1 /* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
3 * This file is part of the LibreOffice project.
5 * This Source Code Form is subject to the terms of the Mozilla Public
6 * License, v. 2.0. If a copy of the MPL was not distributed with this
7 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
9 * This file incorporates work covered by the following license notice:
11 * Licensed to the Apache Software Foundation (ASF) under one or more
12 * contributor license agreements. See the NOTICE file distributed
13 * with this work for additional information regarding copyright
14 * ownership. The ASF licenses this file to you under the Apache
15 * License, Version 2.0 (the "License"); you may not use this file
16 * except in compliance with the License. You may obtain a copy of
17 * the License at http://www.apache.org/licenses/LICENSE-2.0 .
20 #include <sal/types.h>
27 /*ER evaluate macros to pDefOut */
30 * skipnl() skips over input text to the end of the line.
31 * skipws() skips over "whitespace" (spaces or tabs), but
32 * not skip over the end of the line. It skips over
33 * TOK_SEP, however (though that shouldn't happen).
34 * scanid() reads the next token (C identifier) into token[].
35 * The caller has already read the first character of
36 * the identifier. Unlike macroid(), the token is
38 * macroid() reads the next token (C identifier) into token[].
39 * If it is a #defined macro, it is expanded, and
40 * macroid() returns TRUE, otherwise, FALSE.
41 * catenate() Does the dirty work of token concatenation, TRUE if it did.
42 * scanstring() Reads a string from the input stream, calling
43 * a user-supplied function for each character.
44 * This function may be output() to write the
45 * string to the output file, or save() to save
46 * the string in the work buffer.
47 * scannumber() Reads a C numeric constant from the input stream,
48 * calling the user-supplied function for each
49 * character. (output() or save() as noted above.)
50 * save() Save one character in the work[] buffer.
51 * savestring() Saves a string in malloc() memory.
52 * getfile() Initialize a new FILEINFO structure, called when
53 * #include opens a new file, or a macro is to be
55 * getmem() Get a specified number of bytes from malloc memory.
56 * output() Write one character to stdout (calling PUTCHAR) --
57 * implemented as a function so its address may be
58 * passed to scanstring() and scannumber().
59 * lookid() Scans the next token (identifier) from the input
60 * stream. Looks for it in the #defined symbol table.
61 * Returns a pointer to the definition, if found, or NULL
62 * if not present. The identifier is stored in token[].
63 * defnedel() Define enter/delete subroutine. Updates the
65 * get() Read the next byte from the current input stream,
66 * handling end of (macro/file) input and embedded
67 * comments appropriately. Note that the global
68 * instring is -- essentially -- a parameter to get().
69 * cget() Like get(), but skip over TOK_SEP.
70 * unget() Push last gotten character back on the input stream.
71 * cerror(), cwarn(), cfatal(), cierror(), ciwarn()
72 * These routines format an print messages to the user.
73 * cerror & cwarn take a format and a single string argument.
74 * cierror & ciwarn take a format and a single int (char) argument.
75 * cfatal takes a format and a single string argument.
79 * This table must be rewritten for a non-Ascii machine.
81 * Note that several "non-visible" characters have special meaning:
82 * Hex 1D DEF_MAGIC -- a flag to prevent #define recursion.
