is provided on an "as is" basis, with no warranty of any kind. In no event shall the author be liable for any damages arising from its use or distribution. MICRO-C Page: 2 1.1 Compiler Portability MICRO-C is written in standard 'C', and is capable of compiling itself. This allows any system with a 'C' compiler (including MICRO-C) to be used to port MICRO-C to another processor. The parser makes very few assumptions about the target processor or operating system architecture, allowing a code generator to be written for virtually any processor and environment. With the exception of required I/O routines (described later), the MICRO-C compiler uses no library functions, and relies on no system services. Assuming that the code generator is fairly efficent, and that the I/O routines and code generator are not unreasonably large, a full MICRO-C compiler may be implemented on systems with as little as 32K of free ram and a single floppy disk. 1.2 Code Portability With few exceptions, this compiler follows the syntax of the "standard" UNIX compiler (within its subset of the language). Programs written in MICRO-C should compile with few changes under other "standard" compilers. 1.2.1 Unsupported Features: MICRO-C does not currently support the following features of standard 'C': Long / Double / Float / Enumerated data types, Typedef and Bit fields. 1.2.2 Additional Features MICRO-C provides a few additional features which are not always included in "standard" 'C' compilers: Unsigned character variables, Nested comments, 16 bit character constants, Inline assembly code capability. MICRO-C Page: 3 1.3 The compiling process There are five programs which work together to completely compile a MICRO-C program: The PREPROCESSOR (MCP) takes the original 'C' source file, performs MACRO expansion and incorporates the contents of INCLUDE files to get a "pure" 'C' output file. A less powerful pre-processor is also contained inside the COMPILER, which allows this step to be skipped for programs which use only simple pre-processor functions. The COMPILER (MCC) reads a file containing a 'C' source program, and translates it into an equivalent assembly language program. The OPTIMIZER (MCO) reads the assembly language output from the compiler identifying and replacing "general" code produced by the compiler with more efficent code which is equivalent in specific cases. This step is optional, allowing you to choose between greater compile time and program efficency. The ASSEMBLER takes the assembly language produced by the COMPILER or OPTIMIZER, and produces an OBJECT FILE which contains the executable code and external reference map for the program. The LINKER combines the executable code from the OBJECT FILE with code from other programs which it calls (including any LIBRARY functions), to produce a complete stand-alone executable program. The ASSEMBLER and LINKER are NOT included as part of the MICRO-C package. These programs are specific to the particular system, and must be obtained from the system vendor. Some cross ASSEMBLER's do not produce an object file, but generate the executable program directly instead. These systems assume that the ENTIRE program is completely contained in a SINGLE assembly language file. This type of system prevents you from using a object LIBRARY which is an integral feature of any C programming system. To accomodate this type of development system, without sacrificing the ability to use a LIBRARY, a program called the SOURCE LINKER (SLINK) is included with MICRO-C. This program combines your program with library functions at the ASSEMBLY LANGUAGE level, allowing the entire program code to be presented to the assembler in a single file. See the section entitled "THE SOURCE LINKER" for additional information. MICRO-C Page: 4 2. THE COMMAND CO-ORDINATOR 'CC' is a utility which co-ordinates the execution of programs required for each step of the compilation process to provide a simple "one step" compilation command. This MICRO-C package includes a version of the command co-ordinator which handles the commands necessary to compile for the 8086 processor (MCP, MCC, MCO, MASM, LINK and EXE2BIN). This program can be modified as necessary to handle other processors. 2.1 The CC command The format of the 'CC' command is: CC <name> [options] 2.1.1 Command line options -a - produce ASSEMBLER (.ASM) output file -c - Include 'C' source as comments in ASM files -f - Fold identical string constants. -k - KEEP all temporary files (do not delete) -l - produce LINKABLE (.OBJ) output file -o - OPTIMIZE the output code (using MCO) -p - use the extended PRE-PROCESSOR (MCP) -q - QUIET mode (suppress informational messages) -s - produce SMALL-MODEL (.EXE) output file h=mcdir - specify MICRO-C home directory t=mctmp - specify prefix for TEMPORARY files When executing the sub-commands, CC will search the MICRO-C home directory, as well as any directories specified in the 'PATH' environment variable. Libraries are accessed from the MICRO-C home directory only. The environment variable 'MCDIR' is examined to determine the path to the MICRO-C home directory. If MCDIR is not defined, CC will assume the string '\MC'. You may override this directory by using the 'h=' option on the command line. Intermediate results from each command are stored in "temporary" files, which are fed as input to the next command. Temporary files will be deleted once they are no longer needed, except in the case where a command fails. When this happens, any temporary file which was being used as input to that command will not be deleted, allowing you to examine it for the cause of the error. MICRO-C Page: 5 The environment variable 'MCTMP' is examined to determine a prefix to prepend to temporary file names. If MCTMP is not defined, CC assumes the default prefix of '$'. You may override this prefix by using the 't=' option on the command line. Here are example 'SET' commands suitable for inclusion in the AUTOEXEC.BAT file, of an IBM/PC based MICRO-C system which has the home directory in 'C:\MC', and a TEMP directory on a RAMDISK as drive 'D': SET MCDIR=C:\MC SET MCTMP=D:\TEMP\ (Note trailing '\') The '-a' option causes CC to bypass the assembler and linker, and produce an assembly source file (<name>.ASM) as the output file. If the '-c' option is also used, this file will contain the 'C' source code in the form of comments. NOTE: The source code inserted by '-c' will restrict the optimizer to operate only on sections of code produced by a single 'C' line. The '-f' option causes the compiler to "fold" its literal pool (the area of memory where string constants are stored). This means that identical strings not contained in explicit variables, will occur only once in memory. Since most programs never modify such strings, it is usually safe to do this. Note however that this is a violation of the 'C' standard ("The 'C' Programming Language" page 181 - "All strings, even when written identically, are distinct"), and it is possible to write programs which will not work properly when this option is used. The '-l' option causes CC to bypass the link stage, and produce a linkable object module (<name>.OBJ) as the output file. The '-s' option causes CC to link the program using the SMALL memory model, and produce an executable image (<name>.EXE) as the output file. This option applies only to those processors which have a segmented architecture and multiple memory models such as the 8086. If none of '-a', '-l' or '-s' is specified, CC will link the program using the TINY memory model, and produce a command image (<name>.COM) as the output file. MICRO-C Page: 6 2.2 Using multiple object modules The 'LC' command takes one or more object modules produced by 'CC' with the '-l' option, and links them with the libraries to produce an executable module. This module will be given the name of the first file specified in the argument list. When compiling large programs which have more than one source file, you must first compile all modules using 'CC' with the '-l' option, and the use 'LC' to link the resultant object files into the final executable program. The '-s' option may be used as the FIRST argument, to cause LC to link in the SMALL model. eg: CC FIRST -l CC SECOND -l LC FIRST SECOND <-- TINY Model LC -s FIRST SECOND <-- SMALL Model The exact syntax and options available for 'CC' and 'LC' commands may differ in specific implementations of the compiler. See the 'READ.ME' file and any additional technical documentation on your distribution disk(s) for details. MICRO-C Page: 7 3. THE MICRO-C PROGRAMMING LANGUAGE The following pages contain a brief summary of the features and constructs implemented in MICRO-C. This document does not attempt to teach 'C' programming, and assumes that the reader is familiar with the language. 3.1 Constants The following forms of constants are supported by the compiler: <num> - Decimal number (0 - 65535) 0<num> - Octal number (0 - 0177777) 0x<num> - Hexidecimal number (0x0 - 0xffff) '<char>' - Character (1 or 2 chars) "<string>" - Address of literal string. The following "special" characters may be used within character constants or strings: \n - Newline (line-feed) (0x0a) \r - Carriage Return (0x0d) \t - Tab (0x09) \f - Formfeed (0x0c) \b - Backspace (0x08) \<num> - Octal value <num> (Max. three digits) \x<num> - Hex value <num> (Max. two digits) \<char> - Protect character <char> from input scanner. 3.2 Symbols Symbol names may include the characters 'a'-'z', 'A'-'Z', '0'-'9', and '_'. The characters '0'-'9' may not be used as the first character in the symbol name. Symbol names may be any length, however, only the first 15 characters are significant. The "char" modifier may be used to declare a symbol as an 8 bit wide value, otherwise it is assumed to be 16 bits. eg: char input_char; The "int" modifier may be used to declare a symbol as a 16 bit wide value. This is assumed if neither "int" or "char" is given. eg: int abc; The "unsigned" modifier may be used to declare a symbol as an unsigned positive only value. Note that unlike some 'C' compilers, this modifier may be applied to a character (8 bit) variable. eg: unsigned char count; MICRO-C Page: 8 The "extern" modifier causes the compiler to be aware of the existance and type of a global symbol, but not generate a definition for that symbol. This allows the module being compiled to reference a symbol which is defined in another module. This modifier may not be used with local symbols. eg: extern char getc(); A symbol declared as external may be re-declared as a non-external at a later point in the code, in which case a definition for it will be generated. This allows "extern" to be used to inform the compiler of a function or variable type so that it can be referenced properly before that symbol is actually defined. The "static" modifier causes global symbols to be available only in the file where they are defined. Variables or functions declared as "static" will not be accessable as "extern" declarations in other object files, nor will they cause conflicts with duplicate names in those files. eg: static int variable_name; When applied to local symbols, the "static" modifier causes those variables to be allocated in a reserved area of memory, instead of on the processor stack. This has the effect that the contents of the variable will be retained between calls to the function. It also means that the variable may be initialzed at compile time. The "register" modifier indicates to the code generator that this is a high priority variable, and should be kept where it is easy to get at. Since its interpretation depends on the code generator, it is often ignored in simple implementations. See "Functions" for a special use of "register" when defining a function. eg: register unsigned count; Symbols declared with a preceeding '*' are assumed to be 16 bit pointers to the declared type. eg: int *pointer_name; Symbol names declared followed by square brackets are assumed to be arrays with a number of dimensions equal to the number of '[]' pairs that follow. The size of each dimension is identified by a constant value contained within the corresponding square brackets. eg: char array_name[5][10]; MICRO-C Page: 9 3.2.1 Global Symbols Symbols declared outside of a function definition are considered to be global and will have memory permanently reserved for them. Global symbols are defined by name in the output file, allowing other modules to access them. Note that the compiler IS case sensitive, however if the assembler you are using is NOT, you must be careful not to declare any global symbols with names that differ only in case. All non-initialized global variables are generated at the very end of the output file, after the literal pool is dumped. Since non-initialized globals do not generate object code, this allows them to be excluded from the image file when it is saved to disk. Global variables may be initialized with one or more values, which are expressed as a single array of integers REGUARDLESS of the size and shape of the variable. If more than one value is expressed, '{' and '}' must be used. eg: int i = 10, j[2][2] = { 1, 2, 3, 4 }; When arrays are declared, a null dimension may be used as the dimension size, in which case the size of the array will default to the number of initialized values. eg: int array[] = { 1, 2, 3 }; Initialized global variables are automatically saved within the code image, insuring that the initial values will be available at run time. Any non-initialized elements of an array which has been partly initialized will be set to zero. Non-initialized global variables are not preset in any way, and will be undefined at the beginning of program execution. 