| 1 |
65 |
motilito |
|
| 2 |
|
|
|
| 3 |
|
|
c80dos.doc
|
| 4 |
|
|
|
| 5 |
|
|
>>> Small-C Version 1-N Compiler Documentation <<<
|
| 6 |
|
|
|
| 7 |
|
|
NOTE: C80DOS.EXE is the MSDOS compiled binary for running on
|
| 8 |
|
|
a standard PC class machine which emits 8080 assembler
|
| 9 |
|
|
that can then be assembled and loaded on the PC using
|
| 10 |
|
|
lasm.cpm and load.cpm with the zrun.com CP/M emulator.
|
| 11 |
|
|
The final output (or any of the intermediate output in
|
| 12 |
|
|
8080 assembler or Intel HEX format) can then be ported
|
| 13 |
|
|
to the CP/M machine by telecommunicating with a any of
|
| 14 |
|
|
a myriad of programs or by writing the disk directly
|
| 15 |
|
|
using something like the Uniform.exe program or its
|
| 16 |
|
|
equivalent. Hopefully, in the near future, a Z80
|
| 17 |
|
|
opcodeversion of the compiler as well as PC executable
|
| 18 |
|
|
versions of lasm and load will be finished. (RDK)
|
| 19 |
|
|
|
| 20 |
|
|
|
| 21 |
|
|
Available in the directory is a compiler for a
|
| 22 |
|
|
subset of the language C. It consists of the two files C80.C
|
| 23 |
|
|
(compiler) and C80LIB.I80 (runtime library) It is in source
|
| 24 |
|
|
form and is free to anyone wishing to use it.
|
| 25 |
|
|
Characteristics of the compiler are as follows:
|
| 26 |
|
|
|
| 27 |
|
|
(1) It supports a subset of the language C. (see the
|
| 28 |
|
|
book "C A Programming Language", by Brian Kernighan and
|
| 29 |
|
|
Dennis Ritchie.) (2) It is written in C itself. (3) It is
|
| 30 |
|
|
syntactically identical to the C on UNIX (unlike some other
|
| 31 |
|
|
small C compilers and interpreters). (4) It produces as
|
| 32 |
|
|
output a text file suitable for input to an 8080 assembler.
|
| 33 |
|
|
(5) It is a stand-alone single-pass compiler (which means it
|
| 34 |
|
|
does its own syntax checking and parsing and produces no
|
| 35 |
|
|
intermediate files). (6) It can compile itself. This means
|
| 36 |
|
|
any processor supporting C can be used to develop this small
|
| 37 |
|
|
C compiler for any other processor.
|
| 38 |
|
|
|
| 39 |
|
|
The intention behind the writing of this compiler was to
|
| 40 |
|
|
bring the C language to small computers. It was developed
|
| 41 |
|
|
primarily on a 8080 system with 40 K bytes and a single
|
| 42 |
|
|
mini-floppy. Consequently, an effort was made to keep the
|
| 43 |
|
|
compiler small in order to fit within limited memory, and
|
| 44 |
|
|
intermediate files were avoided in order to conserve floppy
|
| 45 |
|
|
space.
|
| 46 |
|
|
|
| 47 |
|
|
|
| 48 |
|
|
COMPILER SPECIFICATIONS
|
| 49 |
|
|
|
| 50 |
|
|
As of this writing, the compiler supports the following:
|
| 51 |
|
|
|
| 52 |
|
|
(1) Data type declarations can be:
|
| 53 |
|
|
|
| 54 |
|
|
- "char" (8 bits)
|
| 55 |
|
|
- "int" (16 bits)
|
| 56 |
|
|
- (by placing an "*" before the variable name, a pointer
|
| 57 |
|
|
can be formed to the respective type of
|
| 58 |
|
|
data element).
|
| 59 |
|
|
|
| 60 |
|
|
(2) Arrays:
|
| 61 |
|
|
|
| 62 |
|
|
- single dimension (vector) arrays can be
|
| 63 |
|
|
of type "char" or "int".