83 * Hex 1E TOK_SEP -- a delimiter for token concatenation
84 * Hex 1F COM_SEP -- a zero-width whitespace for comment concatenation
86 #if TOK_SEP != 0x1E || COM_SEP != 0x1F || DEF_MAGIC != 0x1D
87 << error type table is
not correct
>>
98 char type
[256] = { /* Character type codes Hex */
99 END
, 000, 000, 000, 000, SPA
, 000, 000, /* 00 */
100 000, 000, 000, 000, 000, 000, 000, 000, /* 08 */
101 000, 000, 000, 000, 000, 000, 000, 000, /* 10 */
102 000, 000, 000, 000, 000, LET
, 000, SPA
, /* 18 */
103 000, 000, 000, 000, 000, 000, 000, 000, /* 20 */
104 000, 000, 000, 000, 000, 000, 000, 000, /* 28 */
105 000, 000, 000, 000, 000, 000, 000, 000, /* 30 */
106 000, 000, 000, 000, 000, 000, 000, 000, /* 38 */
107 SPA
, 000, 000, 000, 000, 000, 000, 000, /* 40 */
108 000, 000, 000, DOT
, OP_LT
,OP_LPA
,OP_ADD
, OP_OR
, /* 48 .<(+| */
109 OP_AND
, 000, 000, 000, 000, 000, 000, 000, /* 50 & */
110 000, 000,OP_NOT
, DOL
,OP_MUL
,OP_RPA
, 000,OP_XOR
, /* 58 !$*);^ */
111 OP_SUB
,OP_DIV
, 000, 000, 000, 000, 000, 000, /* 60 -/ */
112 000, 000, 000, 000,OP_MOD
, LET
, OP_GT
,OP_QUE
, /* 68 ,%_>? */
113 000, 000, 000, 000, 000, 000, 000, 000, /* 70 */
114 000, 000,OP_COL
, 000, 000, QUO
, OP_EQ
, QUO
, /* 78 `:#@'=" */
115 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 80 abcdefg */
116 LET
, LET
, 000, 000, 000, 000, 000, 000, /* 88 hi */
117 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 90 jklmnop */
118 LET
, LET
, 000, 000, 000, 000, 000, 000, /* 98 qr */
119 000,OP_NOT
, LET
, LET
, LET
, LET
, LET
, LET
, /* A0 ~stuvwx */
120 LET
, LET
, 000, 000, 000, 000, 000, 000, /* A8 yz [ */
121 000, 000, 000, 000, 000, 000, 000, 000, /* B0 */
122 000, 000, 000, 000, 000, 000, 000, 000, /* B8 ] */
123 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* C0 {ABCDEFG */
124 LET
, LET
, 000, 000, 000, 000, 000, 000, /* C8 HI */
125 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* D0 }JKLMNOP */
126 LET
, LET
, 000, 000, 000, 000, 000, 000, /* D8 QR */
127 BSH
, 000, LET
, LET
, LET
, LET
, LET
, LET
, /* E0 \ STUVWX */
128 LET
, LET
, 000, 000, 000, 000, 000, 000, /* E8 YZ */
129 DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, /* F0 01234567 */
130 DIG
, DIG
, 000, 000, 000, 000, 000, 000, /* F8 89 */
135 char type
[256] = { /* Character type codes Hex */
136 END
, 000, 000, 000, 000, 000, 000, 000, /* 00 */
137 000, SPA
, 000, 000, 000, 000, 000, 000, /* 08 */
138 000, 000, 000, 000, 000, 000, 000, 000, /* 10 */
139 000, 000, 000, 000, 000, LET
, 000, SPA
, /* 18 */
140 SPA
,OP_NOT
, QUO
, 000, DOL
,OP_MOD
,OP_AND
, QUO
, /* 20 !"#$%&' */
141 OP_LPA
,OP_RPA
,OP_MUL
,OP_ADD
, 000,OP_SUB
, DOT
,OP_DIV
, /* 28 ()*+,-./ */
142 DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, DIG
, /* 30 01234567 */
143 DIG
, DIG
,OP_COL
, 000, OP_LT
, OP_EQ
, OP_GT
,OP_QUE
, /* 38 89:;<=>? */
144 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 40 @ABCDEFG */
145 LET
, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 48 HIJKLMNO */
146 LET
, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 50 PQRSTUVW */
147 LET
, LET
, LET
, 000, BSH
, 000,OP_XOR
, LET
, /* 58 XYZ[\]^_ */
148 000, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 60 `abcdefg */
149 LET
, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 68 hijklmno */
150 LET
, LET
, LET
, LET
, LET
, LET
, LET
, LET
, /* 70 pqrstuvw */
151 LET
, LET
, LET
, 000, OP_OR
, 000,OP_NOT
, 000, /* 78 xyz{|}~ */
152 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
153 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
154 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
155 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
156 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
157 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
158 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
159 000, 000, 000, 000, 000, 000, 000, 000, /* 80 .. FF */
165 * C P P S y m b o l T a b l e s
169 * SBSIZE defines the number of hash-table slots for the symbol table.