3.2.2 Local Symbols Symbols declared within a function definition are allocated on the stack, and exist only during the execution of the function. To simplify the allocation and de-allocation of stack space, all local symbols must be declared at the beginning of the function before any code producing statements are encountered. MICRO-C does not support initialization of non-static local variables in the declaration statement. Since local variables have to be initialized every time the function is entered, you can get the same effect using assignment statements at the beginning of the function. No type is assumed for arguments to functions. Arguments must be explicitly declared, otherwise they will be undefined within the scope of the function definition. MICRO-C Page: 10 3.2.3 More Symbol Examples /* Global variables are defined outside of any function */ char a; /* 8 bit signed */ unsigned char b; /* 8 bit unsigned */ int c; /* 16 bit signed */ unsigned int d; /* 16 bit unsigned */ unsigned e; /* also 16 bit unsigned */ extern char f(); /* external function returning char */ static int g; /* 16 bit signed, local to file */ int h[2] = { 1, 2 }; /* initialized array (2 ints) */ main(a, b) /* "int" function containing code */ /* Function arguments are defined between function name and body */ int a; /* 16 bit signed */ unsigned char *b; /* pointer to 8 bit unsigned */ { /* Local variables are defined inside the function body */ /* Note that in MICRO-C, only "static" locals can be initialized */ unsigned c; /* 16 bit unsigned */ static char d[5]; /* 8 bit signed, reserved in memory */ static int e = 1; /* 16 bit signed, initial value is 1 */ /* Function code goes here ... */ c = 0; /* Initialize 'c' to zero */ strcpy(d, "Name"); /* Initialize 'd' to "name" */ } MICRO-C Page: 11 3.3 Arrays & Pointers When MICRO-C passes an array to a function, it actually passes a POINTER to the array. References to arrays which are arguments are automatically performed through the pointer. This allows the use of pointers and arrays to be interchangable through the context of a function call. Ie: An array passed to a function may be declared and used as a pointer, and a pointer passed to a function may be declared and used as an array. 3.4 Functions Functions are essentially initialized global symbols which contain executable code. MICRO-C accepts any valid value as a function reference, allowing some rather unique (although non-standard) function calls. For example: function(); /* call function */ variable(); /* call contents of a variable */ (*var)(); /* call indirect through variable */ (*var[x])(); /* call indirect through indexed array */ 0x5000(); /* call address 0x5000 */ Since this is a single pass compiler, operands to functions are evaluated and pushed on the stack in the order in which they are encountered, leaving the last operand closest to the top of the stack. This is the opposite order from which many 'C' compilers push operands. For functions with a fixed number of arguments, the order of which operands are passed is of no importance, because the compiler looks after generating the proper stack addresses to reference variables. HOWEVER, functions which use a variable number of arguments are affected for two reasons: 1) The location of the LAST arguments are known (as fixed offsets from the stack pointer) instead of the FIRST. 2) The symbols defined as arguments in the function definition represent the LAST arguments instead of the FIRST. If a function is declared as "register", it serves a special purpose and causes the accumulator to be loaded with the number of arguments passed whenever the function is called. This allows the function to know how many arguments were passed and therefore determine the location of the first argument. MICRO-C Page: 12 3.5 Structures & Unions Combinations of other variable types can be organized into STRUCTURES or UNIONS, which allow them to be manipulated as a single entity. In a structure, the individual items occur sequentially in memory, and the total size of the structure is the sum of its elements. Structures are usually used to create "records", in which related items are grouped together. An array of structures is the common method of implementing an "in-memory" database. A union is similar to a structure, except that the individual items are overlaid in memory, and the total size of the union is the size of its largest element. Unions are usually used to allow a single block of memory to be accessed as different 'C' variable types. An example of this would be in handling a message received in memory, in which a "type" byte indicates how the remainder of the message data should be interpreted. Here are some example of how structures are defined and used (unions are defined and used in an identical manner, except that the word 'union' is substituted for 'struct'): /* Create structure named 'data' with 'a,b,c & d' as members */ struct { int a; int b; char c; char d; } data; /* Create structure template named 'mystruc'... */ /* No actual structure variable is defined */ struct mystruc { int a; int b; char c; char d; }; /* Create structure named 'data' using above template */ struct mystruc data; /* Create structure template 'mystruc', AND define a */ /* structure variable named 'data' */ struct mystruc { int a; int b; char c; char d; } data; MICRO-C Page: 13 /* Create an array of structures, a pointer to a structure */ /* and an array of pointers to a structure * struct mystruct array[10], *pointer, *parray[10]; /* To set value in structure variable/members */ data.a = 10; /* Direct access */ array[1].b = 10; /* Direct array access */ pointer->c = 'a'; /* Pointer access */ parray[2]->d = 'b'; /* Pointer array access */ /* To read value in structure variable/members */ value = data.a; /* Direct access */ value = data[1].b; /* Direct array access */ value = pointer->c; /* Pointer access */ value = parray[2]->d; /* Pointer array access */ 3.5.1 Notes on MICRO-C structure implementation: Structures and Unions as implemented in MICRO-C are similar to the implementation of the original UNIX 'C' compiler, and are bound by similar limitations, as well as a few MICRO-C specific ones. Here is a list of major differences when compared to a modern ANSI compiler: All structure and member names MUST be unique within the scope of the definition. A special case exists, where common member names may be used in multiple structure templates if they have EXACTLY the same type and offset into the structure. This also saves symbol table memory, since only one copy of the member definition is actually kept. MICRO-C does NOT pass entire structures to functions on the stack. Like arrays, MICRO-C passes structures by ADDRESS. Structure variables which are function arguments are accessed through pointers. For source code compatibility with compilers which do pass the entire structure, if you declare the argument as a direct (non-pointer) structure, the direct ('.') operator is used to dereference it, even though it is actually a pointer reference. If you MUST have a local copy of the structure, use something like: func(sptr) struct mystruc *sptr; { struct mystruc data; memcpy(data, sptr, sizeof(data)); ... } MICRO-C Page: 14 To obtain the size of a structure from its template name, use the 'struct' keyword in conjunction with the 'sizeof' operator. In the above example, you could replace 'sizeof(data)' with 'sizeof(struct mystruc)'. To obtain the size of a structure member, you must specify it in the context of a structure reference with another symbol: sizeof(variable.member) or sizeof(variable->member) NOTE: The current compiler allows almost any symbol to the left of the '.' or '->' operator in a 'sizeof', however future versions of the compiler may insist on a structure variable or a pointer to structure variable. Pointers to structures are internally stored as pointers to 'char', and therefore will only increment/decrement by 1 when ++/-- is used. To advance a pointer by the size of a structure, use: ptr += sizeof(struct mystruc) The 'struct' and 'union' keywords are not accepted in a TYPECAST. This is most commonly used to setup a pointer to a structure. Since MICRO-C stores its pointers to structures as pointers to char, you can use (char*) as the typecast, and get the same functionality. A structure name by itself (without '.member') acts in a manner similar to a character array name. With no operation, the address of the structure is returned. You can also use '[]' to access individual bytes in the structure by indexing, although doing so is highly non-portable. MICRO-C allows static or global structures to be initialized in the declaration, however the initial values are read as an array of bytes, and are assigned directly to structure memory without regard for the type or size of its members: struct mystruc data = { 0, 1, 0, 2, 3, 4 }; /* A=0:1, B=0:2, C=3, D=4 */ You can use INT and CHAR to switch back and forth between word/byte initialization within the value list: struct mystruct data = { int 0, 1, char 2, 3 }; /* A=0, B=1, C=2, D=3 */ MICRO-C does not WORD ALIGN structure elements. When using a processor which requires word alignment, it is the programmers responsibility to maintain alignment, using filler bytes etc. when necessary. MICRO-C Page: 15 3.6 Control Statements The following control statements are implemented in MICRO-C: if(expression) statement; if(expression) statement; else statement; while(expression) statement; do statement; while expression; for(expression; expression; expression) statement; return; return expression; break; continue; switch(expression) { case constant_expression : statement; ... break; case constant_expression : statement; ... break; . . . default: statement; } label: statement; goto label; asm "..."; asm { ... } MICRO-C Page: 16 3.6.1 Notes on Control Structures 1) Any "statement" may be a single statement or a compound statement enclosed within '{' and '}'. 2) All three "expression"s in the "for" command are optional. 3) If a "case" selection does not end with "break;", it will "fall through" and execute the following case as well. 4) Expressions following 'return' and 'do/while' do not have to be contained in brackets (although this is permitted). 5) Label names may preceed any statement, and must be any valid symbol name, followed IMMEDIATELY by ':' (No spaces are allowed). Labels are considered LOCAL to a function definition and will only be accessable within the scope of that function. 6) The 'asm' statement used to implement the inline assembly language capability of MICRO-C accepts two forms: asm "..."; <- Assemble single line. asm { <- Assemble multiple lines. ... } MICRO-C Page: 17 3.7 Expression Operators The following expression operators are implemented in MICRO-C: 3.7.1 Unary Operators - - Negate ~ - Bitwise Complement ! - Logical complement ++ - Pre or Post increment -- - Pre or post decrement * - Indirection & - Address of sizeof - Size of a object or type (type) - Typecast 3.7.2 Binary Operators + - Addition - - Subtraction * - Multiplication / - Division % - Modulus & - Bitwise AND | - Bitwise OR ^ - Bitwise EXCLUSIVE OR << - Shift left >> - Shift right == - Test for equality != - Test for inequality > - Test for greater than < - Test for less than >= - Test for greater than or equal to <= - Test for less than or equal to && - Logical AND || - Logical OR = - Assignment += - Add to self assignment -= - Subtract from self assignment *= - Multiply by self assignment /= - Divide by and reassign assignment %= - Modular self assignment &= - AND with self assignment |= - OR with self assignment ^= - EXCLUSIVE OR with self assignment <<= - Shift left self assignment >>= - Shift right self assignment MICRO-C Page: 18 NOTES: 1) The expression "a && b" returns 0 if "a" is zero, otherwise the value of "b" is returned. The "b" operand is NOT evaluated if "a" is zero. 2) The expression "a || b" returns the value of "a" if it is not 0, otherwise the value of "b" is returned. The "b" operand is NOT evaluated if "a" is non-zero. 3.7.3 Other Operators ; - Ends a statement. , - Allows several expressions in one statement. + Separates symbol names in multiple declarations. + Separates constants in multi-value initialization. + Separates operands in function calls. ? - Conditional expression (ternary operator). : - Delimits labels, ends CASE and separates conditionals. . - Access a structure member directly. -> - Access a structure member through a pointer. { } - Defines a BLOCK of statements. ( ) - Forces priority in expression, indicates function calls. [ ] - Indexes arrays. If fewer index values are given than the number of dimensions which are defined for the array, the value returned will be a pointer to the appropriate address. Eg: char a[5][2]; a[3] returns address of forth row of two characters. (remember index's start from zero) a[3][0] returns the character at index [3][0]; MICRO-C Page: 19 3.8 Inline Assembly Language Although 'C' is a powerful and flexible language, there are sometimes instances where a particular operation must be peformed at the assembly language level. This most often involves either some processor feature for which there is no corresponding 'C' operation, or a section of very time critical code. MICRO-C provides access to assembly language with the 'asm' statement, which has two basic forms. The first is: asm "..." ; In this form, the entire text contained between the double quote characters (") is output as a single line to the assembler. Note that a semicolon is required, just like any other 'C' statement. Since this is a standard 'C' string, you can use any of the "special" characters, and thus you could output multiple lines by using '\n' within the string. Another important characteristic of it being a string is that it will be protected from pre-processor substitution. The second form of the 'asm' statement is: asm { ... } In this form, all lines between '{' and '}' are output to the assembler. Any text following the opening '{' (on the same line) is ignored. Due to the unknown characteristics of the inline assembly code, the closing '}' will only be recognized when it is the first non-whitespace character on a line. The integral pre-processor will not perform substitution on the inline assembly code, however the external pre-processor (MCP) will substitute in this form. This allows you to create assembly language "macros" using MCP, and have parameters substituted into them when they are expanded: /* * This macro issues a 'SETB' instruction for its parameter */ #define setbit(bit) asm {\ SETB bit\ } /* * This macro WILL NOT WORK, since the 'bit' operand to the SETB * instruction is contained within a string and is therefore * protected from substitution by the pre-processor */ #define setbit(bit) asm " SETB bit"; MICRO-C Page: 20 3.9 Preprocessor Commands The MICRO-C compiler supports the following pre-processor commands. These commands are recognized only if they occur at the beginning of the input line. NOTE: This describes the limited pre-processor which is integral to the compiler, see also the section on the more powerful external processor (MCP). 3.9.1 #define <name> <replacement_text> The "#define" command allows a global name to be defined, which will be replaced with the indicated text whenever it is encountered in the input file. This occurs prior to processing by the compiler. 3.9.2 #file <filename> Sets the filename of the currently procssing file to the given string. This command is used by the external pre-processor (MCP) to insure that error messages indicate the original source file. 3.9.3 #include <filename> This command causes the indicated file to be opened and read in as the source text. When the end of the new file is encountered, processing will continue with the line following "#include" in the original file. 3.9.4 #ifdef <name> Processes the following lines (up to #else or #endif) only if the given name is defined. 