|
| 64 |
|
|
|
| 65 |
|
|
(3) Expressions:
|
| 66 |
|
|
|
| 67 |
|
|
- unary operators:
|
| 68 |
|
|
"-" (minus)
|
| 69 |
|
|
"*" (indirection)
|
| 70 |
|
|
"&" (address of)
|
| 71 |
|
|
"++" (increment, either prefix or postfix)
|
| 72 |
|
|
"--" (decrement, either prefix of postfix)
|
| 73 |
|
|
- binary operators:
|
| 74 |
|
|
"+" (addition)
|
| 75 |
|
|
"-" (subtraction)
|
| 76 |
|
|
"*" (multiplication)
|
| 77 |
|
|
"/" (division)
|
| 78 |
|
|
"%" (mod, i.e. remainder from division)
|
| 79 |
|
|
"|" (inclusive 'or')
|
| 80 |
|
|
"^" (exclusive 'or')
|
| 81 |
|
|
"&" (logical 'and')
|
| 82 |
|
|
"==" (test for equal)
|
| 83 |
|
|
"!=" (test for not equal)
|
| 84 |
|
|
"<" (test for less than)
|
| 85 |
|
|
"<=" (test for less than or equal to)
|
| 86 |
|
|
">" (test for greater than)
|
| 87 |
|
|
">=" (test for greater than or equal to)
|
| 88 |
|
|
"<<" (arithmetic left shift)
|
| 89 |
|
|
">>" (arithmetic right shift)
|
| 90 |
|
|
- primaries:
|
| 91 |
|
|
-array[expression]
|
| 92 |
|
|
-function(arg1, arg2,...,argn)
|
| 93 |
|
|
-constant
|
| 94 |
|
|
-decimal number
|
| 95 |
|
|
-quoted string ("sample string")
|
| 96 |
|
|
-primed string ('a' or 'Z' or 'ab')
|
| 97 |
|
|
-local variable (or pointer)
|
| 98 |
|
|
-global (static) variable (or pointer)
|
| 99 |
|
|
|
| 100 |
|
|
(4) Program control:
|
| 101 |
|
|
|
| 102 |
|
|
-if(expression)statement;
|
| 103 |
|
|
-if(expression) statement;
|
| 104 |
|
|
else statement;
|
| 105 |
|
|
-while (expression) statement;
|
| 106 |
|
|
-break;
|
| 107 |
|
|
-continue;
|
| 108 |
|
|
-return;
|
| 109 |
|
|
-return expression;
|
| 110 |
|
|
-; (null statement)
|
| 111 |
|
|
-{statement; statement; ... statement;}
|
| 112 |
|
|
(compound statement)
|
| 113 |
|
|
|
| 114 |
|
|
(5) Pointers:
|
| 115 |
|
|
|
| 116 |
|
|
-local and static pointers can contain the
|
| 117 |
|
|
address of "char" or "int" data elements.
|
| 118 |
|
|
|
| 119 |
|
|
(6) Compiler commands:
|
| 120 |
|
|
|
| 121 |
|
|
- #define name string (pre-processor will replace
|
| 122 |
|
|
name by string throughout text.)
|
| 123 |
|
|
- #include filename (allows program to include other
|
| 124 |
|
|
files within this compilation.)
|
| 125 |
|
|
- #asm (not supported by standard C)
|
| 126 |
|
|
Allows all code between "#asm" and "#endasm"
|
| 127 |
|
|
to be passed unchanged to the target
|
| 128 |
|
|
assembler. This command is actually a statement
|
| 129 |
|
|
and may appear in the context:
|
| 130 |
|
|
"if (expression) #asm...#endasm else..."
|
| 131 |
|
|
|
| 132 |
|
|
(7) Miscellaneous:
|
| 133 |
|
|
|
| 134 |
|
|
-Expression evaluation maintains the same hierarchy
|
| 135 |
|
|
as standard C.
|
| 136 |
|
|
|
| 137 |
|
|
-Function calls are defined as
|
| 138 |
|
|
any primary followed by an open paren, so legal forms
|
| 139 |
|
|
include:
|
| 140 |
|
|
|
| 141 |
|
|
variable();
|
| 142 |
|
|
array[expression]();
|
| 143 |
|
|
constant();
|
| 144 |
|
|
function()();
|
| 145 |
|
|
|
| 146 |
|
|
-Pointer arithmetic takes into account the data
|
| 147 |
|
|
type of the destination (e.g. pointer++ will increment
|
| 148 |
|
|
by two if pointer was declared "int *pointer").