170 * It must be a power of 2.
175 #define SBMASK (SBSIZE - 1)
176 #if (SBSIZE ^ SBMASK) != ((SBSIZE * 2) - 1)
177 << error
, SBSIZE must be a power of
2 >>
181 static DEFBUF
*symtab
[SBSIZE
]; /* Symbol table queue headers */
186 for( i
= 0; i
< SBSIZE
; i
++ )
194 * Skip to the end of the current input line.
199 do { /* Skip to newline */
201 } while (c
!= '\n' && c
!= EOF_CHAR
);
207 * Skip over whitespace
212 do { /* Skip whitespace */
214 #if COMMENT_INVISIBLE
215 } while (type
[c
] == SPA
|| c
== COM_SEP
);
217 } while (type
[c
] == SPA
);
224 * Get the next token (an id) into the token buffer.
225 * Note: this code is duplicated in lookid().
226 * Change one, change both.
231 if (c
== DEF_MAGIC
) /* Eat the magic token */
232 c
= get(); /* undefiner. */
235 if (bp
< &token
[IDMAX
]) /* token dim is IDMAX+1 */
238 } while (type
[c
] == LET
|| type
[c
] == DIG
);
246 * If c is a letter, scan the id. if it's #defined, expand it and scan
247 * the next character and try again.
249 * Else, return the character. If type[c] is a LET, the token is in token.
254 if (infile
!= NULL
&& infile
->fp
!= NULL
)
256 while (type
[c
] == LET
&& (dp
= lookid(c
)) != NULL
) {
266 * A token was just read (via macroid).
267 * If the next character is TOK_SEP, concatenate the next token
268 * return TRUE -- which should recall macroid after refreshing
269 * macroid's argument. If it is not TOK_SEP, unget() the character
274 register char *token1
;
277 if (get() != TOK_SEP
) { /* Token concatenation */
282 token1
= savestring(token
); /* Save first token */
283 c
= macroid(get()); /* Scan next token */
284 switch(type
[c
]) { /* What was it? */
285 case LET
: /* An identifier, ... */
286 if (strlen(token1
) + strlen(token
) >= NWORK
)
287 cfatal("work buffer overflow doing %s #", token1
);
288 sprintf(work
, "%s%s", token1
, token
);
291 case DIG
: /* A digit string */
292 strcpy(work
, token1
);
293 workp
= work
+ strlen(work
);
296 } while ((c
= get()) != TOK_SEP
);
298 * The trailing TOK_SEP is no longer needed.
303 default: /* An error, ... */
304 #if ! COMMENT_INVISIBLE
306 cierror("Strange character '%c' after #", c
);
308 cierror("Strange character (%d.) after #", c
);
310 strcpy(work
, token1
);
315 * work has the concatenated token and token1 has
316 * the first token (no longer needed). Unget the
317 * new (concatenated) token after freeing token1.
318 * Finally, setup to read the new token.
320 free(token1
); /* Free up memory */
321 ungetstring(work
); /* Unget the new thing, */
325 return (FALSE
); /* Not supported */
330 scanstring(int delim
,
332 void (*outfun
)( int ) /* BP */ /* Output function */
334 void (*outfun
)() /* BP */
338 * Scan off a string. Warning if terminated by newline or EOF.
339 * outfun() outputs the character -- to a buffer if in a macro.
340 * TRUE if ok, FALSE if error.
345 instring
= TRUE
; /* Don't strip comments */
347 while ((c
= get()) != delim
362 cerror("Unterminated string", NULLST
);
368 void scannumber(int c
,
370 register void (*outfun
)( int ) /* BP */ /* Output/store func */
372 register void (*outfun
)() /* BP */
376 * Process a number. We know that c is from 0 to 9 or dot.
377 * Algorithm from Dave Conroy's Decus C.