3.9.5 #ifndef <name> Processes the following lines (up to #else or #endif) only if the given name is NOT defined. 3.9.6 #else Processes the following lines (up to #endif) only if the preceeding #ifdef or #ifndef was false. 3.9.7 #endif Terminates #ifdef and #ifndef NOTE: The integral pre-processor does not support nesting of the #ifdef and #idndef constructs. If you wish to nest these conditionals, you must use the external pre-processor (MCP). MICRO-C Page: 21 3.10 Error Messages When MICRO-C detects an error, it outputs an informational message indicating the type of problem encountered. The error message is preceeded by the filename and line number where the error occured: program.c(5): Syntax error In the above example, the error occured in the file "program.c" at line 5. The following error messages are produced by the compiler: 3.10.1 Compilation aborted The preceeding error was so severe than the compiler cannot proceed. 3.10.2 Constant expression required The compiler requires a constant expression which can be evaluated at compile time (ie: no variables). 3.10.3 Declaration must preceed code. All local variables must be defined at the beginning of the function, before any code producing statements are processed. 3.10.4 Dimension table exhausted The compiler has encountered more active array dimensions than it can handle. 3.10.5 Duplicate local: 'name' You have declared the named local symbol more than once within the same function definition. 3.10.6 Duplicate global: 'name' You have declared the named global symbol more than once. 3.10.7 Expected '<token>' The compiler was expecting the given token, but found something else. 3.10.8 Expression stack overflow The compiler has found a more complicated expression than it can handle. Check that it is of correct syntax, and if so, break it up into two simpler expressions. MICRO-C Page: 22 3.10.9 Expression stack underflow The compiler has made an error in parsing the expression. Check that it is of correct syntax. 3.10.10 Illegal indirection You have attempted to perform an indirect operation ('*' or '[]') on an entity which is not a pointer or array. This error will also result if you attempt to index an array with more indices than it has dimensions. 3.10.11 Illegal initialization Local variables may not be initialized in the declaration statement. Use assignments at the beginning of the function code to perform the initialization. 3.10.12 Illegal nested function You may not declare a function within the definition of another function. 3.10.13 Improper type of symbol: 'name' The named symbol is not of the correct type for the operation that you are attempting. Eg: 'goto' where the symbol is not a label. 3.10.14 Improper #else/#endif A #else or #endif statement is out of place. 3.10.15 Inconsistant member offset: 'name' The named structure member is multiply defined, and has a different offset from its first definition. 3.10.16 Inconsistant member type: 'name' The named structure member is multiply defined, and has a different type from its first definition. 3.10.17 Inconsistant re-declaration: 'name' You have attempted to redefine the named external symbol with a type which does not match its previously declared type. 3.10.18 Incorrect declaration A statement occuring outside of a function definition is not a valid declaration for a function or global variable. MICRO-C Page: 23 3.10.19 Invalid '&' operation You have attempted to reference the address of something that has no address. This error also occurs when you attempt to take the address of an array without giving it a full set of indicies. Since the address is already returned in this case, simply drop the '&'. (The error occurs because you are trying to take the address of an address). 3.10.20 Macro expansion too deep The compiler has encountered a nested macro reference which is too deep to be resolved. 3.10.21 Macro space exhausted The compiler has encountered more macro ("#define") text than it has room to store. 3.10.22 No active loop A "continue" or "break" statement was encountered when no loop is active. 3.10.23 No active switch A "case" or "default" statement was encountered when no "switch" statement is active. 3.10.24 Not an argument: 'name' You have declared the named variable as an argument, but it does not appear in the argument list. 3.10.25 Non-assignable You have attempted an operation which results in assignment of a value to an entity which cannot be assigned. (eg: 1 = 2); 3.10.26 Numeric constant required The compiler requires a constant expression which returns a simple numeric value. 3.10.27 String space exhausted The compiler has encountered more literal strings than it has room store. 3.10.28 Symbol table full The compiler has encountered more symbol definitions than it can handle. MICRO-C Page: 24 3.10.29 Syntax error The statement shown does not follow syntax rules and cannot be parsed. 3.10.30 Too many active cases The compiler has run out of space for storing switch/case tables. Reduce the number of active "cases". 3.10.31 Too many defines The compiler has encountered more '#define' statements than it can handle. Reduce the number of #defines. 3.10.32 Too many errors The compiler is aborting because of excessive errors. 3.10.33 Too many includes The compiler has encountered more nested "#include" files than it can handle. 3.10.34 Too many initializers You have specified more initialization values than there are locations in the global variable. 3.10.35 Type clash You have attempted to use a value in a manner which is inconsistant with its typing information. 3.10.36 Unable to open: 'name' A "#include" command specified the named file, which could not be opened. 3.10.37 Undefined: 'name' You have referenced a name which is not defined as a local or global symbol. 3.10.38 Unknown structure/member: 'name' You have referenced a structure template or member name which is not defined. 3.10.39 Unreferenced: 'name' The named symbol was defined as a local symbol in a function, but was never used in that function. This error will occur at the end of the function definition containing the symbol declaration. It is only a warning, and will not cause the compile to abort. MICRO-C Page: 25 3.10.40 Unresolved: 'name' The named symbol was forward referenced (Such as a GOTO label), and was never defined. This error will occur at the end of the function definition containing the reference. 3.10.41 Unterminated conditional The end of file was encountered when a "#if" or "#else" conditional block was being processed. 3.10.42 Unterminated function The end of the file was encountered when a function definition was still open. MICRO-C Page: 26 3.11 Quirks Due to its background as a highly compact and portable compiler, and its target application in embedded systems, MICRO-C deviates from standard 'C' in some areas. The following is a short summary of the major infractions and quirks: The operands to '#' commands are parsed based on separating spaces, and any portion of the line not required is ignored. In particular, the '#define' command only accepts a definition up to the next space or tab character. eg: #define APLUSONE A+1 <-- uses "A+1" #define APLUSONE A +1 <-- uses "A" Comments are stripped by the token scanner, which occurs AFTER the '#' commands are processed. eg: #define NULL /* comment */ <-- uses "/*" Note that since comments can therefore be included in "#define" symbols, you can use "/**/" to simulate spaces between tokens. eg: #define BYTE unsigned/**/char Include filenames are not delimited by '""' or '<>' and are passed to the operating system exactly as entered. eg: #include \mc\stdio.h NOTE: The above quirks do not apply when the external pre-processor (MCP) is used. The appearance of a variable name in the argument list for a function declaration serves only to identify that variables location on the stack. MICRO-C will not define the variable unless it is explicitly declared (between the argument list and the main function body). In other words, all arguments to a function must be explicitly declared. MICRO-C is more strict about its handling of the ADDRESS operator ('&') than most other compilers. It will produce an error message if you attempt to take the address of a value which is already a fixed address (such as an array name without a full set of indicies). Since an address is already produced in such cases, simply drop the '&'. The 'x' in '0x' and '\x' is accepted in lower case only. MICRO-C Page: 27 When operating on pointers, MICRO-C only scales the increment (++), decrement (--) and index ([]) operations to account for the size of the pointer: eg: char *cptr; /* pointer to character */ int *iptr; /* pointer to integer */ ++cptr; /* Advance one character */ ++iptr; /* Advance one integer */ cptr[10]; /* Access the tenth character */ iptr[10]; /* Access the tenth integer */ cptr += 10; /* Advance 10 characters */ iptr += 10; /* Advance ONLY FIVE integers */ NOTE: A portable way to advance "iptr" by integers is: iptr = &iptr[10]; /* Advance 10 integers */ Incrementing a pointer to a structure will advance only one (1) byte in memory. To advance to the "next" instance of the structure, use: ptr += sizeof(struct template); The INDEXING operator '[]' is not commutative in MICRO-C. In other words 'array[index]' cannot be expressed as 'index[array]'. MICRO-C does not support "complex" declarations which use brackets '()' for other than function parameters. These are most often used in establishing pointers to functions: int (*a)(); /* Pointer to function returning INT */ (*a)(); /* Call address in 'a' */ Since MICRO-C allows you to call any value by following it with '()', you can get the desired effect in the above case, by declaring 'a' as a simple pointer to int, and calling it with the same syntax: int *a; /* Pointer to INT */ (*a)(); /* Call address in 'a' */ MICRO-C will not output external declarations to the output file for any variables or functions which are declared as "extern", unless that symbol is actually referenced in the 'C' source code. This prevents "extern" declarations in system header files (such as "stdio.h") which are used as prototypes for some library functions from causing those functions to be loaded into the object file. Therefore, any "extern" symbols which are referenced only by inline assembly code must be declared in the assembly code, not by the MICRO-C "extern" statement. MICRO-C Page: 28 Unlike some 'C' compilers, MICRO-C will process character expressions using only BYTE values. Character values are not promoted to INT unless there is an INT value involved in the expression. This results in much more efficent code when dealing with characters, particularily on small processors which have limited 16 bit instructions. Consider the statement: return c + 1; On some compilers, this will sign extend the character variable 'c' into an integer value, and then ADD an integer 1 and return the result. MICRO-C will ADD the character variable and a character 1, and then promote the result to INT before returning it (results of expressions as operands to 'return' are always promoted to int). Unfortunately, programs have been written which rely on the automatic promotion of characters to INTs to work properly. The most common source of problems is code which attempts to treat CHAR variables as UNSIGNED values (many older compilers did not support UNSIGNED CHAR). For example: return c & 255; In a compiler which always evaluates character expressions as INT, the above statement will extract the value of 'c' as positive integer ranging from 0 to 255. In MICRO-C, ANDing a character with 255 results in the same character, which gets promoted to an integer value ranging from -128 to 127. To force the promotion within the expression, you could CAST the variable to an INT: return (int)c & 255; The same objective can be achieved in a more efficent (and correct) manner by declaring the variable 'c' as UNSIGNED CHAR, or by CASTing the variable to an UNSIGNED value: return (unsigned)c; Note that this is not only more clearly shows the intent of the programmer, but also results is more efficent code generated. MICRO-C Page: 29 A related quirk arises because most processors do not support a simple efficent method for adding or subtracting a SIGNED 8 bit quantity and a 16 bit quantity. The code generators supplied with MICRO-C make the assumption that character values being added or subtracted to/from integers will contain only POSITIVE values, and thus use UNSIGNED addition/subtraction. This allows much more efficent code to be generated, as the carry/borrow from the low order byte of the operation is simply propagated to the high order byte of the result (an operation supported in hardware by the CPU). For those rare instances where you do wish to add/subtract a potentially negative character value to/from an int, you can force the expression to be performed in less efficent fully 16 bit arithmetic by casting the character to an int. int i; char c; ... i += c; /* Very efficent... B should be positive */ i += (int)c; /* Less efficent... B can be negative */ Read the notes at the end of the section entitled "Structures and Unions" for information on limitations or differences from standards 'C' in MICRO-C's implementation of structures and unions. MICRO-C Page: 30 4. ADVANCED TOPICS This section provides information on the more advanced aspects of MICRO-C, which is generally not needed for casual use of the language. 4.1 Conversion Rules MICRO-C keep track of the "type" of each value used in all expressions. This type identifies certain characteristics of the value, such as size range (8/16 bits), numeric scope (signed/unsigned), reference (value/pointer) etc. When an operation is performed on two values which have identical "types", MICRO-C assigns that same "type" to the result. When the two value "types" involved in an operation are different, MICRO-C calculates the "type" of the result using the following rules: 4.1.1 Size range If both values are direct (not pointer) references, the result will be 8 bits only if both values were 8 bits. If either value was 16 bits, the result will be 16 bits. If one value is a pointer, and the other is direct, the result will be a pointer to the same size value as the original pointer. If both values were pointers, the result will be a pointer to 16 bits only if both original pointers referenced 16 bit values. If either pointer referenced an 8 bit value, the result will reference an 8 bit value. 