|
| 149 |
|
|
|
| 150 |
|
|
-Pointer compares generated unsigned
|
| 151 |
|
|
compares (since addresses are not signed numbers).
|
| 152 |
|
|
|
| 153 |
|
|
-Often used pieces of code
|
| 154 |
|
|
(i.e. storing the primary register indirect through the
|
| 155 |
|
|
top of the stack) generate calls to library routines to
|
| 156 |
|
|
shorten the amount of code generated.
|
| 157 |
|
|
|
| 158 |
|
|
-Generated code is "pure" (i.e. the code may be placed
|
| 159 |
|
|
in Read Only Memory). Code, literals, and variables
|
| 160 |
|
|
are kept in separate sections of memory.
|
| 161 |
|
|
|
| 162 |
|
|
-The generated code is re-entrant. Everytime a function
|
| 163 |
|
|
is called, its local variables refer to a new stack
|
| 164 |
|
|
frame. By way of example, the compiler uses
|
| 165 |
|
|
recursive-descent for most of its parsing, which relies
|
| 166 |
|
|
heavily on re-entrant (recursive) functions.
|
| 167 |
|
|
|
| 168 |
|
|
|
| 169 |
|
|
COMPILER RESTRICTIONS
|
| 170 |
|
|
|
| 171 |
|
|
Since recent stages of compiler check-out have been done
|
| 172 |
|
|
both on an 8080 system and on UNIX, language syntax appears
|
| 173 |
|
|
to be identical (within the given subset) between this small
|
| 174 |
|
|
C compiler and the standard UNIX compiler.
|
| 175 |
|
|
|
| 176 |
|
|
|
| 177 |
|
|
Not supported yet are:
|
| 178 |
|
|
|
| 179 |
|
|
(1) Structures.
|
| 180 |
|
|
(2) Multi-dimensional arrays.
|
| 181 |
|
|
(3) Floating point, long integer, or unsigned data types.
|
| 182 |
|
|
(4) Function calls returning anything but "int".
|
| 183 |
|
|
(5) The unaries "!", "~", and "sizeof".
|
| 184 |
|
|
(6) The control binary operators "&&", "||", and "?:".
|
| 185 |
|
|
(7) The declaration specifiers "auto", "static", "extern",
|
| 186 |
|
|
and "register".
|
| 187 |
|
|
(8) The statements "for", "switch", "case",
|
| 188 |
|
|
and "default."
|
| 189 |
|
|
(9) The use of arguments within a "#define" command.
|
| 190 |
|
|
|
| 191 |
|
|
|
| 192 |
|
|
Compiler restrictions include:
|
| 193 |
|
|
|
| 194 |
|
|
(1) Since it is a single-pass compiler, undefined names
|
| 195 |
|
|
are not detected and are assumed to be function names not yet
|
| 196 |
|
|
defined. If this assumption is incorrect, the undefined
|
| 197 |
|
|
reference will not appear until the compiled program is
|
| 198 |
|
|
assembled.
|
| 199 |
|
|
|
| 200 |
|
|
(2) No optimizing is done. The code produced is sound
|
| 201 |
|
|
and capable of re-entrancy, but no attempt is made to
|
| 202 |
|
|
optimize either for code size or speed. It was assumed a
|
| 203 |
|
|
post-processor optimizer would later be written for the
|
| 204 |
|
|
target machine.
|
| 205 |
|
|
|
| 206 |
|
|
(3) Since the target assembler is of unknown
|
| 207 |
|
|
characteristics, no attempt is made to produce pseudo-ops to
|
| 208 |
|
|
declare static variables as internal or external.
|
| 209 |
|
|
|
| 210 |
|
|
(4) Constants are not evaluated by the compiler. That
|
| 211 |
|
|
is, the line of code:
|
| 212 |
|
|
|
| 213 |
|
|
X = 1+2;
|
| 214 |
|
|
|
| 215 |
|
|
would generated code to add "1" and "2" at runtime. The
|
| 216 |
|
|
results are correct, but unnecessary code is the penalty.