380 register int radix
; /* 8, 10, or 16 */
381 int expseen
; /* 'e' seen in floater */
382 int signseen
; /* '+' or '-' seen */
383 int octal89
; /* For bad octal test */
384 int dotflag
; /* TRUE if '.' was seen */
386 expseen
= FALSE
; /* No exponent seen yet */
387 signseen
= TRUE
; /* No +/- allowed yet */
388 octal89
= FALSE
; /* No bad octal yet */
389 radix
= 10; /* Assume decimal */
390 if ((dotflag
= (c
== '.')) != FALSE
) { /* . something? */
391 (*outfun
)('.'); /* Always out the dot */
392 if (type
[(c
= get())] != DIG
) { /* If not a float numb, */
393 unget(); /* Rescan strange char */
394 return; /* All done for now */
396 } /* End of float test */
397 else if (c
== '0') { /* Octal or hex? */
398 (*outfun
)(c
); /* Stuff initial zero */
399 radix
= 8; /* Assume it's octal */
400 c
= get(); /* Look for an 'x' */
401 if (c
== 'x' || c
== 'X') { /* Did we get one? */
402 radix
= 16; /* Remember new radix */
403 (*outfun
)(c
); /* Stuff the 'x' */
404 c
= get(); /* Get next character */
407 for (;;) { /* Process curr. char. */
409 * Note that this algorithm accepts "012e4" and "03.4"
410 * as legitimate floating-point numbers.
412 if (radix
!= 16 && (c
== 'e' || c
== 'E')) {
413 if (expseen
) /* Already saw 'E'? */
414 break; /* Exit loop, bad nbr. */
415 expseen
= TRUE
; /* Set exponent seen */
416 signseen
= FALSE
; /* We can read '+' now */
417 radix
= 10; /* Decimal exponent */
419 else if (radix
!= 16 && c
== '.') {
420 if (dotflag
) /* Saw dot already? */
421 break; /* Exit loop, two dots */
422 dotflag
= TRUE
; /* Remember the dot */
423 radix
= 10; /* Decimal fraction */
425 else if (c
== '+' || c
== '-') { /* 1.0e+10 */
426 if (signseen
) /* Sign in wrong place? */
427 break; /* Exit loop, not nbr. */
428 /* signseen = TRUE; */ /* Remember we saw it */
430 else { /* Check the digit */
432 case '8': case '9': /* Sometimes wrong */
433 octal89
= TRUE
; /* Do check later */
434 case '0': case '1': case '2': case '3':
435 case '4': case '5': case '6': case '7':
436 break; /* Always ok */
438 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f':
439 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F':
440 if (radix
== 16) /* Alpha's are ok only */
441 break; /* if reading hex. */
442 default: /* At number end */
443 goto done
; /* Break from for loop */
444 } /* End of switch */
445 } /* End general case */
446 (*outfun
)(c
); /* Accept the character */
447 signseen
= TRUE
; /* Don't read sign now */
448 c
= get(); /* Read another char */
449 } /* End of scan loop */
451 * When we break out of the scan loop, c contains the first
452 * character (maybe) not in the number. If the number is an
453 * integer, allow a trailing 'L' for long and/or a trailing 'U'
454 * for unsigned. If not those, push the trailing character back
455 * on the input stream. Floating point numbers accept a trailing
456 * 'L' for "long double".
458 done
: if (dotflag
|| expseen
) { /* Floating point? */
459 if (c
== 'l' || c
== 'L') {
461 get(); /* Ungotten later */
464 else { /* Else it's an integer */
466 * We know that dotflag and expseen are both zero, now:
467 * dotflag signals "saw 'L'", and
468 * expseen signals "saw 'U'".
489 (*outfun
)(c
); /* Got 'L' or 'U'. */
490 c
= get(); /* Look at next, too. */
493 nomore
: unget(); /* Not part of a number */
494 if (octal89
&& radix
== 8)
495 cwarn("Illegal digit in octal number", NULLST
);
500 if (workp
>= &work
[NWORK
]) {
501 work
[NWORK
-1] = '\0';
502 cfatal("Work buffer overflow: %s", work
);
504 else *workp
++ = (char)c
;
508 savestring(char* text
)
510 * Store a string into free memory.
513 register char *result
;
515 result
= getmem(strlen(text
) + 1);
516 strcpy(result
, text
);
521 getfile(int bufsize
, char* name
)
523 * Common FILEINFO buffer initialization for a new file or macro.