4.1.2 Numeric Scope The result of an expression is considered to be signed only if both original values were signed. If either value was an unsigned value, the result is unsigned. 4.1.3 Reference If either of the original values was a pointer, the result will be a pointer. Note that this "calculated" result type is used for partial results within an expression. Whenever a symbol such as a variable or function is referenced, the type of that symbol is taken from its declaration, no matter what "type" of value was last stored (variable) or returned (function). The TYPECAST operation may be used to override the type calculated for the result of an expression if necessary. MICRO-C Page: 31 4.2 Assembly Language Interface Assembly language programs may be called from 'C' functions and vice versa. These programs may be in the form of inline assembly language statements in the 'C' source code, or separately linked modules. When MICRO-C calls any routine ('C' or assembler), it first pushes all arguments to the routine onto the processor stack, in the order in which they occur in the argument list to the function. This means that the LAST argument to the function is LOWEST on the processor stack. Arguments are always pushed as 16 bit values. Character values are extended to 16 bits, and arrays are passed as 16 bit pointers to the array. (MICRO-C knows that arrays which are arguments are actually pointers, and automatically references through the pointer). After pushing the arguments, MICRO-C then generates a machine language subroutine call, thereby executing the code of the routine. Since the compiler uses the ACCUMULATOR and INDEX REGISTER, it will automatically save them (if necessary) during processing of the arguments to the function (even if no arguments are present), and therefore these registers do not have to be preserved by the called routine. NOTE that any other registers used by the code generation routines (register variables etc) will not be saved by MICRO-C, and should be preserved by called functions if their content is to be relied on between function calls. Once the called routine returns, the arguments are removed from the stack by the calling program. This usually consists of simply adding the number of bytes pushed as arguments to the stack pointer. When the called function executes, the first thing usually done is to push any registers which must be preserved, and to reserve space on the stack for any local variables which are required. In some implementations, a "base" register may also be established to provide a stable reference to the local variables and arguments. It is the responsibility of the called function to remove any saved registers and local variable space from the stack before it returns. If a value is to be returned to the calling program, it is expected to be in the ACCUMULATOR register. MICRO-C Page: 32 Local variables in a function may be referenced as direct offsets from the "base" register or stack pointer. Note that if the stack pointer is used, offsets must be adjusted for the number of bytes which are pushed and popped on the stack during the execution of the function. The address of a particular local variable is calculated as: "base register" - (Size of all local variables in bytes) + (size of all preceeding local variables in bytes) -- OR -- "stack pointer" + (# bytes pushed during function execution) + (size of all preceeding local variables in bytes) Arguments to a function may also be referenced as direct offsets from the "base" register or stack pointer, in much the same way as local variables are. The address of a particular argument is calculated as: "base register" + (# bytes pushed at entry of function to preserve registers etc.) + (Size of return address on stack (usually 2)) + (# arguments from LAST argument) * 2 -- OR -- "stack pointer" + (# bytes pushed during function execution) + (size of all local variables in bytes) + (# bytes pushed at entry of function to preserve registers etc.) + (Size of return address on stack (usually 2)) + (# arguments from LAST argument) * 2 NOTE: The (number of bytes pushed at entry of function) is a function of the code generator, and depends on the particular MICRO-C implementation. Examine some assembly output from the compiler to determine the actual number on your system. MICRO-C Page: 33 If a function has been declared as "register", MICRO-C will load the accumulator with the number of arguments which were passed, each time the function is called. This allows the function to determine the location of the first argument. The address of the first argument passed to a "register" function may be calculated as: (accumulator contents) * 2 + "base register" + (# bytes pushed at entry of function to preserve registers etc.) + (Size of return address on stack (usually 2)) -- OR -- (accumulator contents) * 2 + "stack pointer" + (# bytes pushed during function execution) + (size of all local variables in bytes) + (# bytes pushed at entry of function to preserve registers etc.) + (Size of return address on stack (usually 2)) Global variables exist at absolute addresses and may be referenced directly by name from within assembler programs. Keep in mind however, that MICRO-C uses only the first 15 characters of a symbol's name. Also, many code generators will reduce the size of names even further, often using an algorithm to compress the name rather than simply truncating it. For this reason, it is a good idea to avoid using global symbol names which are longer than 8 characters if they are to be referenced from within assembly language programs. MICRO-C Page: 34 4.3 Compiling for ROM Assuming the code generator does not employ "tricky" programming techniques such as self modifying code, the output from the compiler is entirely "clean", and may be placed in Read Only Memory (ROM). The compiler places all initialized global variables in the output file as part of the code image. When the program is stored in ROM, those variables are also stored in ROM, and will not be modifiable. When the program is to be placed in ROM, you may not initialize any variables which you intend to modify later. Those variables must be explicitly initialized by code executed at the beginning of the program. Initialized variables which you do not intend to modify (such as tables etc.) may be initialized in the declaration, and will be permanently saved in the ROM as part of the static code image. The processor stack pointer must be set up before any expressions are evaluated, or any function calls are performed. One way to do this is with inline assembly statements at the top of the file. All non-initialized global variables must be located in RAM. The compiler usually outputs all such variables at the very end of the compilation, just after dumping the literal pool. The global variables may be moved to RAM by editing the output file, and placing an appropriate "ORG" statement at the beginning of the globals. You must also be sure to relocate the temporary location which is defined in the startup code and runtime library ("?temp"), as well as any global variables used in the standard library functions. MICRO-C Page: 35 5. THE MICRO-C COMPILER The heart of the MICRO-C programming environment is the COMPILER. This program reads a file containing a 'C' source program, and translates it into an equivalent assembly language program. The compiler includes its own limited pre-processor, which is suitable for compiling programs requiring only non-parameterized MACRO substitution, simple INCLUDE file capability, and single-level CONDITIONAL processing. 5.1 The MCC command When created using the 'io' routines supplied on the distribution disk, the format of the MICRO-C Compiler command line is: MCC [input_file] [output_file] [options] [input_file] is the name of the source file containing 'C' statements to read. If no filenames are given, MCC will read from standard input. [output_file] is the name of the file to which the generated assembly language code is written. If less than two filenames are specified, MCC will write to standard output. 5.1.1 Command Line Options -c - Includes the 'C' source code in the output file as assembly language comments. -f - Causes the compiler to "Fold" its literal pool. (Identical strings not contained in explicit variables only once in memory). -l - Enables MCC to accept line numbers. (At beginning of line, followed by ':'). -q - Instructs MCC to be quiet, and not display the startup message when it is executed. MICRO-C Page: 36 6. PORTING THE COMPILER There are two major goals to accomplish when porting the MICRO-C compiler to a new machine. The first is to make the compiler run in the new environment, and the second is to make it produce code for that environment. These two goals do not always go hand in hand. For example, it may be desirable to implement a CROSS COMPILER which generates code for a different system from that on which it runs. The usual method of porting a compiler to system A when a version of the compiler is already running on system B, is to first create a compiler which runs on system B, producing code for system A. This "new" compiler may then be used to create a compiler which runs on system A. The compiler consists of a module "compile" which contains, the main compiler, input scanner, regular expression parser, and symbol table management routines. This is the "static" portion of the compiler which does not change in different implementations. To create a working compiler, the above module must be compiled and linked with an "io" module and "code" module. The "io" module performs the necessary initialization and I/O to allow the compiler to run in a particular environment (goal #1). The "code" module generates the assembly language output code for a particular environment (goal #2). The compiler uses NO system library functions, and relies on NO system services (other than those used by the "io" and "code" modules). This allows the compiler to be ported to virtually any system. All fixed compiler parameters (such as table sizes etc) are contained in the header file "compile.h". When porting to systems with very limited RAM (less than 64k), you may have to reduce some of these sizes in order to get it to fit. 6.1 The "io" module The "io" module required by the compiler must contain the following function definitions: The function "main()" is called by the operating system when the MICRO-C compiler is executed. It is responsable for parsing any parmeters and command qualifiers, opening the input and output files, performing any other initializations that might be required, and then invoking the function "compile()" with no arguments. The program filename should also be copied into the array 'file_names' found in compile.c. The "compile" function is internal to the "compile.c" module, and will never return. For UNIX and other systems which support I/O redirection, "stdin" and "stdout" are often used for the input and output files. MICRO-C Page: 37 The function "terminate(rc)" is called when the compiler has finished all processing and wishes to terminate. The value "rc" is a return code: 0 = Success, program compiled without error, -1 = Compile was aborted due to severe errors, n = Compile finished, but had 'n' errors. The function "put_chr(chr, flag)" is passed a character to output, as well as a flag. The flag will always be zero when the character is being sent to the console terminal, or non-zero when the character is to be written to the output file. The function "put_str(*ptr, flag)" is passed a pointer to a zero terminated string, which is to be written to the console or output file as determined by "flag". The function "put_num(number, flag)" is passed a 16 bit number, which is to be written in printable form to the console or output file as determined by "flag". The function "get_lin(*ptr)" is passed a pointer to a character array, and should read a single line from the currently open input file into that array. The manifest definition "LINE_SIZE", found in "compile.h" may be used to determine the maximum line length acceptable to the compiler. Return of a non-zero value indicates the end of file condition. The function "f_open(*ptr)" is passed a pointer to a filename. The new file should be opened, and if successful, the old input file should remain open and be "stacked", so that it can be later returned to, and a non-zero value returned. A zero value indicates to the compiler that the file could not be opened. The function "f_close()" is responsible for closing the currently open input file, and returning to the "stacked" previously open one. It receives no parameters. Note: multiple "opens" may be stacked - see the manifest definition of "INCL_DEPTH" in the "compile.h" file. 6.1.1 Notes on I/O module 1) It is the responsibility of the "put" routines to translate the NEWLINE '\n' (0x0a) character into whatever line termination character(s) are required by the target operating system. 2) If unix "stdin" and "stdout" are not used, it is the responsibility of "main" to display appropriate error messages if the input or output files could not be opened. Refer to the sample "io" modules distributed with the compiler. MICRO-C Page: 38 6.2 The "code" module In order to insure that MICRO-C is portable to virtually any environment, the compiler makes few assumptions about the processor or system software of the target system. The "code" module is relied on to produce all machine instructions and assembler directives written to the output file. The only two real "assumptions" made about the target processor are: 1) It is assumed that the target processor has an "accumulator" register in which all math operations are performed, and that the lower 8 bits of this register may be accessed independantly. 