|
| 217 |
|
|
|
| 218 |
|
|
|
| 219 |
|
|
ASSEMBLY LANGUAGE INTERFACE
|
| 220 |
|
|
|
| 221 |
|
|
Interfacing to assembly language is relatively
|
| 222 |
|
|
straight-forward. The "#asm ... #endasm" construct allows
|
| 223 |
|
|
the user to place assembly language code directly into the
|
| 224 |
|
|
control context. Since it is considered by the compiler to
|
| 225 |
|
|
be a single statement, it may appear in such forms as:
|
| 226 |
|
|
|
| 227 |
|
|
while(1) #asm ... #endasm
|
| 228 |
|
|
|
| 229 |
|
|
or
|
| 230 |
|
|
|
| 231 |
|
|
if (expression) #asm...#endasm else...
|
| 232 |
|
|
|
| 233 |
|
|
Due to the workings of the preprocessor which must be
|
| 234 |
|
|
suppressed in this construct, the pseudo-op "#asm" must be
|
| 235 |
|
|
the last item before the carriage return on the end of the
|
| 236 |
|
|
line (i.e. the text between #asm and the is thrown
|
| 237 |
|
|
away), and the "#endasm" pseudo-op must appear on a line by
|
| 238 |
|
|
itself (i.e. everything after #endasm is also thrown away).
|
| 239 |
|
|
Since the parser is completely free-format outside of these
|
| 240 |
|
|
execeptions, the expected format is as follows:
|
| 241 |
|
|
|
| 242 |
|
|
if (expression) #asm
|
| 243 |
|
|
...
|
| 244 |
|
|
...
|
| 245 |
|
|
#endasm
|
| 246 |
|
|
else statement;
|
| 247 |
|
|
|
| 248 |
|
|
Note a semicolon is not required after the #endasm since
|
| 249 |
|
|
the end of context is obvious to the compiler. Assembly
|
| 250 |
|
|
language code within the "#asm ... #endasm" context has
|
| 251 |
|
|
access to all global symbols and functions by name. It is up
|
| 252 |
|
|
to the programmer to know the data type of the symbol
|
| 253 |
|
|
(whether "char" or "int" implies a byte access or a word
|
| 254 |
|
|
access). Stack locals and arguments may be retrieved by
|
| 255 |
|
|
offset (see STACK FRAME). External assembly language
|
| 256 |
|
|
routines invoked by function calls from the c-code have
|
| 257 |
|
|
access to all registers and do not have to restore them prior
|
| 258 |
|
|
to exit. They may push items on the stack as well, but must
|
| 259 |
|
|
pop them off before exit. It is the responsibility of the
|
| 260 |
|
|
calling program to remove arguments from the stack after a
|
| 261 |
|
|
function call. This must not be done by the function itself.
|
| 262 |
|
|
There is no limit to the number of bytes the function may
|
| 263 |
|
|
push onto the stack, providing they are removed prior to
|
| 264 |
|
|
returning. Since parameters are passed by value, the
|
| 265 |
|
|
paramters on the stack may be modified by the called program.
|
| 266 |
|
|
|
| 267 |
|
|
|
| 268 |
|
|
|
| 269 |
|
|
STACK FRAME
|
| 270 |
|
|
|
| 271 |
|
|
The stack is used extensively by the compiler. Function
|
| 272 |
|
|
arguments are pushed onto the stack as they are encountered
|
| 273 |
|
|
between parentheses (note, this is opposite that of standard
|
| 274 |
|
|
C, which means routines expressly retrieving arguments from
|
| 275 |
|
|
the stack rather than declaring them by name must beware).
|
| 276 |
|
|
By the definition of the language, parameter passing is "call
|
| 277 |
|
|
by value". For example the following code would be produced
|
| 278 |
|
|
for the C statement:
|
| 279 |
|
|
|
| 280 |
|
|
function(X, Y, z());
|
| 281 |
|
|
|
| 282 |
|
|
LHLD X
|
| 283 |
|
|
PUSH H
|
| 284 |
|
|
LHLD Y
|
| 285 |
|
|
PUSH H
|
| 286 |
|
|
CALL z
|
| 287 |
|
|
PUSH H
|
| 288 |
|
|
CALL function
|
| 289 |
|
|
POP B
|
| 290 |
|
|
POP B
|
| 291 |
|
|
POP B
|
| 292 |
|
|
|
| 293 |
|
|
Notice, the compiler cleans up the stack after the call
|
| 294 |
|
|
using a simple algorithm to use the least number of bytes.