526 register FILEINFO
*file
;
529 size
= strlen(name
); /* File/macro name */
530 file
= (FILEINFO
*) getmem(sizeof (FILEINFO
) + bufsize
+ size
);
531 file
->parent
= infile
; /* Chain files together */
532 file
->fp
= NULL
; /* No file yet */
533 file
->filename
= savestring(name
); /* Save file/macro name */
534 file
->progname
= NULL
; /* No #line seen yet */
535 file
->unrecur
= 0; /* No macro fixup */
536 file
->bptr
= file
->buffer
; /* Initialize line ptr */
537 file
->buffer
[0] = EOS
; /* Force first read */
538 file
->line
= 0; /* (Not used just yet) */
539 if (infile
!= NULL
) /* If #include file */
540 infile
->line
= line
; /* Save current line */
541 infile
= file
; /* New current file */
542 line
= 1; /* Note first line */
543 return (file
); /* All done. */
549 * Get a block of free memory.
552 register char *result
;
554 if ((result
= malloc((unsigned) size
)) == NULL
)
555 cfatal("Out of memory", NULLST
);
562 * Look for the next token in the symbol table. Returns token in "token".
563 * If found, returns the table pointer; Else returns NULL.
570 int isrecurse
; /* For #define foo foo */
574 if (0 != (isrecurse
= (c
== DEF_MAGIC
))) /* If recursive macro */
575 c
= get(); /* hack, skip DEF_MAGIC */
577 if (np
< &token
[IDMAX
]) { /* token dim is IDMAX+1 */
578 *np
++ = (char)c
; /* Store token byte */
579 nhash
+= c
; /* Update hash value */
581 c
= get(); /* And get another byte */
582 } while (type
[c
] == LET
|| type
[c
] == DIG
);
583 unget(); /* Rescan terminator */
584 *np
= EOS
; /* Terminate token */
585 if (isrecurse
) /* Recursive definition */
586 return (NULL
); /* undefined just now */
587 nhash
+= (np
- token
); /* Fix hash value */
588 dp
= symtab
[nhash
& SBMASK
]; /* Starting bucket */
589 while (dp
!= (DEFBUF
*) NULL
) { /* Search symbol table */
590 if (dp
->hash
== nhash
/* Fast precheck */
591 && (temp
= strcmp(dp
->name
, token
)) >= 0)
593 dp
= dp
->link
; /* Nope, try next one */
595 return ((temp
== 0) ? dp
: NULL
);
599 defendel(char* name
, int delete)
601 * Enter this name in the lookup table (delete = FALSE)
602 * or delete this name (delete = TRUE).
603 * Returns a pointer to the define block (delete = FALSE)
604 * Returns NULL if the symbol wasn't defined (delete = TRUE).
608 register DEFBUF
**prevp
;
614 for (nhash
= 0, np
= name
; *np
!= EOS
;)
618 prevp
= &symtab
[nhash
& SBMASK
];
619 while ((dp
= *prevp
) != (DEFBUF
*) NULL
) {
620 if (dp
->hash
== nhash
621 && (temp
= strcmp(dp
->name
, name
)) >= 0) {
623 dp
= NULL
; /* Not found */
625 *prevp
= dp
->link
; /* Found, unlink and */
626 if (dp
->repl
!= NULL
) /* Free the replacement */
627 free(dp
->repl
); /* if any, and then */
628 free((char *) dp
); /* Free the symbol */
635 dp
= (DEFBUF
*) getmem(sizeof (DEFBUF
) + size
);
641 strcpy(dp
->name
, name
);
646 #if OSL_DEBUG_LEVEL > 1
648 void dumpdef(char *why
)
651 register DEFBUF
**syp
;
652 FILE *pRememberOut
= NULL
;
654 if ( bDumpDefs
) /*ER */
656 pRememberOut
= pCppOut
;
659 fprintf( pCppOut
, "CPP symbol table dump %s\n", why
);
660 for (syp
= symtab
; syp
< &symtab
[SBSIZE
]; syp
++) {
661 if ((dp
= *syp
) != (DEFBUF
*) NULL
) {
662 fprintf( pCppOut
, "symtab[%" SAL_PRI_PTRDIFFT
"d]\n", (syp
- symtab
));
664 dumpadef((char *) NULL
, dp
);