2) It is assumed that the target processor has one "index" register which may be loaded with a 16 bit value, and that memory references may be made indirectly through this register. If the target processor does not support the above features, it may be possible to write a code generator for it using some other features of the processor. For example, if an "index" register does not exist, it may often be implemented using two bytes of reserved memory. The code generation module required by the compiler must contain the following function definitions: The function "do_comment(*ptr)" is passed a pointer to a character string, which is should write to the output file as a single line COMMENT (used by '-c' option). This function is also called with a NULL (0) pointer anytime the compiler wants to insure that the code generator is reset to begin a new line of output. The function "do_asm(*ptr)" is passed a pointer to a character string, which it should write to the output file as a single line STATEMENT (used for inline assembly language). The function "def_module()" is called at the beginning of compilation, before any other code generator functions are called. It is used to output any pre-amble needed by the assembler. The function "end_module()" is called at the very end of compilation, and is the last code generator function called. It is used to output any post-amble needed by the assembler. MICRO-C Page: 39 The function "def_static(symbol)" is passed am index into the compiler symbol tables for a global variable, which is about to be initialized as static storage. The call to this function will be immediately followed by a call to "init_static" or "end_static". The "init_static(token, value, word)" function is passed a token and value, with which it should initialize a single element of static storage. The "word" flag will be non-zero if the value is a 16 bit element. Only the "constant" tokens (NUMBER, STRING, LABEL) and SYMBOL need be handled by this routine. When SYMBOL is used, the constant ADDRESS of the passed symbol should be generated. The "end_static()" function is called to terminate the definition of static storage. NOTE: "init_static" should not rely on "def_static" being called first, since it is also called immediately following "def_label" to define "switch" tables (See "do_switch"). After the table is defined, the compiler will call "end_static()" to terminate the initialization and set up for the next. The function "def_global(symbol, size)" is called at the end of the compile, once for each non-static global variable which was defined. The "size" paremeter indicates the number of BYTES of memory to be reserved for that variable. The function "def_extern(symbol)" is called at the end of the compile, once for each unresolved external symbol which was defined. This routine should examine the type of the symbol and output the appropriate assembler directives to allow it to be referenced in another module. The "def_func(symbol, size)" routine is called to start a function definition. The "symbol" parameter is an index into the compiler symbol tables for the function entry being defined. The "size" parameter indicates how many bytes of memory should be allocated on the stack for local variables. When defining "register" functions, care must be taken that the entry code preserves the contents of the accumulator. The "end_func()" routine is called to terminate a function definition. It should remove anything placed on the stack (including the local variable space allocated by "def_func"), and terminate the function with a "return" instruction. The "def_label(label)" function is called whenever the compiler wants to generate a label in the output file. Each label generated by the compiler is identified by a 16 bit unsigned number. It is up to the code generator to generate a unique label suitable for the target assembler. MICRO-C Page: 40 The "def_literal(*ptr, size)" function is called at the end of the compile, just before non-initialized global symbols are generated. This routine is given a pointer to the compiler's "literal pool", which contains all the character strings occuring during the compilation. The "size" parameter indicates how many characters are in the pool. This pool must be generated in the output file as a string of byte constants. The "call(token, value, type, clean)" function is used to generate a machine language subroutine call to the entity indicated by the "token, value & type" parameters (See later). The "clean" parameter indicates how many entries were pushed onto the stack as arguments, which should be removed following the function call. Note: Since stack entries are TWO bytes in most implementations, the "clean" value must be multiplied by two to get the actual number of bytes to be removed from the stack. The function "jump(label, ljmp)" is called to generate an unconditional jump instruction to the indicated compiler generated label. The "ljmp" flag will be set to zero if the jump references code within the same expression from which it is generated, and non-zero if one or more statements may occur between the jump and the destination label. This allows the code generator to take advantage of "short" jumps if they are available on the target. The function "jump_if(cond, label, ljmp) is called to generate a conditional jump to a compiler generated label. The "cond" value indicates the condition: FALSE = Jump if accumulator is zero, TRUE = jump if accumulator is non-zero. Remaining parameters are the same as above. The function "do_switch(label)" is passed the address of a "switch" table, which contains 16 bit entries, and is of the following format: label-1, value-1, label-2, value-2, .... label-n, value-n, 0, default_label This routine should search the table for the value currently in the accumulator, and if found, it should proceed to the label associated with that value. If the value is not found before the end of the table is encountered (identified by a label value of zero), execution should proceed at the address identified by "default_label". The "index_ptr(token, value, type)" routine should load the index register with the value of the entity indicated by the "token, value & type" parameters. This will always be a 16 bit wide quantity. MICRO-C Page: 41 The "index_adr(token, value, type)" routine should load the index register with the 16 bit address of the object represented by "token, value & type". The only tokens passed to this routine are SYMBOL and ION_STACK. The routine "expand(type)" is called following the evaluation of expressions in "return" and "switch" statements, and is used to insure that the result is a 16 bit value. The routine "accop(oper, type)" is called to perform a "no operand" operation on the accumulator. See "compile.h" for a list of these operations. The "type" passed indicates the type of value expected as a result. The routine "accval(oper, rtype, token, value, type)" is called to perform a "one operand" operation on the accumulator. See "compile.h" for a list of these operations. "rtype" indicates the type of value expected as a result. "token", "value" and "type" indicate the location and type of operand which is being processed. MICRO-C Page: 42 6.2.1 Notes on code generation The meaning of the individual bits in the 16 bit "type" value which is passed to many of the code generation routines, is documented in the "compile.h" file. The meaning of "token" is documented in the "compile.h" file. For each type of "token", "value" has a different meaning: Token Meaning of "value" ---------------------------------------------------------- NUMBER The numeric value of the constant. STRING The offset into the literal pool. LABEL The value of the compiler generated label. SYMBOL Index into global symbol tables. All others Undefined. When token is a SYMBOL, the "value" passed is used to index into the global symbol tables (contained within the "compile" module) to determine information about the variable. When "value" is less than the value of the compiler internal variable 'global_top', the symbol is GLOBAL, otherwise it is LOCAL: s_name[value] - Name of symbol (up to SYMBOL_SIZE chars) s_type[value] - Type of symbol (bits defined in "compile.h"). s_index[value] - Variable index: Local (auto): Stack offset from def_func. Argument: Stack offset from last argument pushed. Static: Unique variable number for symbol. When calculating the stack offset for ARGUMENTs, you must add the number of bytes placed on the stack when the function was called. This includes the local variables, the return address, and any other values that "def_func" might push. There are two popular ways of providing addressability to local variables: If the processor has many registers, you can reserve one as a "base" pointer, and point it at the top of the stack when the function is entered. This allows local variables to be referenced as negative offsets from that "base" register, and arguments to be referenced as positive offsets from it. This approach also allows the stack to be restored directly from this base register when the function terminates. See the section on assembly language interfacing. Another approach is to have the code generator "remember" exactly how many bytes have been pushed onto the stack since "def_func", and adjust the offsets it generates based on the stack contents. This has the advantage of not tying up a register. MICRO-C Page: 43 If you intend to use the MICRO-C Source Linker (SLINK), then you have to insure that whatever variable you use to reference the "literal pool" qualifies as a "compiler generated" label, allowing SLINK to adjust it when processing the source files. Compiler generated labels normally consist of a single character (such as '?'), followed by a decimal number. Since the compiler begins generating its labels with the value '1', you may safely use '0' as the literal pool variable (Ie: '?0'). It is the responsibility of the code generator to keep track of the validity of the upper 8 bits of the accumulator. Appropriate sign extension or clearing of high bits must be performed as necessary to convert signed and unsigned character values when 16 bit results when required. To improve the effiency of conditional statements, the code generator should keep track of the validity of the "zero" flag in the processor's condition code register, and generate appropriate "test" instructions only if necessary when a conditional jump is compiled. For processors not supporting operations with stack contents, the "ON_STACK" and "ION_STACK" tokens may implemented by first popping the top of the stack into a register. The "ISTACK_TOP" token is a special case, because the address on the top of the stack must not be lost. This token may be efficiently implemented, because "ISTACK_TOP" is only used for read/write operations to a stacked calculated address. For example: array1[i] += array2[i]; This statement calculates the address of "array1[i]" (in the index register). Since the "index" register is again used in calculating the address of "array2[i]", the first "index" will be placed on the stack. Once the contents of "array2[i]" are retrieved, it will be added using "ISTACK_TOP", and then re-stored using "ION_STACK". The "ISTACK_TOP" token may thus pop the address into a register, and set a flag indicating to the code generator that the next "ION_STACK" token is to go through that register rather than the top of the stack. Note that since an arithmetic operation may be performed between the two references, you must not use a register which is modified in code generated for arithmetic operations. Refer to the sample code generators distributed with the compiler. MICRO-C Page: 44 6.3 The "compile" module The "compile" module contains the main statement analyser, input scanner, expression parser and symbol table managment routines for the MICRO-C compiler. This module is common to all implementations, and should NOT require changes in most cases. The source code for this module is contained in the "compile.c" file on your distribution diskette, and may be examined for insight into the internal operation of the compiler. This module must be compiled and linked with your I/O and code generation routines to generate a complete executable MICRO-C compiler. To test your code generator, the file "test.c" is provided which when compiled using the new compiler, performs a number of simple tests to verify your code generator. This is not a comprehensive analysis, as it makes no assumptions about the processor, however, it provides a good indication that your code generator is on the right track. A number of fixed parameters to the compiler (such as table sizes etc.) are contained in the header file "compile.h". The most common reason for changing these parameters is to reduce the memory requirements, in order to get MICRO-C to fit into very small systems. If any of the parameters in "compile.h" are to be changed, you must make the changes BEFORE compiling any of the modules (compile, io or codegen). The file "tokens.h" contains symbolic definitions for the tokens parsed by the compiler, the text of which is in the character array variable "tokens". If you make any changes to this token table, MAKE SURE that the contents of the "tokens" variable and the "tokens.h" file agree, otherwise, you will end up with a compiler for a very strange language. 6.3.1 **NOTE for 32 bit compiler users COMPILE.C has one potential portability problem if it is compiled on a machine where an 'int' is not 16 bits. The routine "skip_comment()" uses a 16 bit 'int' to hold the last two characters from the source file, in order to test for the "/*" and "*/" sequences. If an 'int' can contain more that TWO characters, you must modify this function so that only the last two characters are retained. MICRO-C Page: 45 6.4 Porting without a linker It is possible to port MICRO-C using a system which does not support a linker. To do this, you must concatinate all the source files "compile.c", "code.c" and "io.c" into one large file (in that order), and compile them all as one program. When this is done, the ".h" include files need only be included once, and external definitions of variables occuring in other source files should not be used. The source programs on the distribution disk all contain conditional compilation statements (#ifndef), which only perform the necessary #include and external definition statements when compiling as a single file. 6.5 Optimization Techniques The MICRO-C compiler performs the following machine independant optimizations of the output file: 1) All constant expressions are evaluated at compile time, and expressed as a single constant value in the output code. 2) Commutative operations may be reversed to take advantage of partial results already in the accumulator. This includes operations which have an equivalent reverse (ie: "a>(b-1)" may become "(b-1)<a"). 3) Redundant jumps as a result of "return" or "break" statements are suppressed. 