|
| 295 |
|
|
|
| 296 |
|
|
Local variables allocate as much stack space as is
|
| 297 |
|
|
needed, and are then assigned the current value of the stack
|
| 298 |
|
|
pointer (after the allocation) as their address.
|
| 299 |
|
|
|
| 300 |
|
|
int X;
|
| 301 |
|
|
|
| 302 |
|
|
would produce:
|
| 303 |
|
|
|
| 304 |
|
|
PUSH B
|
| 305 |
|
|
|
| 306 |
|
|
which merely allocates room on the stack for 2 bytes (not
|
| 307 |
|
|
initialized to any value). References to the local variable
|
| 308 |
|
|
X will now be made to the stack pointer + 0. If another
|
| 309 |
|
|
declaration is made:
|
| 310 |
|
|
|
| 311 |
|
|
char array[3];
|
| 312 |
|
|
|
| 313 |
|
|
the code would be:
|
| 314 |
|
|
|
| 315 |
|
|
DCX SP
|
| 316 |
|
|
PUSH B
|
| 317 |
|
|
|
| 318 |
|
|
Array[0] would be at SP+0, array[1] would be at SP+1,
|
| 319 |
|
|
array[2] would be at SP+2, and X would now be at SP+3. Thus,
|
| 320 |
|
|
assembly language code using "#asm...#endasm" cannot access
|
| 321 |
|
|
local variables by name, but must know how many intervening
|
| 322 |
|
|
bytes have been allocated between the declaration of the
|
| 323 |
|
|
variable and its use. It is worth pointing out local
|
| 324 |
|
|
declarations allocate only as much stack space as is
|
| 325 |
|
|
required, including an odd number of bytes, whereas function
|
| 326 |
|
|
arguments always consist of two bytes apiece. In the event
|
| 327 |
|
|
the argument was type "char" (8 bits), the most significant
|
| 328 |
|
|
byte of the 2-byte value is a sign-extension of the lower
|
| 329 |
|
|
byte.
|
| 330 |
|
|
|
| 331 |
|
|
|
| 332 |
|
|
|
| 333 |
|
|
OPERATING THE COMPILER
|
| 334 |
|
|
|
| 335 |
|
|
The small C compiler begins by asking the user for a
|
| 336 |
|
|
number of options regarding the expected compilation. Since
|
| 337 |
|
|
it was easier to ask questions than to pull arguments from a
|
| 338 |
|
|
command line (which is in no way similar between the 8080
|
| 339 |
|
|
developmental system and UNIX), this was the preferred
|
| 340 |
|
|
method.
|
| 341 |
|
|
|
| 342 |
|
|
The questions asked are as follows:
|
| 343 |
|
|
|
| 344 |
|
|
Do you want the c-text to appear?
|
| 345 |
|
|
|
| 346 |
|
|
This gives the user the option of interleaving the
|
| 347 |
|
|
source code into the output file. Response is Y or N. If Y,
|
| 348 |
|
|
a semicolon will be placed at the start of each input line
|
| 349 |
|
|
(to force a comment to the 8080 assembler) and the input
|
| 350 |
|
|
lines will be printed where appropriate. If the answer is N,
|
| 351 |
|
|
only the generated 8080 code will be output.
|
| 352 |
|
|
|
| 353 |
|
|
Do you wish the globals to be defined?
|
| 354 |
|
|
|
| 355 |
|
|
This question is primarily a developmental aid between
|
| 356 |
|
|
machines. If the answer is Y, all static symbols will
|
| 357 |
|
|
allocate storage within the module being compiled. This is
|
| 358 |
|
|
the normal method. If N, no storage will be allocated, but
|
| 359 |
|
|
symbol references will still be made in the normal way.
|
| 360 |
|
|
Essentially, this question allows the user to specify all or
|
| 361 |
|
|
none of the static symbols external. It is to be considered
|
| 362 |
|
|
a temporary measure.
|
| 363 |
|
|
|
| 364 |
|
|
Starting number for labels?
|
| 365 |
|
|
|
| 366 |
|
|
This lets the user supply the first label number
|
| 367 |
|
|
generated by the compiler for it internal labels (which will
|
| 368 |
|
|
typically be "ccXXXXX", where XXXXX is a decimal number
|
| 369 |
|
|
increasing with each label). This option allows modules to
|
| 370 |
|
|
be compiled separately and later appended on the source level
|
| 371 |
|
|
without generating multi-defined labels.