665 } while ((dp
= dp
->link
) != (DEFBUF
*) NULL
);
670 fprintf( pCppOut
, "\n");
671 pCppOut
= pRememberOut
;
675 void dumpadef(char *why
, register DEFBUF
*dp
)
679 FILE *pRememberOut
= NULL
;
681 /*ER dump #define's to pDefOut */
684 pRememberOut
= pCppOut
;
687 fprintf( pCppOut
, " \"%s\" [%d]", dp
->name
, dp
->nargs
);
689 fprintf( pCppOut
, " (%s)", why
);
690 if (dp
->repl
!= NULL
) {
691 fprintf( pCppOut
, " => ");
692 for (cp
= dp
->repl
; (c
= *cp
++ & 0xFF) != EOS
;) {
696 if( c
== EOS
) break;
697 fprintf( pCppOut
, "<%%%d>", c
- MAC_PARM
);
700 if (c
>= MAC_PARM
&& c
<= (MAC_PARM
+ PAR_MAC
))
701 fprintf( pCppOut
, "<%%%d>", c
- MAC_PARM
);
703 else if (isprint(c
) || c
== '\n' || c
== '\t')
706 fprintf( pCppOut
, "<^%c>", c
+ '@');
708 fprintf( pCppOut
, "<\\0%o>", c
);
710 /*ER evaluate macros to pDefOut */
712 if ( bDumpDefs
&& !bIsInEval
&& dp
->nargs
<= 0 )
714 FILEINFO
*infileSave
= infile
;
715 char *tokenSave
= savestring( token
);
716 char *workSave
= savestring( work
);
718 int wronglineSave
= wrongline
;
719 int recursionSave
= recursion
;
724 infile
= NULL
; /* start from scrap */
730 file
= getfile( strlen( dp
->repl
), dp
->name
);
731 strcpy( file
->buffer
, dp
->repl
);
732 fprintf( pCppOut
, " ===> ");
734 cppmain(); /* get() frees also *file */
736 if ( 0 == evaluate( EvalBuf
, &valEval
) )
739 if ( valEval
!= (EVALTYPE
)((long)valEval
) )
740 fprintf( pCppOut
, " ==eval=> %f", valEval
);
743 fprintf( pCppOut
, " ==eval=> %ld", (long)valEval
);
745 recursion
= recursionSave
;
746 wrongline
= wronglineSave
;
748 strcpy( work
, workSave
);
750 strcpy( token
, tokenSave
);
758 fprintf( pCppOut
, ", no replacement.");
762 pCppOut
= pRememberOut
;
773 * Return the next character from a macro or the current file.
774 * Handle end of file from #include files.
778 register FILEINFO
*file
;
779 register int popped
; /* Recursion fixup */
783 if ((file
= infile
) == NULL
)
788 * Read a character from the current input line or macro.
789 * At EOS, either finish the current macro (freeing temp.
790 * storage) or read another line from the current input file.
791 * At EOF, exit the current file (#include) or, at EOF from
792 * the cpp input file, return EOF_CHAR to finish processing.
794 if ((c
= *file
->bptr
++ & 0xFF) == EOS
) {
796 * Nothing in current line or macro. Get next line (if
797 * input from a file), or do end of file/macro processing.
798 * In the latter case, jump back to restart from the top.
800 if (file
->fp
== NULL
) { /* NULL if macro */
802 recursion
-= file
->unrecur
;
805 infile
= file
->parent
; /* Unwind file chain */
807 else { /* Else get from a file */
808 if ((file
->bptr
= fgets(file
->buffer
, NBUFF
, file
->fp
))
810 #if OSL_DEBUG_LEVEL > 1
811 if (debug
> 1) { /* Dump it to stdout */
812 fprintf( pCppOut
, "\n#line %d (%s), %s",
813 line
, file
->filename
, file
->buffer
);
816 goto newline
; /* process the line */
819 if( file
->fp
!= stdin
)
820 fclose(file
->fp
); /* Close finished file */
821 if ((infile
= file
->parent
) != NULL
) {
823 * There is an "ungotten" newline in the current
824 * infile buffer (set there by doinclude() in
825 * cpp1.c). Thus, we know that the mainline code
826 * is skipping over blank lines and will do a
827 * #line at its convenience.