4) The sense of jumps in conditional statements are reversed for logically negated expressions, code for '!' is only generated if the value returned by that operator is actually used. 5) Jumps between code generated within a single expression are flagged as "short". Although the MICRO-C compiler produces fairly reasonable code for the processor model it uses, that model is necessarily limited, in order that it might fit a large number of physical targets. There are several simple optimizations which may be performed to further enhance the code generated by the compiler in a specific implementation. MICRO-C Page: 46 6.5.1 Register Usage MICRO-C assumes a single accumulator, and a single index register. Additional terms in complicated expressions are handled by placing temporary results on the processor stack, and re-using those registers. All data placed on the stack is accessed on a "last in - first out" basis. If the target processor has a full compliment of general purpose registers, an optimization may be performed by simply selecting another general purpose register as the accumulator or index register instead of placing the value on the stack. The code generator must keep track of the order in which registers are selected, and which register represents the "top" of the stack. If the number of values "pushed" exceeds the number of available registers, the "oldest" register should be placed on the stack, thereby allowing it to be re-used. 6.5.2 Jump Optimization Although MICRO-C identifies jumps to instructions which span more than one expression as "long", often the addresses are close enough together that short jumps may actually be used. This optimization is particularily useful for processors such as the 8086, which does not support "long" conditional jumps, and therefore must simulate them with a short conditional jump of the opposite sense around a long unconditional jump. 6.5.3 Redundant Load Elimination Since MICRO-C evaluates and processes each statement individually, it does not carry partial results from one statement to another. Consider the following statements: a = x; b = a + 1; MICRO-C generates the code: LOAD x STORE a LOAD a ADD 1 STORE b An optimization may be performed by recognizing that the second "load" instruction is redundant, and can be eliminated. Note: A more efficent (but less readable) way of coding the above statements which would result in the latter code without optimization is: b = (a = x) + 1; MICRO-C Page: 47 6.5.4 Peephole Optimization Consider the statement: a = *++ptr; MICRO-C generates the code: LOAD ptr INCREMENT STORE ptr MOVE ACCUMULATOR TO INDEX LOAD [INDEX] STORE a For a processor supporting a rich set of direct memory addressing modes, the above sequence can be shortened to: INCREMENT_MEMORY ptr LOAD [ptr] STORE a One of the most successful techniques of optimizing the output code is also one of the simplest. Known as "peephole" optimization, the method consists of keeping a window of the last few instructions generated, and scanning the list for known patterns every time a new instruction is added to it. As long as the instructions in the list at least partially match one or more of the "predefined" patterns, additional instructions are collected until either a complete pattern match occurs, or all known patterns are eliminated. If no matches occur, the "oldest" instruction is written to the output file, and the next one becomes the first in the "window". Whenever a pattern is discovered, it is replaced by a new series of instructions which perform the same function, but in a more efficent manner. The new instruction sequences are replaced on the list, which may then be again scanned, allowing further possible reductions to be discovered. Handling of labels in the "window" and their corresponding placement in the output file must be carefully done, in order to preserve the "logical" context of the original code. MICRO-C Page: 48 7. THE MICRO-C PREPROCESSOR The MICRO-C Preprocessor is a source code filter, which provides greater capabilities than the preprocessor which is integral to the MICRO-C compiler. It has been implemented as a stand alone utility program which processes the source code before it is compiled. Due to the higher complexity of this preprocessor, it operates slightly slower than the the integral MICRO-C preprocessor. This is mainly due to the fact that it reads each line from the file and then copies it to a new line while performing the macro substitution. This is necessary since each macro may contain parameters which must be replaced "on the fly" when it is referenced. The integral MICRO-C preprocessor is very FAST, because it does not copy the input line. When it encounters a '#define'd symbol, it simply adjusts the input scanner pointer to point to the definition of that symbol. Keeping the extended preprocessor as a stand alone utility allows you to choose between greater MACRO capability and faster compilation. It also allows the system to continue to run on very small hardware platforms. The additional capabilities of the extended preprocessor are: - Parameterized MACROs. - Multiple line MACRO's. - Nested conditionals. - Ability to undefine MACRO symbols. - Library reference in include file names. MICRO-C Page: 49 7.1 The MCP command The format of the MICRO-C Preprocessor command line is: MCP [input_file] [output_file] [options] [input_file] is the name of the source file containing 'C' statements to read. If no filenames are given, MCP will read from standard input. [output_file] is the name of the file to which the processed source code is written. If less than two filenames are specified, MCP will write to standard output. 7.1.1 Command Line Options MCP accepts the following command line [options]: -c - Instructs MCP to keep comments from the input file (except for those in '#' statements which are always removed). Normally, MCP will remove all comments. -l - Causes the output file to contain line numbers. Each line in the output file will be prefixed with the line number of the originating line from the input file. l=path - Defines the directory path which will be taken to reference "library" files when '<>' are used around an '#include' file name. Unless otherwise specified, the path defaults to: '\mc' -q - Instructs MCP to be quiet, and not display the startup message when it is executed. <name>=<text> - Pre-defines a non-parameterized macro of the specified <name> with the string value <text>. MICRO-C Page: 50 7.2 Preprocesor Commands The following commands are recognized by the MCP utility, only if they occur at the beginning of the source file line: 7.2.1 #define <name>(parameters) <replacement text> Defines a global macro name which will be replaced with the indicated <replacement text> wherever it occurs in the source file. Macro names may be any length, and may contain the characters 'a'-'z', 'A'-'Z', '0'-'9' and '_'. Names must not begin with the characters '0'-'9'. If the macro name is IMMEDIATELY followed by a list of up to 10 parameter names contained in brackets, those parameter names will be substituted with parameters passed to the macro when it is referenced. Parameter names follow the same rules as macro names. eg: #define min(a, b) (a < b ? a : b) If any spaces exist between the macro name and the opening '(', the macro will not be parameterized, and all following text (including '(' and ')') will be entered into the macro definition. If the very last character of a macro definition line is '\', MCP will continue the definition with the next line (The '\' is not included). Pre-processor statements included as part of a macro definition will not be processed by MCP, but will be passed on and handled by the integral MICRO-C preprocessor. 7.2.2 #undef <symbol> Undefines the named macro symbol. further references to this symbol will not be replaced. NOTE: With MCP, macro definitions operate on a STACK. IE: If you define a macro symbol, and then re-define it (without '#undef'ing it first), subsequently '#undef'ing it will cause it to revert to its previous definition. A second '#undef' would then cause it to be completely undefined. MICRO-C Page: 51 7.2.3 #forget <symbol> Similar to '#undef', except that the symbol and ALL SUBSEQUENTLY DEFINED SYMBOLS will be undefined. Useful for releasing any local symbols (used only within a single include file). For example: #define GLOBAL "xxx" /* first global symbol */ ... /* more globals */ #define LOCAL "xxx" /* first local symbol */ ... /* more locals */ /* body of include file goes here */ #forget LOCAL /* release locals */ 7.2.4 #ifdef <symbol> Causes the following lines (up to '#else' of '#endif') to be processed and included in the source file only if the named symbol is defined as a macro. 7.2.5 #ifndef <symbol> Causes the following lines (up to '#else' of '#endif') to be processed and included in the source file only if the named symbol is NOT defined as a macro. NOTE: '#ifdef/#ifndef#else/#endif' may be nested. 7.2.6 #else Toggles the state of the "if_flag", controlling conditional processing. Only has effect in the highest level of suspended processing. IE: Nested conditionals will work properly. If the previous '#ifdef/#ifndef' failed, processing will begin again following the '#else'. If the previous '#ifdef/#ifndef' passed, processing will be suspended until the '#endif' is encountered. NOTE: Since '#else' acts as a toggle, it may be used outside of any '#ifdef/#ifndef' to unconditionally suspend processing up to '#endif'. You can also use multiple '#else's in a single conditional, to swap back and forth between true/false processing without re-testing the condition. 7.2.7 #endif Resets the "if_flag" controlling conditionals, causing processing to resume. Only has effect in the highest level of suspended processing. IE: Nested conditionals will work properly. MICRO-C Page: 52 7.2.8 #include <filename> Causes MCP to open the named file and include its contents as part of the input source. If the filename is contained within '"' characters, it will be opened exactly as specified, and (unless it contains a directory path) will reference a file in the current directory. If the filename is contained within the characters '<' and '>', it will be prefixed with the library path (See 'l=' option), and will therefore reference a file in that library directory. The default library directory is assumed to be '\mc'. For example: #include "header.h" /* from current directory */ #include <stdio.h> /* from library directory */ MICRO-C Page: 53 7.3 Error messages When MCP detects an error during processing of an include file, it displays an error message, which is preceeded by the filename and line number where the error occurs. If more than 10 errors are encountered, MCP will terminate. The following error messages are reported by MCP: 7.3.1 Cannot open include file A '#include' statement on the indicated line specified a file which could not be opened for reading. 7.3.2 Invalid include file name A '#include' statement on the indicated line specified a file name which was not contained within '"' or '<>' characters. 7.3.3 Invalid macro name A '#define' statement on the indicated line contains a macro name which does not follow the name rules. 7.3.4 Invalid macro parameter A '#define' statement on the indicated line contains a macro parameter name which does not follow the name rules. A reference to a macro does not have a proper ')' character to terminate the parameter list. 7.3.5 Too many errors More than 10 errors has been encountered and MCP is terminating. 7.3.6 Too many macro definitions MCP has encountered more '#define' statements than it can handle. 7.3.7 Too many macro parameters A '#define' statement on the indicated line specifies more parameters to the macro than MCP can handle. 7.3.8 Too many include files MCP has encountered more nested '#include' statements than it can handle. MICRO-C Page: 54 7.3.9 Undefined macro A '#undef' or '#forget' statement on the indicated line references a macro name which has not been defined. 7.3.10 Unterminated comment The END OF FILE has been encountered while processing a comment. 7.3.11 Unterminated string A quoted string on the indicated line has no end. To continue a string to the next line, use '\' as the last character on the line. The '\' will not be included in the string. MICRO-C Page: 55 8. THE MICRO-C OPTIMIZER The MICRO-C optimizer is an output code filter which examines the assembly code produced by the compiler, recognizing known patterns of inefficent code (using the "peephole" technique), and replaces them with more optimal code which performs the same function. It is entirely table driven, allowing it to be modified for virtually any processor. Due its many table lookup operations, the optimizer may perform quite slowly when processing a large file. For this reason, most people prefer not to optimize during the debugging of a program, and utilize the optimizer only when creating the final copy. 8.1 The MCO command The format of the MICRO-C Optimizer command line is: MCO [input_file] [output_file] [options] [input_file] is the name of the source file containing assembly statements to read. If no filenames are given, MCO will read from standard input. [output_file] is the name of the file to which the optimized assembly code is written. If less than two filenames are specified, MCO will write to standard output. 