|
| 372 |
|
|
|
| 373 |
|
|
Output filename?
|
| 374 |
|
|
|
| 375 |
|
|
This question gets from the user the name of the file to
|
| 376 |
|
|
be created. A null line sends output to the user's terminal.
|
| 377 |
|
|
|
| 378 |
|
|
Input filename?
|
| 379 |
|
|
|
| 380 |
|
|
This question gets from the user the name of the C
|
| 381 |
|
|
module to use as input. The question will be repeated each
|
| 382 |
|
|
time a name is supplied, allowing the user to create an
|
| 383 |
|
|
output file consisting of many separate input files (it
|
| 384 |
|
|
behaves as if the user had appended them together and
|
| 385 |
|
|
submitted only the one file). A null line response ends the
|
| 386 |
|
|
compilation process.
|
| 387 |
|
|
|
| 388 |
|
|
|
| 389 |
|
|
COMPILING THE COMPILER
|
| 390 |
|
|
|
| 391 |
|
|
The power of the compiler lies in the fact it can
|
| 392 |
|
|
compile itself. This allows a user to "bootstrap" the
|
| 393 |
|
|
compiler onto a new machine without excessive recoding.
|
| 394 |
|
|
|
| 395 |
|
|
To compile the compiler under the UNIX operating system,
|
| 396 |
|
|
the appropriate command is:
|
| 397 |
|
|
|
| 398 |
|
|
% cc C80.c -lS
|
| 399 |
|
|
|
| 400 |
|
|
which will invoke the UNIX C-compiler and the UNIX linker to
|
| 401 |
|
|
create the runnable file "a.out". This file may be renamed
|
| 402 |
|
|
as needed and used. No other files are needed.
|
| 403 |
|
|
|
| 404 |
|
|
In order to create a compiler for a new machine, the
|
| 405 |
|
|
user will need to compile the compiler into the language of
|
| 406 |
|
|
the destination processor. The procedure currently used to
|
| 407 |
|
|
create the compiler for my 8080 system is as follows:
|
| 408 |
|
|
|
| 409 |
|
|
(1) Edit the file C80.c to modify two lines of code:
|
| 410 |
|
|
|
| 411 |
|
|
-change the line of code
|
| 412 |
|
|
|
| 413 |
|
|
#include
|
| 414 |
|
|
to
|
| 415 |
|
|
#define NULL 0
|
| 416 |
|
|
|
| 417 |
|
|
(this is done since the "stdio.h" I/O header file
|
| 418 |
|
|
contains unparsable lines for the small compiler, and the
|
| 419 |
|
|
line defining NULL is the only line of "stdio.h" needed by
|
| 420 |
|
|
the compiler).
|
| 421 |
|
|
|
| 422 |
|
|
-change the line of code
|
| 423 |
|
|
|
| 424 |
|
|
#define eol 10
|
| 425 |
|
|
to
|
| 426 |
|
|
#define eol 13
|
| 427 |
|
|
|
| 428 |
|
|
(this is done since my 8080 system uses for the end
|
| 429 |
|
|
of line character, and UNIX uses the "newline" character).
|
| 430 |
|
|
|
| 431 |
|
|
|
| 432 |
|
|
(2) Invoke the compiler (by typing "a.out" or whatever other
|
| 433 |
|
|
name it was given.
|
| 434 |
|
|
|
| 435 |
|
|
(3) Answer the questions by the compiler to use the file
|
| 436 |
|
|
C80.c as input and to produce the file C80.I80
|
| 437 |
|
|
as output.
|
| 438 |
|
|
|
| 439 |
|
|
(4) Append the files C80.I80 and C80LIB.I80 (the code for the
|
| 440 |
|
|
compiler and the code for the runtime library,
|
| 441 |
|
|
respectively).
|
| 442 |
|
|
|
| 443 |
|
|
(5) Assemble the combined file using some 8080 assembler.
|
| 444 |
|
|
|
| 445 |
|
|
(6) Execute the created run file.
|
| 446 |
|
|
|
| 447 |
|
|
Currently, the 8080 assembler used must possess the
|
| 448 |
|
|
abilities to handle symbol names unique to 8 characters and
|
| 449 |
|
|
to recognize lower-case symbol names as unique from their
|
| 450 |
|
|
upper-case equivalent. This is due to the fact the compiler
|
| 451 |
|
|
recognizes 8-character names and passes all static variable
|
| 452 |
|
|
and function names intact to the assembler. There are a few
|
| 453 |
|
|
symbol names within the compiler which are not unique until
|
| 454 |
|
|
the 7th character and which have "upper-case twins". These
|
| 455 |
|
|
discourage the use of the KL-10's MACN80 since it folds
|
| 456 |
|
|
lower-case to upper case and does not recognize 8-character
|
| 457 |
|
|
names. It may be used, however, if the user is aware of
|
| 458 |
|
|
these limitations and chooses symbol names within these
|
| 459 |
|
|
restrictions.
|
| 460 |
|
|
|
| 461 |
|
|
|
| 462 |
|
|
THE FUTURE OF THE COMPILER
|
| 463 |
|
|
|
| 464 |
|
|
That part of the compiler which produces code for the
|
| 465 |
|
|
8080 is all together in the final section of the compiler.
|
| 466 |
|
|
Routines used by the compiler to produce code are kept short
|
| 467 |
|
|
and are commented. Changing this compiler to produce code
|
| 468 |
|
|
for any other machine is a matter of changing only these few
|
| 469 |
|
|
routines, and does not entail digging around through the
|
| 470 |
|
|
internals of the program. I would expect the change to
|
| 471 |
|
|
another machine could be made in an afternoon providing the
|
| 472 |
|
|
target machine had the following attributes:
|
| 473 |
|
|
|
| 474 |
|
|
(1) A stack, preferably running backwards as items
|
| 475 |
|
|
are pushed onto it.
|
| 476 |
|
|
|
| 477 |
|
|
(2) Two sixteen-bit registers. In the 8080 these
|
| 478 |
|
|
are the HL register pair (the primary register
|
| 479 |
|
|
to the compiler) and the DE register pair (the
|
| 480 |
|
|
secondary register).
|
| 481 |
|
|
|
| 482 |
|
|
(3) An assembler (or cross-assembler).
|
| 483 |
|
|
|
| 484 |
|
|
|
| 485 |
|
|
Since the compiler is just now on its feet and subject
|
| 486 |
|
|
to feedback from users, it is expected many changes will be
|
| 487 |
|
|
made to it. Already planned changes (in order of expected
|
| 488 |
|
|
addition) are:
|
| 489 |
|
|
|
| 490 |
|
|
(1) Constants will be pre-evaluated by the
|
| 491 |
|
|
compiler. Something like x=1+2*3 will become
|
| 492 |
|
|
x=7 prior to generating any code.
|
| 493 |
|
|
|
| 494 |
|
|
(2) Structures will be added. This is one of the
|
| 495 |
|
|
powers of C. Its omission has always been
|
| 496 |
|
|
considered temporary.
|
| 497 |
|
|
|
| 498 |
|
|
(3) Assignment operators (+=, &=, etc.) will be
|
| 499 |
|
|
added.
|
| 500 |
|
|
|
| 501 |
|
|
(4) Missing unary and binary operators and
|
| 502 |
|
|
statements will be added.
|
| 503 |
|
|
|
| 504 |
|
|
(5) The expression parser will create intermediate
|
| 505 |
|
|
tree-structures of the expressions and will
|
| 506 |
|
|
walk through them before generating any code.
|
| 507 |
|
|
This will allow some optimization and will
|
| 508 |
|
|
allow the function arguments to be passed on
|
| 509 |
|
|
the stack in the same sequence as UNIX.
|
| 510 |
|
|
|
| 511 |
|
|
(6) A peep-hole optimizer will be added to improve
|
| 512 |
|
|
the generated code.
|
| 513 |
|
|
|
| 514 |
|
|
Many of these things represent a wish-list. Time will
|
| 515 |
|
|
be spent only when it becomes available. Any volunteer help
|
| 516 |
|
|
in any of these areas would be appreciated.
|
| 517 |
|
|
|
| 518 |
|
|
Questions should be directed to Ron Cain here at SRI
|
| 519 |
|
|
either at extension 3860 or at CAIN@SRI-KL.
|
| 520 |
|
|
|
| 521 |
|
|
|
| 522 |
|
|
|
| 523 |
|
|
|
| 524 |
|
|
|