829 wrongline
= TRUE
; /* Need a #line now */
834 * Free up space used by the (finished) file or macro and
835 * restart input from the parent file/macro, if any.
837 free(file
->filename
); /* Free name and */
838 if (file
->progname
!= NULL
) /* if a #line was seen, */
839 free(file
->progname
); /* free it, too. */
840 free((char *) file
); /* Free file space */
841 if (infile
== NULL
) /* If at end of file */
842 return (EOF_CHAR
); /* Return end of file */
843 line
= infile
->line
; /* Reset line number */
844 goto get_from_file
; /* Get from the top. */
847 * Common processing for the new character.
849 if (c
== DEF_MAGIC
&& file
->fp
!= NULL
) /* Don't allow delete */
850 goto newline
; /* from a file */
851 if (file
->parent
!= NULL
) { /* Macro or #include */
853 file
->parent
->unrecur
+= popped
;
855 recursion
-= file
->parent
->unrecur
;
858 file
->parent
->unrecur
= 0;
861 #if (HOST == SYS_UNIX)
862 /*ER*/ if (c
== '\r')
863 /*ER*/ return get(); /* DOS fuck */
865 if (c
== '\n') /* Maintain current */
866 ++line
; /* line counter */
867 if (instring
) /* Strings just return */
868 return (c
); /* the character. */
869 else if (c
== '/') { /* Comment? */
870 instring
= TRUE
; /* So get() won't loop */
873 /*MM*/ if ((c
!= '*') && (c
!= '/')) { /* Next byte '*'? */
874 instring
= FALSE
; /* Nope, no comment */
875 unget(); /* Push the char. back */
876 return ('/'); /* Return the slash */
878 if (keepcomments
) { /* If writing comments */
879 PUTCHAR('/'); /* Write out the */
881 /*MM*/ if( '*' == c
)
887 /*MM*/ if( '*' == c
){
888 for (;;) { /* Eat a comment */
890 test
: if (keepcomments
&& c
!= EOF_CHAR
)
894 cerror("EOF in comment", NULLST
);
898 if ((c
= get()) != '*') /* Don't let comments */
899 goto test
; /* Nest. */
900 #ifdef STRICT_COMMENTS
901 cwarn("Nested comments", NULLST
);
903 /* Fall into * stuff */
905 if ((c
= get()) != '/') /* If comment doesn't */
906 goto test
; /* end, look at next */
907 instring
= FALSE
; /* End of comment, */
908 if (keepcomments
) { /* Put out the comment */
909 cput(c
); /* terminator, too */
912 * A comment is syntactically "whitespace" --
913 * however, there are certain strange sequences
915 * #define foo(x) (something)
916 * foo|* comment *|(123)
918 * where just returning space (or COM_SEP) will cause
919 * problems. This can be "fixed" by overwriting the
920 * '/' in the input line buffer with ' ' (or COM_SEP)
921 * but that may mess up an error message.
922 * So, we peek ahead -- if the next character is
923 * "whitespace" we just get another character, if not,
924 * we modify the buffer. All in the name of purity.
926 if (*file
->bptr
== '\n'
927 || type
[*file
->bptr
& 0xFF] == SPA
)
929 #if COMMENT_INVISIBLE
931 * Return magic (old-fashioned) syntactic space.