8.1.1 Command Line Options MCO accepts the following command line [options]: -d - Instructs MCO to produce a 'debug' display on standard output showing the source code which it is removing and replacing in the input file. NOTE: If you do not specify an explict output file, you will get the debug statements intermixed with the optimized code on standard output. -q - Instructs MCO to be quiet, and not display the startup message when it is executed. MICRO-C Page: 56 8.2 Porting MCO to a new processor The MICRO-C Optimizer is completely table driven, and is fairly easy to port to new processors. The peephole optimization table is called 'peep_table', and consists of two sequential character string entries for each optimization, and is terminated by single NULL pointer. The first is the "take" entry, and represents an instruction sequence which (if found) is to be removed from the output file. The second is the "give" entry, and defines a sequence of instructions to be placed in the output file at that point. NOTE that due to the way the optimizer removes and replaces instruction in the circular "peephole" buffer, the instructions in the "give" entry ARE CODED IN REVERSE ORDER. Characters in the "take" entry with the high bit set ('\200' and higher) are special characters which represent any variable string which may occur in the instruction sequence, and will be replaced with the same string wherever that character occurs in the "give" entry. If the same special character (eg: '\200') occurs more than once in the "take" entry, the corresponding strings must be exactly the same or else the entire sequence will not be matched. There may be up to 8 different such variable strings within each entry, the lower three bits ('\201'-\'207') indicate which variable is referenced. The 4'th bit ('\21x') when used in a "take" entry, allows a match only if the variable string contains only the numeric digits ('0'-'9'). If any other character occurs, the instruction sequence will not be matched. This bit has no effect when used in a "give" entry. The 5'th bit ('\22x') causes the variable string to access through the "complement" table (see later). When used in a "take" entry, this bit allows a match only if the variable string occurs in that table. When used in a "give" entry, causes the complementary string to be output from the table. The processor will stop scanning a variable string when it encounters the character which immediately follows the special character in the "take" entry, or at the end of the input line. MICRO-C Page: 57 There are cases in which you may not want the optimizer to further reduce code substitued for a particular optimization. In this case, simply include a single characer comment in the GIVE entry, in such a position as to prevent the optimizer from recognizing that with any further TAKE entries. The optimzer also includes a "complement" table, which provides the ability to define and use "opposites" within an optimization. The most common use for this feature is in the processing of conditional jumps, in cases when an optimization reverses the logcal sense of a jump. The complement table is called "not_table", and contains an even number of character string entries. The table is terminated by a single NULL pointer. Each even/odd pair of entries in this table are considered to be complementary strings. When a "give" entry specifies that the complement of a variable string should be output, this table is scanned, and if the original string is found, the corresponding complementary entry is output instead. If the string is not found, it is output without modification. Note that you do not have to access the complementary table in both the "give" and "take" optimization entries. For example, you could specify a "special character" of '\221' in the "take" entry, indicating that you are using variable location 1, and that the string must occur in the complement table. You could then use "\201" in the "give" entry to output the original string, or "\221" to output the complement. MICRO-C Page: 58 9. THE SOURCE LINKER Many small development environments have assemblers which do not directly support an object linker. This causes a problem with 'C' development, because the library functions must be included in the source code, with several drawbacks: 1) There is no way to automatically tell which library functions to include, therefore, you must do it manually. 2) 'C' library functions must be re-compiled every time, in order to avoid conflict between compiler generated labels. The MICRO-C Source Linker (SLINK) helps overcome these problems, by automatically joining previously compiled (assembly language) source code from the library to your programs compiler output. Only those files containing functions which you reference are joined (Taking into consideration of course any functions called by the included library functions etc...). As the files are joined, compiler generated labels are automatically adjusted to be unique within each file. 9.1 The SLINK Command The format of the SLINK command line is: SLINK [input_file] [output_file] [options] [input_file] is the name of the source file containing the compiler output from your program. If no filenames is given, SLINK will read from standard input. [output_file] is the name of the file to which the linked source code is written. If less that two filenames are given, SLINK will write to standard output. 9.1.1 Command line options SLINK accepts the following command line [options]: ? - Display command line help summary. i=name - Specify name of the External Index File. -l - Instructs SLINK to list each library used. l=path - Identifies the directory path which will be taken to reference "library" files. If not specified, it defaults to: '\mc\slib' p=char - Identifies the PREFIX character for compiler generated symbols. Defaults is '?'. -q - Inhibit display of the startup message. t=string- Prefix to prepend to temporary filenames. If not specified, default is "$". MICRO-C Page: 59 9.2 The External Index File SLINK uses a special file from the library to determine which symbols are in which files. This files is called the EXTERNAL INDEX FILE, and is found in the library directory (see 'l=' option), under the name "EXTINDEX.LIB". This file contains entries which cross-reference external symbols to files. Each entry is as follows: 1) Any lines beginning with '<' contain the names of files which are to be processed and included at the BEGINNING of the program (Before your source file). This the best way to include the startup code and any runtime library routines which are required at all times, and also provides a method of initializing any segments used. eg: <6809rl.asm 2) Any lines beginning with '^' contain the names of files which are to be processed AFTER the program and library source files, but BEFORE any uninitialized data areas are output. This is the best way to set up the location and storage class of the uninitialized data if it does not immediately follow the executable program code, and also providing any postamble needed by the segments. 3) Any lines beginning with '>' contain the names of files which are to be processed at the END of the program, after all other information is output. This is the best way to define heap memory storage, and to provide any post-amble needed by the assembler. 4) Any lines beginning with '-' contain the names of files which are to be included if any of the following symbols (Up to another '<', '^', '>', '-' or '$') are referenced. eg: -printf.asm format.asm fgets.asm fget.asm NOTE: In most cases, the library functions will contain indications of any external references that they do, in which case SLINK will automatically include those files even of the names are not mentioned on the '-' line. In the example above, the following would suffice: -printf.asm 5) The names of each symbol which may be referenced externally must follow the '<', '^', '>' or '-' entry. Symbols must occur one per line, with no leading or trailing spaces. eg: printf fprintf sprintf MICRO-C Page: 60 6) A line beginning with '$' is used to define the pseudo- opcode used by SLINK to reserve uninitialized data at the end of the output file. Only one line beginning with '$' should be entered into the EXTINDEX.LIB file. The remainder of this line, including all spaces etc. is entered between each symbol name, and the decimal size (in bytes) which is written to the output file. eg: '$ RMB ' <- Quotes are for clarity A complete example: -printf.asm printf fprintf sprintf -scanf.asm scanf fscanf sscanf <PREFIX.asm ^MIDDLE.ASM >SUFFIX.ASM $ RMB In summary, the output file is written from: 1 - The '<' (prefix) files * 2 - The program source files *\ 3 - Library files referenced (if any) * > See note 4 - Segments 1-9 from above ... */ 5 - The '^' (middle) files 6 - Segments 1-9 from middle files * See note 7 - Uninitialized data definitions (if any) 8 - The '>' (suffix) files 9 - Segments 1-9 from sufix file(s) * See note * NOTE: If these files contain multiple segments (see later), all segments are grouped and written in seguential order. IE: Seg 0 from all files is written, followed by Seg 1, etc. 9.3 Multiple source files SLINK processes only one source file, and resolves external references only to the library files. This is because the public symbol information is available in the EXTINDEX.LIB files, which contains no entries for the user supplied programs. If you wish to use source code linking for a program which contains multiple source files, use "#include" statements at the end of your main program to include all of the other parts into a single source file. MICRO-C Page: 61 9.4 Source file information 9.4.1 SLINK Directives SLINK interpretes several "directives" which may be inserted in the input source files to control the source linking process. These directives must be on a separate line, beginning in column 1, and must be in uppercase. They are removed by SLINK during processing, and thus will not cause conflict with the normal syntax used by your assembler. $SE:<0-9> The '$SE' directive is used by SLINK to define multiple output segments. Up to 10 segments are allowed, with segment 0 being the default which is selected when a file is first encountered. Other segments (1-9) when selected via this directive are written to temporary files, and re-joined at the end of processing in sequential order. this allows you to separate sections of the source file (such as initialized data, literal pool etc.) into distinct areas of memory. $DD:<symbol> <size> The '$DD' directive is used to define uninitialized data storage areas, which will be allocated by SLINK between the '^' (middle) and '>' (suffix) files. This allows you to allocate unitialized data outside of the bounds of the executable image, and thus exclude it from being saved to disk. This action may be thought of as an additional (11'th) segment which is available for uninitialized data only, and which avoids the temporary file read/write overhead associated with use of the other segments. This directive is normally placed into the source file by the "def_global" routine in the MICRO-C code generator. $EX:<symbol> The '$EX' directuve is used by SLINK to identify any symbols which are externally referenced. Whenever a '$EX' directive is found, SLINK searches the EXTINDX.LIB file for the named symbol, and marks the corresponding files for inclusion in the program. This directive is normally placed into the source file by the "def_extern" routine in the MICRO-C code generator. If you are writing assembly language programs for the library, be sure to include "$SE:<0-9>" directives for any segments you wish to access, "$DD:<symbol> <size>" directives for any uninitialized data you wish to allocate, and "$EX:<symbol>" directives for any symbols which you externally reference. If you wish to place an assembly language comment on the same line, make sure it is separated from the remainder of the directive by at least one space or tab character. MICRO-C Page: 62 Also, note that '<' (prefix) files may contain "$SE", "$DD" and "$EX" directives, '^' (middle) files may contain "$SE" and "$DD" but not "$EX", and '>' (suffix) files may only contains "$SE" directives. Basically, the rule is that "$EX" cannot be used after the libraries are included and "$DD" cannot be used after the uninitialized data is output. Since the middle and suffix files are ALWAYS included, simply insure that all external references and data declarations needed by any of them are performed in the '<' (prefix) file. 9.4.2 Compiler generated labels As it processes each source file, SLINK scans each line for symbols which consist of the '?' character (See 'p=' option), followed by a number. If it finds such as symbol, it inserts a two character sequence ranging from 'AA' to 'ZZ' between the '?', and the number. This sequence will be incremented for each source file processed, and thus insures that the compiler generated symbols will be unique for each file. If you are writing assembly language programs for the library, you must be careful to avoid using identical local symbols in any of the library files, one way to do this is to use symbols which meet the above criteria. 9.5 The SCONVERT command SCONVERT is a utility which assists in converting existing assembly language source files into a format which is more suitable for use by the SLINK. Two main functions are performed: 1) All comments are removed, and all spacing is reduced to a single space. This minimizes the size of the file, and helps decrease linkage time. 2) All symbols defined in the file which are not identified as "keep" symbols are converted to resemble the MICRO-C compiler generated symbols. This allows SLINK to adjust them to be unique within each source file. The format of the SCONVERT command line is: SCONVERT [input_file] [output_file] [options] [input_file] is the name of the source file containing the original assembly language program. If no filenames is given, SCONVERT will read from standard input. [output_file] is the name of the file to which the converted source code is written. If less that two filenames are given, SCONVERT will write to standard output. MICRO-C Page: 63 9.5.1 Command line options SCONVERT accepts the following command line [options]: ? - Display command line help summary. c=char - Identifies the character used to begin a comment at the trailing end of a source line. If no 'c=' is defined, SCONVERT will terminate processing at the first blank or tab which follows the operand field. C=char - Identifies the charcter which indicates a comment line in the source code. Defaults to '*'. k=name - Identifies a symbol name to KEEP. This symbol will not be converted. Multiple 'k=' are permitted. K=file - Identifies a file containing the names of symbols to KEEP, one per line. Multiple 'K=' are permitted. p=char - Identifies the PREFIX character which is to be used for the converted symbols. Defaults to '?'. -q - Instructs SCONVERT to be quiet, and not issue its startup message. SCONVERT identifies symbols in the input source file as any string beginning with 'A-Z', 'a-z', '_' or '?', and containing these charcters plus the digits '0-9'. If your assembler source files uses any other characters in its symbols, you must edit your sources and change the symbols. 9.6 The SRENUM command SRENUM is a small utility which re-numbers the compiler generated symbols within a assembly language source file. This is useful if you have made added symbols to the file by hand, and wish to make it "pretty" before adding it to the library etc. The format of the SRENUM command is: SRENUM [input_file] [output_file] [options] 9.6.1 Command line options SRENUM accepts the following command line [options]: ? - Display command line help summary. p=char - Identifies the PREFIX character which is to be used to recognize compiler generated symbols. -q - Instructs SRENUM to be quiet, and not issue its startup message. MICRO-C Page: 64 9.7 The SINDEX command SINDEX is a utility which assists in the creation of the EXTINDEX.LIB file used by SLINK. When you run SINDEX, it examines all of the '.ASM' files in the current directory, and writes a EXTINDEX.LIB file which contains a '-' type entry for each file, and external symbol entries for any labels which it finds conforming to the 'C' naming conventions (Starts with 'a-z', 'A-Z' or '_', and contains only 'a-z', 'A-Z', '0-9' or '_'). Once you have run SINDEX, you must manually edit the EXTINDEX.LIB file, and remove any file or symbol entries which you do not wish to have available as external references, as well as insert any necessary entries for '<', '^', '>' and '$' commands. 9.7.1 Command line options SINDEX accepts the following command line options: ? - Display command line help summary. i=name - Specify name for index file to be written. Dafault is "EXTINDEX.LIB". You may also instruct SINDEX to search for a file pattern other than '*.ASM' by passing it as a command line parameter. eg: SINDEX *.A86 9.8 The SLIB command Once you have constructed your source library, you may from time to time want to make minor changes to it, either adding new functions, or removing old ones ones. You could make such changes simply by editing the EXTINDEX.LIB file, however you would have to be very careful not to add or delete the wrong entry, and you would have to manually determine if adding or removing the file would adversly affect the remainder of the library. For example, you could accidently add a duplicate of another symbol name, or remove a symbol which is referred to by another file. To simplify maintenance of source libraries, you can make use of the "Source Librarian", a utility program which automates the addition and removal of source files, and automatically reports of any inconsistancies occuring in the source library. To use SLIB, you must first position yourself to the directory containing the source library. And then execute SLIB using the command options described later to indicate the action to be taken on the library. If you do not specify any actions, SLIB will simply examine the library and report its size and content. MICRO-C Page: 65 You may use multiple command options in a single SLIB command if you wish to add and/or remove more than one file at a time. After executing any command, SLIB will report on any inconsistancies which it finds in the library. If any are found, and you have used a command which caused changes, SLIB will prompt for permission before writing the updated library file. 9.8.1 Command line options The following options control the action(s) which will be performed by SLIB on the source library index file. ? - Display command line help summary. ?=file - Display information about named source file. a=file - Add specified file as standard functions. i=index - Use the specified INDEX file, if not specified, the filename 'EXTINDEX.LIB' is assumed. m=file - Add named source file as a MIDDLE file. p=file - Add named source file as a PREFIX file. -q - Quiet mode: SLIB will not display informational messages or ask for permission to update index. r=file - Remove the named file from the index. s=file - Add named source file as a SUFFIX file. -w - Write inhibit: SLIB will not write the updated library index. This is useful if you just want to see what would happen if you add or remove a file. MICRO-C Page: 66 9.9 Making a source library To make a complete source linkable library, follow these basic steps: 1) If you have assembly language library functions, run the SINDEX utility to create an index file with the names of any symbols defined in them. Edit this file and remove all names which are LOCAL to the files. Only the global function and variable names should remain. NOTE: Also leave in any other symbols or names which you don't want changed by SCONVERT. 2) Use SCONVERT to convert the assembly language sources into library format, using the index file created above as your KEEP file (K=EXTINDEX.LIB). You may send the output files directly to your source library directory. 3) Edit the converted assembly language sources to change any declarations for uninitialized data into '$DD' directives, and to add any '$SE' and '$EX' directives which may be needed. 4) Compile all of your 'C' library functions to assembly language, using a code generator which outputs the appropriate directives for SLINK. Send the ASM output files to your source library directory. 5) From within your library directory, run the SINDEX utility again. This will create an index file (EXTINDEX.LIB) which contains the names of all global symbols. 6) Edit the index file and remove any non-public symbol names. You should also change the headers for the '<', '^' and '>' files, and add the '$' record for reserved memory information. 7) Run the SLIB utility, to check the new library for duplicate symbols, unresolved external references and any other inconsistancies. MICRO-C Page: 67 10. THE MAKE UTILITY The MAKE utility provides a method of automating the building of larger programs consisting of more that one module. The main benefit of MAKE is that it keeps track of the files that each module is dependant on, and will rebuild a module if any of those files have been modified since the module was last built. This frees the programmer from the task of remembering which files have been changed, and the commands needed to rebuild the dependant modules. 10.1 MAKEfiles To use MAKE, you must first create a MAKEFILE, which is a text file containing entries for each module in the program. Each entry consists of a DEPENDANCY list, and a series of COMMANDS. 10.1.1 MAKEfile Entries A dependency list in MAKE is a line which contains the name of the module, followed by a ':', followed by the names of any files on which it depends. The module name MUST begin in column 1. When MAKE is invoked, it will process each dependancy list, and will execute any following commands (up to another dependancy list) if (1) the module does not exist, or (2) if any of the files to the right of the ':' have a timestamp which is later than that of the module. For example: main.obj : main.c main.h \\mc\\stdio.h \\mc\\mcc main.c main.asm masm/ml main; -del main.asm In the above example, the 'main.obj'would be rebuilt (by compiling and assembling 'main.c') if either it did not already exist, or any of 'main.c', 'main.h' or '\mc\stdio.h' was found to have a later timestamp. The '-' preceeding the 'del' command prevents it from being displayed. Unless the '-q' option is enabled, MAKE will display any commands not preceeded by '-' as they are executed. NOTE: To enter a single '\' in the MAKEFILE, you must use '\\', this is because like 'C', MAKE uses '\' to "protect" special characters which otherwise are used for special functions (such as '\', '$' and '#'). The first '\' "protects" the second one, allowing it to pass through as source text. MICRO-C Page: 68 10.1.2 Macro Substitutions Sometimes in a MAKEFILE, you have a single file or directory path that you use over and over again. If it is a long directory path, this may involve a lot of typing, and it becomes inconvenient to change that name (if you want to use a different directory etc.) because it is repeated many times. MAKE includes a MACRO facility, which allows you to define variable names which will be replaced with a text string when used in subsequent MAKEfile lines. Names are defined by placing them in the MAKEfile, followed by '=', and the text string. Macro names being defined MUST begin in column one, and may consist of the characters ('a'-'z', 'A'-'Z', '0'-'9', and '_'). Whenever MAKE encounters a '$' in the file, it takes the name immediately following, and performs the macro replacement: mcdir = \\mc main.obj : main.c main.h $mcdir\\stdio.h $mcdir\\mcc main.c main.asm masm/ml main; del main.asm When a macro name is immediately followed by alphanumeric text, use a single '\' to separate it from the text. This "protects" the first character of the text from being interpreted as part of the macro name: mcdir = \\mc\\ main.obj : main.c main.h $mcdir\stdio.h $mcdir\mcc main.c main.asm masm/ml main; del main.asm There are several predefined macro symbols which are available: $* = The full name of the dependant module (name.type). $@ = The name only of the dependant module. $. = The full name of each file in the dependancy list, separated from each other by a single space. $, = The full name of each file in the dependancy list, separated from each other by a single comma. $: = The name only of each file in the dependancy list, separated from each other by a single space. $; = The name only of each file in the dependancy list, separated from each other by a single comma. MICRO-C Page: 69 File names in the dependancy list which are preceeded by '-' will not be included in the '$. $, $: $;' macro expansions: mcdir = \\mc main.obj : main.c -main.h -$mcdir\\stdio.h $mcdir\\mcc $. $@.ASM masm/ml $@; del $@.ASM 10.1.3 MAKEfile Comments Whenever MAKE encounters the '#' character in the MAKEFILE, it treats the remainder of the line as a comment, and does not process it: # Define Directories mcdir = \\mc # Build the MAIN module main.obj : main.c -main.h -$mcdir\\stdio.h # Dependants $mcdir\\mcc $. $@.ASM # Compile masm/ml $@; # Assemble del $@.ASM # Delete tmp 10.1.4 Ordering the MAKEfile MAKE processes the MAKEfile is sequential fashion, with the entries near the top being processed before the entries near the bottom. To insure that each module is built properly, any files appearing in the dependancy list for a module which are themselves dependant on other files, should have MAKEfile entries which occur BEFORE the entries for the modules which are dependant on them: # Define Directories mcdir = \\mc # Build the MAIN module main.obj : main.c -main.h -$mcdir\\stdio.h $mcdir\\mcc $. $@.ASM masm/ml $@; del $@.ASM # Build the SUB module sub.obj : sub.c -sub.h $mcdir\\mcc $. $@.ASM masm/ml $@; del $@.ASM # Bind the executable file # NOTE: If either of the above modules is rebuilt, # this entry will be guarenteed to execute. main.exe : main.obj sub.obj LINK $:; MICRO-C Page: 70 10.2 Directives Like 'C', MAKE recognizes several "directives" in the MAKEfile. These directives are only recognized if they occur at the beginning of the input line: 10.2.1 @include <filename> This command causes the indicated file to be opened and read in as the source text. When the end of the new file is encountered, processing will continue with the line following "@include" in the original MAKEfile. 10.2.2 @ifdef <name> Processes the following lines (up to @else of @endif) only if the given MACRO name is defined. NOTE: <name> should not be preceeded by '$', otherwise its CONTENTS will be tested. 10.2.3 @ifndef <name> Processes the following lines (up to @else of @endif) only if the given MACRO name is NOT defined. 10.2.4 @ifeq <word1> <word2> Processes the following lines (up to @else of @endif) only if the following two words match exactly. This is useful for testing the value of a defined MACRO symbol. 10.2.5 @ifne <word1> <word2> Processes the following lines (up to @else or @endif) only if the following two words do not match. 10.2.6 @else Processes the following lines (up to @endif) only if the preceeding @ifdef, @ifndef, @ifeq or @ifne was false. 10.2.7 @endif Terminates @ifdef, @ifndef, @ifeq and @ifne. 10.2.8 @type <text> Displays the following text. 10.2.9 @abort [text] Terminates MAKE with an 'Aborted!' message. Any text on the remainder of the line will be appended to the message. MICRO-C Page: 71 10.3 The MAKE command The format of the MAKE command line is: MAKE [makefile] [options] [makefile] is the name of the MAKEfile to process. If no name is given, MAKE assumes the default name 'MAKEFILE'. 10.3.1 Command Line Options MAKE accepts the following command line [options]: ? - Causes MAKE to output a short summary of the available command line options. -d - Instructs MAKE to operate in "debug" mode, and display the commands which it would execute, without actually executing them. This provides a method of quickly testing the MAKEFILE. -q - Instructs MAKE to be quiet, and not display the informational messages and commands executed as it progresses. <name>=<text> - Pre-defines a macro of the specified <name> with the string value <text>. This OVERRIDES any definition within the MAKEfile, which may be used to establish a "default" value. MICRO-C Page: 72 10.4 The TOUCH command TOUCH is a small utility program which sets the timestamp of one or more files to the current or specified time/date. It is useful as a method of forcing MAKE to recognize a file as "changed", even when it has not. For example, if you had decided to "undo" several recent changes by restoring a backup of 'main.c', the restored file will probably have a timestamp which is older than the last module which was built. In this case, MAKE would be unaware that the file has changed, and would therefore not rebuild the module. The TOUCH command could then be used to "update" the timestamp of 'main.c' to the current time, causing MAKE to recognize it as a changed file. TOUCH main.c You could also use TOUCH to force rebuilding of several files: TOUCH main.c sub1.c sub2.c Or even ALL '.C' files: TOUCH *.c TOUCH can also be used to set the timestamp of a file to an arbritrary value, this may be useful to PREVENT a change from causing an update: TOUCH main.c t=0:00 d=31/10/80 NOTE: Use of the 't=' or 'd=' parameters to TOUCH allows the possibility that a changed file will go unnoticed. CAUTION is advised. The MSDOS implementation of TOUCH supports '-h' and '-s' options, which cause it to set the timestamp of HIDDEN and/or SYSTEM files. If these options are not used, TOUCH will not affect those types of files. MICRO-C TABLE OF CONTENTS Page 1. INTRODUCTION 1 1.1 Compiler Portability 2 1.2 Code Portability 2 1.3 The compiling process 3 2. THE COMMAND CO-ORDINATOR 4 2.1 The CC command 4 2.2 Using multiple object modules 6 3. THE MICRO-C PROGRAMMING LANGUAGE 7 3.1 Constants 7 3.2 Symbols 7 3.3 Arrays & Pointers 11 3.4 Functions 11 3.5 Structures & Unions 12 3.6 Control Statements 15 3.7 Expression Operators 17 3.8 Inline Assembly Language 19 3.9 Preprocessor Commands 20 3.10 Error Messages 21 3.11 Quirks 26 4. ADVANCED TOPICS 30 4.1 Conversion Rules 30 4.2 Assembly Language Interface 31 4.3 Compiling for ROM 34 5. THE MICRO-C COMPILER 35 5.1 The MCC command 35 6. PORTING THE COMPILER 36 6.1 The "io" module 36 6.2 The "code" module 38 6.3 The "compile" module 44 6.4 Porting without a linker 45 6.5 Optimization Techniques 45 7. THE MICRO-C PREPROCESSOR 48 7.1 The MCP command 49 7.2 Preprocesor Commands 50 MICRO-C Table of Contents Page 7.3 Error messages 53 8. THE MICRO-C OPTIMIZER 55 8.1 The MCO command 55 8.2 Porting MCO to a new processor 56 9. THE SOURCE LINKER 58 9.1 The SLINK Command 58 9.2 The External Index File 59 9.3 Multiple source files 60 9.4 Source file information 61 9.5 The SCONVERT command 62 9.6 The SRENUM command 63 9.7 The SINDEX command 64 9.8 The SLIB command 64 9.9 Making a source library 66 10. THE MAKE UTILITY 67 10.1 MAKEfiles 67 10.2 Directives 70 10.3 The MAKE command 71 10.4 The TOUCH command 72