933 return ((file
->bptr
[-1] = COM_SEP
));
935 return ((file
->bptr
[-1] = ' '));
938 case '\n': /* we'll need a #line */
940 wrongline
= TRUE
; /* later... */
941 default: /* Anything else is */
942 break; /* Just a character */
944 } /* End comment loop */
946 else{ /* c++ comment */
948 for (;;) { /* Eat a comment */
950 if (keepcomments
&& c
!= EOF_CHAR
)
954 else if( '\n' == c
){
955 instring
= FALSE
; /* End of comment, */
960 } /* End if in comment */
961 else if (!inmacro
&& c
== '\\') { /* If backslash, peek */
962 if ((c
= get()) == '\n') { /* for a <nl>. If so, */
966 else { /* Backslash anything */
967 unget(); /* Get it later */
968 return ('\\'); /* Return the backslash */
971 else if (c
== '\f' || c
== VT
) /* Form Feed, Vertical */
972 c
= ' '; /* Tab are whitespace */
973 else if (c
== 0xef) /* eat up UTF-8 BOM */
975 if((c
= get()) == 0xbb)
977 if((c
= get()) == 0xbf)
995 return (c
); /* Just return the char */
1000 * Backup the pointer to reread the last character. Fatal error
1001 * (code bug) if we backup too far. unget() may be called,
1002 * without problems, at end of file. Only one character may
1003 * be ungotten. If you need to unget more, call ungetstring().
1006 register FILEINFO
*file
;
1008 if ((file
= infile
) == NULL
)
1009 return; /* Unget after EOF */
1010 if (--file
->bptr
< file
->buffer
)
1011 cfatal("Too much pushback", NULLST
);
1012 if (*file
->bptr
== '\n') /* Ungetting a newline? */
1013 --line
; /* Unget the line number, too */
1016 void ungetstring(char* text
)
1018 * Push a string back on the input stream. This is done by treating
1019 * the text as if it were a macro.
1022 register FILEINFO
*file
;
1023 extern FILEINFO
*getfile();
1024 file
= getfile(strlen(text
) + 1, "");
1025 strcpy(file
->buffer
, text
);
1031 * Get one character, absorb "funny space" after comments or
1032 * token concatenation
1039 #if COMMENT_INVISIBLE
1040 } while (c
== TOK_SEP
|| c
== COM_SEP
);
1042 } while (c
== TOK_SEP
);
1048 * Error messages and other hacks. The first byte of severity
1049 * is 'S' for string arguments and 'I' for int arguments. This
1050 * is needed for portability with machines that have int's that
1051 * are shorter than char *'s.
1054 static void domsg(char* severity
, char* format
, void* arg
)
1056 * Print filenames, macro names, and line numbers for error messages.
1060 register FILEINFO
*file
;
1062 fprintf(stderr
, "%sline %d, %s: ", MSG_PREFIX
, line
, &severity
[1]);
1063 if (*severity
== 'S')
1064 fprintf(stderr
, format
, (char *)arg
);
1066 fprintf(stderr
, format
, *((int *)arg
) );
1068 if ((file
= infile
) == NULL
)
1069 return; /* At end of file */
1070 if (file
->fp
!= NULL
) {
1071 tp
= file
->buffer
; /* Print current file */
1072 fprintf(stderr
, "%s", tp
); /* name, making sure */
1073 if (tp
[strlen(tp
) - 1] != '\n') /* there's a newline */
1076 while ((file
= file
->parent
) != NULL
) { /* Print #includes, too */
1077 if (file
->fp
== NULL
)
1078 fprintf(stderr
, "from macro %s\n", file
->filename
);
1081 fprintf(stderr
, "from file %s, line %d:\n%s",
1082 (file
->progname
!= NULL
)
1083 ? file
->progname
: file
->filename
,
1085 if (tp
[strlen(tp
) - 1] != '\n')
1091 void cerror(char* format
, char* sarg
)
1093 * Print a normal error message, string argument.
1096 domsg("SError", format
, sarg
);
1100 void cierror(char* format
, int narg
)
1102 * Print a normal error message, numeric argument.
1105 domsg("IError", format
, &narg
);
1109 void cfatal(char* format
, char* sarg
)
1114 domsg("SFatal error", format
, sarg
);
1118 void cwarn(char* format
, char* sarg
)
1120 * A non-fatal error, string argument.
1123 domsg("SWarning", format
, sarg
);
1126 void ciwarn(char* format
, int narg
)
1128 * A non-fatal error, numeric argument.
1131 domsg("IWarning", format
, &narg
);
1134 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */