version 1.15, 2012/03/04 04:05:15 |
version 1.16, 2014/04/18 21:11:34 |
|
|
/* $OpenBSD$ */ |
|
/* |
/* |
* Copyright (c) 2002, 2003 Tony Finch <dot@dotat.at> |
* Copyright (c) 2002 - 2014 Tony Finch <dot@dotat.at> |
* Copyright (c) 1985, 1993 |
|
* The Regents of the University of California. All rights reserved. |
|
* |
* |
* This code is derived from software contributed to Berkeley by |
|
* Dave Yost. Support for #if and #elif was added by Tony Finch. |
|
* |
|
* Redistribution and use in source and binary forms, with or without |
* Redistribution and use in source and binary forms, with or without |
* modification, are permitted provided that the following conditions |
* modification, are permitted provided that the following conditions |
* are met: |
* are met: |
|
|
* 2. Redistributions in binary form must reproduce the above copyright |
* 2. Redistributions in binary form must reproduce the above copyright |
* notice, this list of conditions and the following disclaimer in the |
* notice, this list of conditions and the following disclaimer in the |
* documentation and/or other materials provided with the distribution. |
* documentation and/or other materials provided with the distribution. |
* 3. Neither the name of the University nor the names of its contributors |
|
* may be used to endorse or promote products derived from this software |
|
* without specific prior written permission. |
|
* |
* |
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND |
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND |
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE |
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE |
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
|
|
/* |
/* |
* unifdef - remove ifdef'ed lines |
* unifdef - remove ifdef'ed lines |
* |
* |
|
* This code was derived from software contributed to Berkeley by Dave Yost. |
|
* It was rewritten to support ANSI C by Tony Finch. The original version |
|
* of unifdef carried the 4-clause BSD copyright licence. None of its code |
|
* remains in this version (though some of the names remain) so it now |
|
* carries a more liberal licence. |
|
* |
* Wishlist: |
* Wishlist: |
* provide an option which will append the name of the |
* provide an option which will append the name of the |
* appropriate symbol after #else's and #endif's |
* appropriate symbol after #else's and #endif's |
* provide an option which will check symbols after |
* provide an option which will check symbols after |
* #else's and #endif's to see that they match their |
* #else's and #endif's to see that they match their |
* corresponding #ifdef or #ifndef |
* corresponding #ifdef or #ifndef |
* generate #line directives in place of deleted code |
|
* |
* |
* The first two items above require better buffer handling, which would |
* These require better buffer handling, which would also make |
* also make it possible to handle all "dodgy" directives correctly. |
* it possible to handle all "dodgy" directives correctly. |
*/ |
*/ |
|
|
#include <ctype.h> |
#include "unifdef.h" |
#include <err.h> |
|
#include <stdarg.h> |
|
#include <stdbool.h> |
|
#include <stdio.h> |
|
#include <stdlib.h> |
|
#include <string.h> |
|
#include <unistd.h> |
|
|
|
|
static const char copyright[] = |
|
#include "version.h" |
|
"@(#) $Author$\n" |
|
"@(#) $URL: http://dotat.at/prog/unifdef $\n" |
|
; |
|
|
/* types of input lines: */ |
/* types of input lines: */ |
typedef enum { |
typedef enum { |
LT_TRUEI, /* a true #if with ignore flag */ |
LT_TRUEI, /* a true #if with ignore flag */ |
|
|
LT_DODGY_LAST = LT_DODGY + LT_ENDIF, |
LT_DODGY_LAST = LT_DODGY + LT_ENDIF, |
LT_PLAIN, /* ordinary line */ |
LT_PLAIN, /* ordinary line */ |
LT_EOF, /* end of file */ |
LT_EOF, /* end of file */ |
|
LT_ERROR, /* unevaluable #if */ |
LT_COUNT |
LT_COUNT |
} Linetype; |
} Linetype; |
|
|
|
|
"DODGY IF", "DODGY TRUE", "DODGY FALSE", |
"DODGY IF", "DODGY TRUE", "DODGY FALSE", |
"DODGY ELIF", "DODGY ELTRUE", "DODGY ELFALSE", |
"DODGY ELIF", "DODGY ELTRUE", "DODGY ELFALSE", |
"DODGY ELSE", "DODGY ENDIF", |
"DODGY ELSE", "DODGY ENDIF", |
"PLAIN", "EOF" |
"PLAIN", "EOF", "ERROR" |
}; |
}; |
|
|
|
#define linetype_if2elif(lt) ((Linetype)(lt - LT_IF + LT_ELIF)) |
|
#define linetype_2dodgy(lt) ((Linetype)(lt + LT_DODGY)) |
|
|
/* state of #if processing */ |
/* state of #if processing */ |
typedef enum { |
typedef enum { |
IS_OUTSIDE, |
IS_OUTSIDE, |
|
|
C_COMMENT, /* in a comment like this one */ |
C_COMMENT, /* in a comment like this one */ |
CXX_COMMENT, /* between // and end of line */ |
CXX_COMMENT, /* between // and end of line */ |
STARTING_COMMENT, /* just after slash-backslash-newline */ |
STARTING_COMMENT, /* just after slash-backslash-newline */ |
FINISHING_COMMENT /* star-backslash-newline in a C comment */ |
FINISHING_COMMENT, /* star-backslash-newline in a C comment */ |
|
CHAR_LITERAL, /* inside '' */ |
|
STRING_LITERAL /* inside "" */ |
} Comment_state; |
} Comment_state; |
|
|
static char const * const comment_name[] = { |
static char const * const comment_name[] = { |
"NO", "C", "CXX", "STARTING", "FINISHING" |
"NO", "C", "CXX", "STARTING", "FINISHING", "CHAR", "STRING" |
}; |
}; |
|
|
/* state of preprocessor line parser */ |
/* state of preprocessor line parser */ |
|
|
*/ |
*/ |
#define MAXDEPTH 64 /* maximum #if nesting */ |
#define MAXDEPTH 64 /* maximum #if nesting */ |
#define MAXLINE 4096 /* maximum length of line */ |
#define MAXLINE 4096 /* maximum length of line */ |
#define MAXSYMS 4096 /* maximum number of symbols */ |
#define MAXSYMS 16384 /* maximum number of symbols */ |
|
|
/* |
/* |
* Sometimes when editing a keyword the replacement text is longer, so |
* Sometimes when editing a keyword the replacement text is longer, so |
|
|
* Globals. |
* Globals. |
*/ |
*/ |
|
|
|
static bool compblank; /* -B: compress blank lines */ |
|
static bool lnblank; /* -b: blank deleted lines */ |
static bool complement; /* -c: do the complement */ |
static bool complement; /* -c: do the complement */ |
static bool debugging; /* -d: debugging reports */ |
static bool debugging; /* -d: debugging reports */ |
|
static bool inplace; /* -m: modify in place */ |
static bool iocccok; /* -e: fewer IOCCC errors */ |
static bool iocccok; /* -e: fewer IOCCC errors */ |
|
static bool strictlogic; /* -K: keep ambiguous #ifs */ |
static bool killconsts; /* -k: eval constant #ifs */ |
static bool killconsts; /* -k: eval constant #ifs */ |
static bool lnblank; /* -l: blank deleted lines */ |
static bool lnnum; /* -n: add #line directives */ |
static bool symlist; /* -s: output symbol list */ |
static bool symlist; /* -s: output symbol list */ |
|
static bool symdepth; /* -S: output symbol depth */ |
static bool text; /* -t: this is a text file */ |
static bool text; /* -t: this is a text file */ |
|
|
static const char *symname[MAXSYMS]; /* symbol name */ |
static const char *symname[MAXSYMS]; /* symbol name */ |
|
|
static FILE *input; /* input file pointer */ |
static FILE *input; /* input file pointer */ |
static const char *filename; /* input file name */ |
static const char *filename; /* input file name */ |
static int linenum; /* current line number */ |
static int linenum; /* current line number */ |
|
static const char *linefile; /* file name for #line */ |
|
static FILE *output; /* output file pointer */ |
|
static const char *ofilename; /* output file name */ |
|
static const char *backext; /* backup extension */ |
|
static char *tempname; /* avoid splatting input */ |
|
|
static char tline[MAXLINE+EDITSLOP];/* input buffer plus space */ |
static char tline[MAXLINE+EDITSLOP];/* input buffer plus space */ |
static char *keyword; /* used for editing #elif's */ |
static char *keyword; /* used for editing #elif's */ |
|
|
|
/* |
|
* When processing a file, the output's newline style will match the |
|
* input's, and unifdef correctly handles CRLF or LF endings whatever |
|
* the platform's native style. The stdio streams are opened in binary |
|
* mode to accommodate platforms whose native newline style is CRLF. |
|
* When the output isn't a processed input file (when it is error / |
|
* debug / diagnostic messages) then unifdef uses native line endings. |
|
*/ |
|
|
|
static const char *newline; /* input file format */ |
|
static const char newline_unix[] = "\n"; |
|
static const char newline_crlf[] = "\r\n"; |
|
|
static Comment_state incomment; /* comment parser state */ |
static Comment_state incomment; /* comment parser state */ |
static Line_state linestate; /* #if line parser state */ |
static Line_state linestate; /* #if line parser state */ |
static Ifstate ifstate[MAXDEPTH]; /* #if processor state */ |
static Ifstate ifstate[MAXDEPTH]; /* #if processor state */ |
static bool ignoring[MAXDEPTH]; /* ignore comments state */ |
static bool ignoring[MAXDEPTH]; /* ignore comments state */ |
static int stifline[MAXDEPTH]; /* start of current #if */ |
static int stifline[MAXDEPTH]; /* start of current #if */ |
static int depth; /* current #if nesting */ |
static int depth; /* current #if nesting */ |
static bool keepthis; /* don't delete constant #if */ |
static int delcount; /* count of deleted lines */ |
|
static unsigned blankcount; /* count of blank lines */ |
|
static unsigned blankmax; /* maximum recent blankcount */ |
|
static bool constexpr; /* constant #if expression */ |
|
static bool zerosyms; /* to format symdepth output */ |
|
static bool firstsym; /* ditto */ |
|
|
|
static int exitmode; /* exit status mode */ |
static int exitstat; /* program exit status */ |
static int exitstat; /* program exit status */ |
|
|
static void addsym(bool, bool, char *); |
static void addsym1(bool, bool, char *); |
|
static void addsym2(bool, const char *, const char *); |
|
static char *astrcat(const char *, const char *); |
|
static void cleantemp(void); |
|
static void closeio(void); |
static void debug(const char *, ...); |
static void debug(const char *, ...); |
|
static void debugsym(const char *, int); |
|
static bool defundef(void); |
|
static void defundefile(const char *); |
|
static void done(void); |
static void error(const char *); |
static void error(const char *); |
static int findsym(const char *); |
static int findsym(const char **); |
static void flushline(bool); |
static void flushline(bool); |
static Linetype get_line(void); |
static void hashline(void); |
|
static void help(void); |
static Linetype ifeval(const char **); |
static Linetype ifeval(const char **); |
static void ignoreoff(void); |
static void ignoreoff(void); |
static void ignoreon(void); |
static void ignoreon(void); |
|
static void indirectsym(void); |
static void keywordedit(const char *); |
static void keywordedit(const char *); |
|
static const char *matchsym(const char *, const char *); |
static void nest(void); |
static void nest(void); |
|
static Linetype parseline(void); |
static void process(void); |
static void process(void); |
|
static void processinout(const char *, const char *); |
|
static const char *skipargs(const char *); |
static const char *skipcomment(const char *); |
static const char *skipcomment(const char *); |
|
static const char *skiphash(void); |
|
static const char *skipline(const char *); |
static const char *skipsym(const char *); |
static const char *skipsym(const char *); |
static void state(Ifstate); |
static void state(Ifstate); |
static int strlcmp(const char *, const char *, size_t); |
static void unnest(void); |
static void usage(void); |
static void usage(void); |
|
static void version(void); |
|
static const char *xstrdup(const char *, const char *); |
|
|
#define endsym(c) (!isalpha((unsigned char)c) && !isdigit((unsigned char)c) && c != '_') |
#define endsym(c) (!isalnum((unsigned char)c) && c != '_') |
|
|
/* |
/* |
* The main program. |
* The main program. |
|
|
{ |
{ |
int opt; |
int opt; |
|
|
while ((opt = getopt(argc, argv, "i:D:U:I:cdeklst")) != -1) |
while ((opt = getopt(argc, argv, "i:D:U:f:I:M:o:x:bBcdehKklmnsStV")) != -1) |
switch (opt) { |
switch (opt) { |
case 'i': /* treat stuff controlled by these symbols as text */ |
case 'i': /* treat stuff controlled by these symbols as text */ |
/* |
/* |
|
|
*/ |
*/ |
opt = *optarg++; |
opt = *optarg++; |
if (opt == 'D') |
if (opt == 'D') |
addsym(true, true, optarg); |
addsym1(true, true, optarg); |
else if (opt == 'U') |
else if (opt == 'U') |
addsym(true, false, optarg); |
addsym1(true, false, optarg); |
else |
else |
usage(); |
usage(); |
break; |
break; |
case 'D': /* define a symbol */ |
case 'D': /* define a symbol */ |
addsym(false, true, optarg); |
addsym1(false, true, optarg); |
break; |
break; |
case 'U': /* undef a symbol */ |
case 'U': /* undef a symbol */ |
addsym(false, false, optarg); |
addsym1(false, false, optarg); |
break; |
break; |
case 'I': |
case 'I': /* no-op for compatibility with cpp */ |
/* no-op for compatibility with cpp */ |
|
break; |
break; |
|
case 'b': /* blank deleted lines instead of omitting them */ |
|
case 'l': /* backwards compatibility */ |
|
lnblank = true; |
|
break; |
|
case 'B': /* compress blank lines around removed section */ |
|
compblank = true; |
|
break; |
case 'c': /* treat -D as -U and vice versa */ |
case 'c': /* treat -D as -U and vice versa */ |
complement = true; |
complement = true; |
break; |
break; |
|
|
case 'e': /* fewer errors from dodgy lines */ |
case 'e': /* fewer errors from dodgy lines */ |
iocccok = true; |
iocccok = true; |
break; |
break; |
|
case 'f': /* definitions file */ |
|
defundefile(optarg); |
|
break; |
|
case 'h': |
|
help(); |
|
break; |
|
case 'K': /* keep ambiguous #ifs */ |
|
strictlogic = true; |
|
break; |
case 'k': /* process constant #ifs */ |
case 'k': /* process constant #ifs */ |
killconsts = true; |
killconsts = true; |
break; |
break; |
case 'l': /* blank deleted lines instead of omitting them */ |
case 'm': /* modify in place */ |
lnblank = true; |
inplace = true; |
break; |
break; |
|
case 'M': /* modify in place and keep backup */ |
|
inplace = true; |
|
backext = optarg; |
|
break; |
|
case 'n': /* add #line directive after deleted lines */ |
|
lnnum = true; |
|
break; |
|
case 'o': /* output to a file */ |
|
ofilename = optarg; |
|
break; |
case 's': /* only output list of symbols that control #ifs */ |
case 's': /* only output list of symbols that control #ifs */ |
symlist = true; |
symlist = true; |
break; |
break; |
|
case 'S': /* list symbols with their nesting depth */ |
|
symlist = symdepth = true; |
|
break; |
case 't': /* don't parse C comments */ |
case 't': /* don't parse C comments */ |
text = true; |
text = true; |
break; |
break; |
|
case 'V': |
|
version(); |
|
break; |
|
case 'x': |
|
exitmode = atoi(optarg); |
|
if(exitmode < 0 || exitmode > 2) |
|
usage(); |
|
break; |
default: |
default: |
usage(); |
usage(); |
} |
} |
argc -= optind; |
argc -= optind; |
argv += optind; |
argv += optind; |
if (nsyms == 0 && !symlist) { |
if (compblank && lnblank) |
warnx("must -D or -U at least one symbol"); |
errx(2, "-B and -b are mutually exclusive"); |
usage(); |
if (symlist && (ofilename != NULL || inplace || argc > 1)) |
|
errx(2, "-s only works with one input file"); |
|
if (argc > 1 && ofilename != NULL) |
|
errx(2, "-o cannot be used with multiple input files"); |
|
if (argc > 1 && !inplace) |
|
errx(2, "multiple input files require -m or -M"); |
|
if (argc == 0) |
|
argc = 1; |
|
if (argc == 1 && !inplace && ofilename == NULL) |
|
ofilename = "-"; |
|
indirectsym(); |
|
|
|
atexit(cleantemp); |
|
if (ofilename != NULL) |
|
processinout(*argv, ofilename); |
|
else while (argc-- > 0) { |
|
processinout(*argv, *argv); |
|
argv++; |
} |
} |
if (argc > 1) { |
switch(exitmode) { |
errx(2, "can only do one file"); |
case(0): exit(exitstat); |
} else if (argc == 1 && strcmp(*argv, "-") != 0) { |
case(1): exit(!exitstat); |
filename = *argv; |
case(2): exit(0); |
if ((input = fopen(filename, "r")) != NULL) { |
default: abort(); /* bug */ |
process(); |
} |
(void) fclose(input); |
} |
} else |
|
err(2, "can't open %s", *argv); |
/* |
} else { |
* File logistics. |
|
*/ |
|
static void |
|
processinout(const char *ifn, const char *ofn) |
|
{ |
|
struct stat st; |
|
|
|
if (ifn == NULL || strcmp(ifn, "-") == 0) { |
filename = "[stdin]"; |
filename = "[stdin]"; |
input = stdin; |
linefile = NULL; |
|
input = fbinmode(stdin); |
|
} else { |
|
filename = ifn; |
|
linefile = ifn; |
|
input = fopen(ifn, "rb"); |
|
if (input == NULL) |
|
err(2, "can't open %s", ifn); |
|
} |
|
if (strcmp(ofn, "-") == 0) { |
|
output = fbinmode(stdout); |
process(); |
process(); |
|
return; |
} |
} |
|
if (stat(ofn, &st) < 0) { |
|
output = fopen(ofn, "wb"); |
|
if (output == NULL) |
|
err(2, "can't create %s", ofn); |
|
process(); |
|
return; |
|
} |
|
|
exit(exitstat); |
tempname = astrcat(ofn, ".XXXXXX"); |
|
output = mktempmode(tempname, st.st_mode); |
|
if (output == NULL) |
|
err(2, "can't create %s", tempname); |
|
|
|
process(); |
|
|
|
if (backext != NULL) { |
|
char *backname = astrcat(ofn, backext); |
|
if (rename(ofn, backname) < 0) |
|
err(2, "can't rename \"%s\" to \"%s\"", ofn, backname); |
|
free(backname); |
|
} |
|
if (replace(tempname, ofn) < 0) |
|
err(2, "can't rename \"%s\" to \"%s\"", tempname, ofn); |
|
free(tempname); |
|
tempname = NULL; |
} |
} |
|
|
|
/* |
|
* For cleaning up if there is an error. |
|
*/ |
static void |
static void |
|
cleantemp(void) |
|
{ |
|
if (tempname != NULL) |
|
remove(tempname); |
|
} |
|
|
|
/* |
|
* Self-identification functions. |
|
*/ |
|
|
|
static void |
|
version(void) |
|
{ |
|
const char *c = copyright; |
|
for (;;) { |
|
while (*++c != '$') |
|
if (*c == '\0') |
|
exit(0); |
|
while (*++c != '$') |
|
putc(*c, stderr); |
|
putc('\n', stderr); |
|
} |
|
} |
|
|
|
static void |
|
synopsis(FILE *fp) |
|
{ |
|
fprintf(fp, |
|
"usage: unifdef [-bBcdehKkmnsStV] [-x{012}] [-Mext] [-opath] \\\n" |
|
" [-[i]Dsym[=val]] [-[i]Usym] [-fpath] ... [file] ...\n"); |
|
} |
|
|
|
static void |
usage(void) |
usage(void) |
{ |
{ |
fprintf(stderr, |
synopsis(stderr); |
"usage: unifdef [-ceklst] [-Dsym[=val]] [-Ipath] [-iDsym[=val]] " |
|
"[-iUsym] [-Usym]\n" |
|
"\t[file]\n"); |
|
exit(2); |
exit(2); |
} |
} |
|
|
|
static void |
|
help(void) |
|
{ |
|
synopsis(stdout); |
|
printf( |
|
" -Dsym=val define preprocessor symbol with given value\n" |
|
" -Dsym define preprocessor symbol with value 1\n" |
|
" -Usym preprocessor symbol is undefined\n" |
|
" -iDsym=val \\ ignore C strings and comments\n" |
|
" -iDsym ) in sections controlled by these\n" |
|
" -iUsym / preprocessor symbols\n" |
|
" -fpath file containing #define and #undef directives\n" |
|
" -b blank lines instead of deleting them\n" |
|
" -B compress blank lines around deleted section\n" |
|
" -c complement (invert) keep vs. delete\n" |
|
" -d debugging mode\n" |
|
" -e ignore multiline preprocessor directives\n" |
|
" -h print help\n" |
|
" -Ipath extra include file path (ignored)\n" |
|
" -K disable && and || short-circuiting\n" |
|
" -k process constant #if expressions\n" |
|
" -Mext modify in place and keep backups\n" |
|
" -m modify input files in place\n" |
|
" -n add #line directives to output\n" |
|
" -opath output file name\n" |
|
" -S list #if control symbols with nesting\n" |
|
" -s list #if control symbols\n" |
|
" -t ignore C strings and comments\n" |
|
" -V print version\n" |
|
" -x{012} exit status mode\n" |
|
); |
|
exit(0); |
|
} |
|
|
/* |
/* |
* A state transition function alters the global #if processing state |
* A state transition function alters the global #if processing state |
* in a particular way. The table below is indexed by the current |
* in a particular way. The table below is indexed by the current |
* processing state and the type of the current line. A NULL entry |
* processing state and the type of the current line. |
* indicates that processing is complete. |
|
* |
* |
* Nesting is handled by keeping a stack of states; some transition |
* Nesting is handled by keeping a stack of states; some transition |
* functions increase or decrease the depth. They also maintain the |
* functions increase or decrease the depth. They also maintain the |
|
|
* When we have processed a group that starts off with a known-false |
* When we have processed a group that starts off with a known-false |
* #if/#elif sequence (which has therefore been deleted) followed by a |
* #if/#elif sequence (which has therefore been deleted) followed by a |
* #elif that we don't understand and therefore must keep, we edit the |
* #elif that we don't understand and therefore must keep, we edit the |
* latter into a #if to keep the nesting correct. |
* latter into a #if to keep the nesting correct. We use memcpy() to |
|
* overwrite the 4 byte token "elif" with "if " without a '\0' byte. |
* |
* |
* When we find a true #elif in a group, the following block will |
* When we find a true #elif in a group, the following block will |
* always be kept and the rest of the sequence after the next #elif or |
* always be kept and the rest of the sequence after the next #elif or |
|
|
typedef void state_fn(void); |
typedef void state_fn(void); |
|
|
/* report an error */ |
/* report an error */ |
static void |
static void Eelif (void) { error("Inappropriate #elif"); } |
Eelif(void) |
static void Eelse (void) { error("Inappropriate #else"); } |
{ |
static void Eendif(void) { error("Inappropriate #endif"); } |
error("Inappropriate #elif"); |
static void Eeof (void) { error("Premature EOF"); } |
} |
static void Eioccc(void) { error("Obfuscated preprocessor control line"); } |
|
|
static void |
|
Eelse(void) |
|
{ |
|
error("Inappropriate #else"); |
|
} |
|
|
|
static void |
|
Eendif(void) |
|
{ |
|
error("Inappropriate #endif"); |
|
} |
|
|
|
static void |
|
Eeof(void) |
|
{ |
|
error("Premature EOF"); |
|
} |
|
|
|
static void |
|
Eioccc(void) |
|
{ |
|
error("Obfuscated preprocessor control line"); |
|
} |
|
|
|
/* plain line handling */ |
/* plain line handling */ |
static void |
static void print (void) { flushline(true); } |
print(void) |
static void drop (void) { flushline(false); } |
{ |
|
flushline(true); |
|
} |
|
|
|
static void |
|
drop(void) |
|
{ |
|
flushline(false); |
|
} |
|
|
|
/* output lacks group's start line */ |
/* output lacks group's start line */ |
static void |
static void Strue (void) { drop(); ignoreoff(); state(IS_TRUE_PREFIX); } |
Strue(void) |
static void Sfalse(void) { drop(); ignoreoff(); state(IS_FALSE_PREFIX); } |
{ |
static void Selse (void) { drop(); state(IS_TRUE_ELSE); } |
drop(); |
|
ignoreoff(); |
|
state(IS_TRUE_PREFIX); |
|
} |
|
|
|
static void |
|
Sfalse(void) |
|
{ |
|
drop(); |
|
ignoreoff(); |
|
state(IS_FALSE_PREFIX); |
|
} |
|
|
|
static void |
|
Selse(void) |
|
{ |
|
drop(); |
|
state(IS_TRUE_ELSE); |
|
} |
|
|
|
/* print/pass this block */ |
/* print/pass this block */ |
static void |
static void Pelif (void) { print(); ignoreoff(); state(IS_PASS_MIDDLE); } |
Pelif(void) |
static void Pelse (void) { print(); state(IS_PASS_ELSE); } |
{ |
static void Pendif(void) { print(); unnest(); } |
print(); |
|
ignoreoff(); |
|
state(IS_PASS_MIDDLE); |
|
} |
|
|
|
static void |
|
Pelse(void) |
|
{ |
|
print(); |
|
state(IS_PASS_ELSE); |
|
} |
|
|
|
static void |
|
Pendif(void) |
|
{ |
|
print(); |
|
--depth; |
|
} |
|
|
|
/* discard this block */ |
/* discard this block */ |
static void |
static void Dfalse(void) { drop(); ignoreoff(); state(IS_FALSE_TRAILER); } |
Dfalse(void) |
static void Delif (void) { drop(); ignoreoff(); state(IS_FALSE_MIDDLE); } |
{ |
static void Delse (void) { drop(); state(IS_FALSE_ELSE); } |
drop(); |
static void Dendif(void) { drop(); unnest(); } |
ignoreoff(); |
|
state(IS_FALSE_TRAILER); |
|
} |
|
|
|
static void |
|
Delif(void) |
|
{ |
|
drop(); |
|
ignoreoff(); |
|
state(IS_FALSE_MIDDLE); |
|
} |
|
|
|
static void |
|
Delse(void) |
|
{ |
|
drop(); |
|
state(IS_FALSE_ELSE); |
|
} |
|
|
|
static void |
|
Dendif(void) |
|
{ |
|
drop(); |
|
--depth; |
|
} |
|
|
|
/* first line of group */ |
/* first line of group */ |
static void |
static void Fdrop (void) { nest(); Dfalse(); } |
Fdrop(void) |
static void Fpass (void) { nest(); Pelif(); } |
{ |
static void Ftrue (void) { nest(); Strue(); } |
nest(); |
static void Ffalse(void) { nest(); Sfalse(); } |
Dfalse(); |
|
} |
|
|
|
static void |
|
Fpass(void) |
|
{ |
|
nest(); |
|
Pelif(); |
|
} |
|
|
|
static void |
|
Ftrue(void) |
|
{ |
|
nest(); |
|
Strue(); |
|
} |
|
|
|
static void |
|
Ffalse(void) |
|
{ |
|
nest(); |
|
Sfalse(); |
|
} |
|
|
|
/* variable pedantry for obfuscated lines */ |
/* variable pedantry for obfuscated lines */ |
static void |
static void Oiffy (void) { if (!iocccok) Eioccc(); Fpass(); ignoreon(); } |
Oiffy(void) |
static void Oif (void) { if (!iocccok) Eioccc(); Fpass(); } |
{ |
static void Oelif (void) { if (!iocccok) Eioccc(); Pelif(); } |
if (iocccok) |
|
Fpass(); |
|
else |
|
Eioccc(); |
|
ignoreon(); |
|
} |
|
|
|
static void |
|
Oif(void) |
|
{ |
|
if (iocccok) |
|
Fpass(); |
|
else |
|
Eioccc(); |
|
} |
|
|
|
static void |
|
Oelif(void) |
|
{ |
|
if (iocccok) |
|
Pelif(); |
|
else |
|
Eioccc(); |
|
} |
|
|
|
/* ignore comments in this block */ |
/* ignore comments in this block */ |
static void |
static void Idrop (void) { Fdrop(); ignoreon(); } |
Idrop(void) |
static void Itrue (void) { Ftrue(); ignoreon(); } |
{ |
static void Ifalse(void) { Ffalse(); ignoreon(); } |
Fdrop(); |
/* modify this line */ |
ignoreon(); |
static void Mpass (void) { memcpy(keyword, "if ", 4); Pelif(); } |
} |
static void Mtrue (void) { keywordedit("else"); state(IS_TRUE_MIDDLE); } |
|
static void Melif (void) { keywordedit("endif"); state(IS_FALSE_TRAILER); } |
|
static void Melse (void) { keywordedit("endif"); state(IS_FALSE_ELSE); } |
|
|
static void |
|
Itrue(void) |
|
{ |
|
Ftrue(); |
|
ignoreon(); |
|
} |
|
|
|
static void |
|
Ifalse(void) |
|
{ |
|
Ffalse(); |
|
ignoreon(); |
|
} |
|
|
|
/* edit this line */ |
|
static void |
|
Mpass (void) |
|
{ |
|
strncpy(keyword, "if ", 4); |
|
Pelif(); |
|
} |
|
|
|
static void |
|
Mtrue (void) |
|
{ |
|
keywordedit("else\n"); |
|
state(IS_TRUE_MIDDLE); |
|
} |
|
|
|
static void |
|
Melif (void) |
|
{ |
|
keywordedit("endif\n"); |
|
state(IS_FALSE_TRAILER); |
|
} |
|
|
|
static void |
|
Melse (void) |
|
{ |
|
keywordedit("endif\n"); |
|
state(IS_FALSE_ELSE); |
|
} |
|
|
|
static state_fn * const trans_table[IS_COUNT][LT_COUNT] = { |
static state_fn * const trans_table[IS_COUNT][LT_COUNT] = { |
/* IS_OUTSIDE */ |
/* IS_OUTSIDE */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Eendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Eendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Eendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Eendif, |
print, NULL }, |
print, done, abort }, |
/* IS_FALSE_PREFIX */ |
/* IS_FALSE_PREFIX */ |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Strue, Sfalse,Selse, Dendif, |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Strue, Sfalse,Selse, Dendif, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Eioccc,Eioccc,Eioccc,Eioccc, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Mpass, Eioccc,Eioccc,Eioccc,Eioccc, |
drop, Eeof }, |
drop, Eeof, abort }, |
/* IS_TRUE_PREFIX */ |
/* IS_TRUE_PREFIX */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Dfalse,Dfalse,Dfalse,Delse, Dendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Dfalse,Dfalse,Dfalse,Delse, Dendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eioccc,Eioccc,Eioccc,Eioccc,Eioccc, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eioccc,Eioccc,Eioccc,Eioccc,Eioccc, |
print, Eeof }, |
print, Eeof, abort }, |
/* IS_PASS_MIDDLE */ |
/* IS_PASS_MIDDLE */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Pelif, Mtrue, Delif, Pelse, Pendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Pelif, Mtrue, Delif, Pelse, Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Pelif, Oelif, Oelif, Pelse, Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Pelif, Oelif, Oelif, Pelse, Pendif, |
print, Eeof }, |
print, Eeof, abort }, |
/* IS_FALSE_MIDDLE */ |
/* IS_FALSE_MIDDLE */ |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Pelif, Mtrue, Delif, Pelse, Pendif, |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Pelif, Mtrue, Delif, Pelse, Pendif, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eioccc,Eioccc,Eioccc,Eioccc,Eioccc, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eioccc,Eioccc,Eioccc,Eioccc,Eioccc, |
drop, Eeof }, |
drop, Eeof, abort }, |
/* IS_TRUE_MIDDLE */ |
/* IS_TRUE_MIDDLE */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Melif, Melif, Melif, Melse, Pendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Melif, Melif, Melif, Melse, Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eioccc,Eioccc,Eioccc,Eioccc,Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eioccc,Eioccc,Eioccc,Eioccc,Pendif, |
print, Eeof }, |
print, Eeof, abort }, |
/* IS_PASS_ELSE */ |
/* IS_PASS_ELSE */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Pendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Pendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Pendif, |
print, Eeof }, |
print, Eeof, abort }, |
/* IS_FALSE_ELSE */ |
/* IS_FALSE_ELSE */ |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Dendif, |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Dendif, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Eioccc, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Eelif, Eelif, Eelif, Eelse, Eioccc, |
drop, Eeof }, |
drop, Eeof, abort }, |
/* IS_TRUE_ELSE */ |
/* IS_TRUE_ELSE */ |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Dendif, |
{ Itrue, Ifalse,Fpass, Ftrue, Ffalse,Eelif, Eelif, Eelif, Eelse, Dendif, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Eioccc, |
Oiffy, Oiffy, Fpass, Oif, Oif, Eelif, Eelif, Eelif, Eelse, Eioccc, |
print, Eeof }, |
print, Eeof, abort }, |
/* IS_FALSE_TRAILER */ |
/* IS_FALSE_TRAILER */ |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Dendif, |
{ Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Dendif, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Eioccc, |
Idrop, Idrop, Fdrop, Fdrop, Fdrop, Dfalse,Dfalse,Dfalse,Delse, Eioccc, |
drop, Eeof } |
drop, Eeof, abort } |
/*TRUEI FALSEI IF TRUE FALSE ELIF ELTRUE ELFALSE ELSE ENDIF |
/*TRUEI FALSEI IF TRUE FALSE ELIF ELTRUE ELFALSE ELSE ENDIF |
TRUEI FALSEI IF TRUE FALSE ELIF ELTRUE ELFALSE ELSE ENDIF (DODGY) |
TRUEI FALSEI IF TRUE FALSE ELIF ELTRUE ELFALSE ELSE ENDIF (DODGY) |
PLAIN EOF */ |
PLAIN EOF ERROR */ |
}; |
}; |
|
|
/* |
/* |
|
|
static void |
static void |
ignoreoff(void) |
ignoreoff(void) |
{ |
{ |
|
if (depth == 0) |
|
abort(); /* bug */ |
ignoring[depth] = ignoring[depth-1]; |
ignoring[depth] = ignoring[depth-1]; |
} |
} |
|
|
static void |
static void |
ignoreon(void) |
ignoreon(void) |
{ |
{ |
ignoring[depth] = true; |
ignoring[depth] = true; |
} |
} |
|
|
static void |
static void |
keywordedit(const char *replacement) |
keywordedit(const char *replacement) |
{ |
{ |
strlcpy(keyword, replacement, tline + sizeof(tline) - keyword); |
snprintf(keyword, tline + sizeof(tline) - keyword, |
|
"%s%s", replacement, newline); |
print(); |
print(); |
} |
} |
|
|
static void |
static void |
nest(void) |
nest(void) |
{ |
{ |
depth += 1; |
if (depth > MAXDEPTH-1) |
if (depth >= MAXDEPTH) |
abort(); /* bug */ |
|
if (depth == MAXDEPTH-1) |
error("Too many levels of nesting"); |
error("Too many levels of nesting"); |
|
depth += 1; |
stifline[depth] = linenum; |
stifline[depth] = linenum; |
} |
} |
|
|
static void |
static void |
|
unnest(void) |
|
{ |
|
if (depth == 0) |
|
abort(); /* bug */ |
|
depth -= 1; |
|
} |
|
static void |
state(Ifstate is) |
state(Ifstate is) |
{ |
{ |
ifstate[depth] = is; |
ifstate[depth] = is; |
} |
} |
|
|
/* |
/* |
|
* The last state transition function. When this is called, |
|
* lineval == LT_EOF, so the process() loop will terminate. |
|
*/ |
|
static void |
|
done(void) |
|
{ |
|
if (incomment) |
|
error("EOF in comment"); |
|
closeio(); |
|
} |
|
|
|
/* |
* Write a line to the output or not, according to command line options. |
* Write a line to the output or not, according to command line options. |
|
* If writing fails, closeio() will print the error and exit. |
*/ |
*/ |
static void |
static void |
flushline(bool keep) |
flushline(bool keep) |
{ |
{ |
if (symlist) |
if (symlist) |
return; |
return; |
if (keep ^ complement) |
if (keep ^ complement) { |
fputs(tline, stdout); |
bool blankline = tline[strspn(tline, " \t\r\n")] == '\0'; |
else { |
if (blankline && compblank && blankcount != blankmax) { |
if (lnblank) |
delcount += 1; |
putc('\n', stdout); |
blankcount += 1; |
|
} else { |
|
if (lnnum && delcount > 0) |
|
hashline(); |
|
if (fputs(tline, output) == EOF) |
|
closeio(); |
|
delcount = 0; |
|
blankmax = blankcount = blankline ? blankcount + 1 : 0; |
|
} |
|
} else { |
|
if (lnblank && fputs(newline, output) == EOF) |
|
closeio(); |
exitstat = 1; |
exitstat = 1; |
|
delcount += 1; |
|
blankcount = 0; |
} |
} |
|
if (debugging && fflush(output) == EOF) |
|
closeio(); |
} |
} |
|
|
/* |
/* |
|
* Format of #line directives depends on whether we know the input filename. |
|
*/ |
|
static void |
|
hashline(void) |
|
{ |
|
int e; |
|
|
|
if (linefile == NULL) |
|
e = fprintf(output, "#line %d%s", linenum, newline); |
|
else |
|
e = fprintf(output, "#line %d \"%s\"%s", |
|
linenum, linefile, newline); |
|
if (e < 0) |
|
closeio(); |
|
} |
|
|
|
/* |
|
* Flush the output and handle errors. |
|
*/ |
|
static void |
|
closeio(void) |
|
{ |
|
/* Tidy up after findsym(). */ |
|
if (symdepth && !zerosyms) |
|
printf("\n"); |
|
if (output != NULL && (ferror(output) || fclose(output) == EOF)) |
|
err(2, "%s: can't write to output", filename); |
|
fclose(input); |
|
} |
|
|
|
/* |
* The driver for the state machine. |
* The driver for the state machine. |
*/ |
*/ |
static void |
static void |
process(void) |
process(void) |
{ |
{ |
Linetype lineval; |
Linetype lineval = LT_PLAIN; |
state_fn *trans; |
/* When compressing blank lines, act as if the file |
|
is preceded by a large number of blank lines. */ |
for (;;) { |
blankmax = blankcount = 1000; |
linenum++; |
zerosyms = true; |
lineval = get_line(); |
newline = NULL; |
trans = trans_table[ifstate[depth]][lineval]; |
linenum = 0; |
if (trans == NULL) |
while (lineval != LT_EOF) { |
break; |
lineval = parseline(); |
trans(); |
trans_table[ifstate[depth]][lineval](); |
debug("process %s -> %s depth %d", |
debug("process line %d %s -> %s depth %d", |
linetype_name[lineval], |
linenum, linetype_name[lineval], |
ifstate_name[ifstate[depth]], depth); |
ifstate_name[ifstate[depth]], depth); |
} |
} |
if (incomment) |
|
error("EOF in comment"); |
|
} |
} |
|
|
/* |
/* |
* Parse a line and determine its type. We keep the preprocessor line |
* Parse a line and determine its type. We keep the preprocessor line |
* parser state between calls in a global variable. |
* parser state between calls in the global variable linestate, with |
|
* help from skipcomment(). |
*/ |
*/ |
static Linetype |
static Linetype |
get_line(void) |
parseline(void) |
{ |
{ |
const char *cp; |
const char *cp; |
int cursym; |
int cursym; |
int kwlen; |
|
Linetype retval; |
Linetype retval; |
Comment_state wascomment; |
Comment_state wascomment; |
|
|
if (fgets(tline, MAXLINE, input) == NULL) |
|
return (LT_EOF); |
|
retval = LT_PLAIN; |
|
wascomment = incomment; |
wascomment = incomment; |
cp = skipcomment(tline); |
cp = skiphash(); |
if (linestate == LS_START) { |
if (cp == NULL) |
if (*cp == '#') { |
return (LT_EOF); |
linestate = LS_HASH; |
if (newline == NULL) { |
cp = skipcomment(cp + 1); |
if (strrchr(tline, '\n') == strrchr(tline, '\r') + 1) |
} else if (*cp != '\0') |
newline = newline_crlf; |
linestate = LS_DIRTY; |
else |
|
newline = newline_unix; |
} |
} |
if (!incomment && linestate == LS_HASH) { |
if (*cp == '\0') { |
keyword = tline + (cp - tline); |
retval = LT_PLAIN; |
cp = skipsym(cp); |
goto done; |
kwlen = cp - keyword; |
} |
|
keyword = tline + (cp - tline); |
|
if ((cp = matchsym("ifdef", keyword)) != NULL || |
|
(cp = matchsym("ifndef", keyword)) != NULL) { |
|
cp = skipcomment(cp); |
|
if ((cursym = findsym(&cp)) < 0) |
|
retval = LT_IF; |
|
else { |
|
retval = (keyword[2] == 'n') |
|
? LT_FALSE : LT_TRUE; |
|
if (value[cursym] == NULL) |
|
retval = (retval == LT_TRUE) |
|
? LT_FALSE : LT_TRUE; |
|
if (ignore[cursym]) |
|
retval = (retval == LT_TRUE) |
|
? LT_TRUEI : LT_FALSEI; |
|
} |
|
} else if ((cp = matchsym("if", keyword)) != NULL) |
|
retval = ifeval(&cp); |
|
else if ((cp = matchsym("elif", keyword)) != NULL) |
|
retval = linetype_if2elif(ifeval(&cp)); |
|
else if ((cp = matchsym("else", keyword)) != NULL) |
|
retval = LT_ELSE; |
|
else if ((cp = matchsym("endif", keyword)) != NULL) |
|
retval = LT_ENDIF; |
|
else { |
|
cp = skipsym(keyword); |
/* no way can we deal with a continuation inside a keyword */ |
/* no way can we deal with a continuation inside a keyword */ |
if (strncmp(cp, "\\\n", 2) == 0) |
if (strncmp(cp, "\\\r\n", 3) == 0 || |
|
strncmp(cp, "\\\n", 2) == 0) |
Eioccc(); |
Eioccc(); |
if (strlcmp("ifdef", keyword, kwlen) == 0 || |
cp = skipline(cp); |
strlcmp("ifndef", keyword, kwlen) == 0) { |
retval = LT_PLAIN; |
cp = skipcomment(cp); |
goto done; |
if ((cursym = findsym(cp)) < 0) |
} |
retval = LT_IF; |
cp = skipcomment(cp); |
else { |
if (*cp != '\0') { |
retval = (keyword[2] == 'n') |
cp = skipline(cp); |
? LT_FALSE : LT_TRUE; |
if (retval == LT_TRUE || retval == LT_FALSE || |
if (value[cursym] == NULL) |
retval == LT_TRUEI || retval == LT_FALSEI) |
retval = (retval == LT_TRUE) |
retval = LT_IF; |
? LT_FALSE : LT_TRUE; |
if (retval == LT_ELTRUE || retval == LT_ELFALSE) |
if (ignore[cursym]) |
retval = LT_ELIF; |
retval = (retval == LT_TRUE) |
} |
? LT_TRUEI : LT_FALSEI; |
/* the following can happen if the last line of the file lacks a |
} |
newline or if there is too much whitespace in a directive */ |
cp = skipsym(cp); |
if (linestate == LS_HASH) { |
} else if (strlcmp("if", keyword, kwlen) == 0) |
long len = cp - tline; |
retval = ifeval(&cp); |
if (fgets(tline + len, MAXLINE - len, input) == NULL) { |
else if (strlcmp("elif", keyword, kwlen) == 0) |
if (ferror(input)) |
retval = ifeval(&cp) - LT_IF + LT_ELIF; |
err(2, "can't read %s", filename); |
else if (strlcmp("else", keyword, kwlen) == 0) |
/* append the missing newline at eof */ |
retval = LT_ELSE; |
strcpy(tline + len, newline); |
else if (strlcmp("endif", keyword, kwlen) == 0) |
cp += strlen(newline); |
retval = LT_ENDIF; |
linestate = LS_START; |
else { |
} else { |
linestate = LS_DIRTY; |
linestate = LS_DIRTY; |
retval = LT_PLAIN; |
|
} |
} |
cp = skipcomment(cp); |
|
if (*cp != '\0') { |
|
linestate = LS_DIRTY; |
|
if (retval == LT_TRUE || retval == LT_FALSE || |
|
retval == LT_TRUEI || retval == LT_FALSEI) |
|
retval = LT_IF; |
|
if (retval == LT_ELTRUE || retval == LT_ELFALSE) |
|
retval = LT_ELIF; |
|
} |
|
if (retval != LT_PLAIN && (wascomment || incomment)) { |
|
retval += LT_DODGY; |
|
if (incomment) |
|
linestate = LS_DIRTY; |
|
} |
|
/* skipcomment should have changed the state */ |
|
if (linestate == LS_HASH) |
|
abort(); /* bug */ |
|
} |
} |
if (linestate == LS_DIRTY) { |
if (retval != LT_PLAIN && (wascomment || linestate != LS_START)) { |
while (*cp != '\0') |
retval = linetype_2dodgy(retval); |
cp = skipcomment(cp + 1); |
linestate = LS_DIRTY; |
} |
} |
debug("parser %s comment %s line", |
done: |
|
debug("parser line %d state %s comment %s line", linenum, |
comment_name[incomment], linestate_name[linestate]); |
comment_name[incomment], linestate_name[linestate]); |
return (retval); |
return (retval); |
} |
} |
|
|
/* |
/* |
* These are the operators that are supported by the expression |
* These are the binary operators that are supported by the expression |
* evaluator. Note that if support for division is added then we also |
* evaluator. |
* need short-circuiting booleans because of divide-by-zero. |
|
*/ |
*/ |
static int |
static Linetype op_strict(long *p, long v, Linetype at, Linetype bt) { |
op_lt(int a, int b) |
if(at == LT_IF || bt == LT_IF) return (LT_IF); |
{ |
return (*p = v, v ? LT_TRUE : LT_FALSE); |
return (a < b); |
|
} |
} |
|
static Linetype op_lt(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a < b, at, bt); |
op_gt(int a, int b) |
|
{ |
|
return (a > b); |
|
} |
} |
|
static Linetype op_gt(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a > b, at, bt); |
op_le(int a, int b) |
|
{ |
|
return (a <= b); |
|
} |
} |
|
static Linetype op_le(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a <= b, at, bt); |
op_ge(int a, int b) |
|
{ |
|
return (a >= b); |
|
} |
} |
|
static Linetype op_ge(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a >= b, at, bt); |
op_eq(int a, int b) |
|
{ |
|
return (a == b); |
|
} |
} |
|
static Linetype op_eq(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a == b, at, bt); |
op_ne(int a, int b) |
|
{ |
|
return (a != b); |
|
} |
} |
|
static Linetype op_ne(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
return op_strict(p, a != b, at, bt); |
op_or(int a, int b) |
|
{ |
|
return (a || b); |
|
} |
} |
|
static Linetype op_or(long *p, Linetype at, long a, Linetype bt, long b) { |
static int |
if (!strictlogic && (at == LT_TRUE || bt == LT_TRUE)) |
op_and(int a, int b) |
return (*p = 1, LT_TRUE); |
{ |
return op_strict(p, a || b, at, bt); |
return (a && b); |
|
} |
} |
|
static Linetype op_and(long *p, Linetype at, long a, Linetype bt, long b) { |
|
if (!strictlogic && (at == LT_FALSE || bt == LT_FALSE)) |
|
return (*p = 0, LT_FALSE); |
|
return op_strict(p, a && b, at, bt); |
|
} |
|
|
/* |
/* |
* An evaluation function takes three arguments, as follows: (1) a pointer to |
* An evaluation function takes three arguments, as follows: (1) a pointer to |
|
|
* value of the expression; and (3) a pointer to a char* that points to the |
* value of the expression; and (3) a pointer to a char* that points to the |
* expression to be evaluated and that is updated to the end of the expression |
* expression to be evaluated and that is updated to the end of the expression |
* when evaluation is complete. The function returns LT_FALSE if the value of |
* when evaluation is complete. The function returns LT_FALSE if the value of |
* the expression is zero, LT_TRUE if it is non-zero, or LT_IF if the |
* the expression is zero, LT_TRUE if it is non-zero, LT_IF if the expression |
* expression could not be evaluated. |
* depends on an unknown symbol, or LT_ERROR if there is a parse failure. |
*/ |
*/ |
struct ops; |
struct ops; |
|
|
typedef Linetype eval_fn(const struct ops *, int *, const char **); |
typedef Linetype eval_fn(const struct ops *, long *, const char **); |
|
|
static eval_fn eval_table, eval_unary; |
static eval_fn eval_table, eval_unary; |
|
|
|
|
* element of the table. Innermost expressions have special non-table-driven |
* element of the table. Innermost expressions have special non-table-driven |
* handling. |
* handling. |
*/ |
*/ |
static const struct ops { |
struct op { |
|
const char *str; |
|
Linetype (*fn)(long *, Linetype, long, Linetype, long); |
|
}; |
|
struct ops { |
eval_fn *inner; |
eval_fn *inner; |
struct op { |
struct op op[5]; |
const char *str; |
}; |
int (*fn)(int, int); |
static const struct ops eval_ops[] = { |
} op[5]; |
|
} eval_ops[] = { |
|
{ eval_table, { { "||", op_or } } }, |
{ eval_table, { { "||", op_or } } }, |
{ eval_table, { { "&&", op_and } } }, |
{ eval_table, { { "&&", op_and } } }, |
{ eval_table, { { "==", op_eq }, |
{ eval_table, { { "==", op_eq }, |
|
|
{ ">", op_gt } } } |
{ ">", op_gt } } } |
}; |
}; |
|
|
|
/* Current operator precedence level */ |
|
static long prec(const struct ops *ops) |
|
{ |
|
return (ops - eval_ops); |
|
} |
|
|
/* |
/* |
* Function for evaluating the innermost parts of expressions, |
* Function for evaluating the innermost parts of expressions, |
* viz. !expr (expr) defined(symbol) symbol number |
* viz. !expr (expr) number defined(symbol) symbol |
* We reset the keepthis flag when we find a non-constant subexpression. |
* We reset the constexpr flag in the last two cases. |
*/ |
*/ |
static Linetype |
static Linetype |
eval_unary(const struct ops *ops, int *valp, const char **cpp) |
eval_unary(const struct ops *ops, long *valp, const char **cpp) |
{ |
{ |
const char *cp; |
const char *cp; |
char *ep; |
char *ep; |
int sym; |
int sym; |
|
bool defparen; |
|
Linetype lt; |
|
|
cp = skipcomment(*cpp); |
cp = skipcomment(*cpp); |
if (*cp == '!') { |
if (*cp == '!') { |
debug("eval%d !", ops - eval_ops); |
debug("eval%d !", prec(ops)); |
cp++; |
cp++; |
if (eval_unary(ops, valp, &cp) == LT_IF) |
lt = eval_unary(ops, valp, &cp); |
return (LT_IF); |
if (lt == LT_ERROR) |
*valp = !*valp; |
return (LT_ERROR); |
|
if (lt != LT_IF) { |
|
*valp = !*valp; |
|
lt = *valp ? LT_TRUE : LT_FALSE; |
|
} |
} else if (*cp == '(') { |
} else if (*cp == '(') { |
cp++; |
cp++; |
debug("eval%d (", ops - eval_ops); |
debug("eval%d (", prec(ops)); |
if (eval_table(eval_ops, valp, &cp) == LT_IF) |
lt = eval_table(eval_ops, valp, &cp); |
return (LT_IF); |
if (lt == LT_ERROR) |
|
return (LT_ERROR); |
cp = skipcomment(cp); |
cp = skipcomment(cp); |
if (*cp++ != ')') |
if (*cp++ != ')') |
return (LT_IF); |
return (LT_ERROR); |
} else if (isdigit((unsigned char)*cp)) { |
} else if (isdigit((unsigned char)*cp)) { |
debug("eval%d number", ops - eval_ops); |
debug("eval%d number", prec(ops)); |
*valp = strtol(cp, &ep, 0); |
*valp = strtol(cp, &ep, 0); |
cp = skipsym(cp); |
if (ep == cp) |
} else if (strncmp(cp, "defined", 7) == 0 && endsym(cp[7])) { |
return (LT_ERROR); |
|
lt = *valp ? LT_TRUE : LT_FALSE; |
|
cp = ep; |
|
} else if (matchsym("defined", cp) != NULL) { |
cp = skipcomment(cp+7); |
cp = skipcomment(cp+7); |
debug("eval%d defined", ops - eval_ops); |
if (*cp == '(') { |
if (*cp++ != '(') |
cp = skipcomment(cp+1); |
return (LT_IF); |
defparen = true; |
|
} else { |
|
defparen = false; |
|
} |
|
sym = findsym(&cp); |
cp = skipcomment(cp); |
cp = skipcomment(cp); |
sym = findsym(cp); |
if (defparen && *cp++ != ')') { |
if (sym < 0 && !symlist) |
debug("eval%d defined missing ')'", prec(ops)); |
return (LT_IF); |
return (LT_ERROR); |
*valp = (value[sym] != NULL); |
} |
cp = skipsym(cp); |
if (sym < 0) { |
cp = skipcomment(cp); |
debug("eval%d defined unknown", prec(ops)); |
if (*cp++ != ')') |
lt = LT_IF; |
return (LT_IF); |
} else { |
keepthis = false; |
debug("eval%d defined %s", prec(ops), symname[sym]); |
|
*valp = (value[sym] != NULL); |
|
lt = *valp ? LT_TRUE : LT_FALSE; |
|
} |
|
constexpr = false; |
} else if (!endsym(*cp)) { |
} else if (!endsym(*cp)) { |
debug("eval%d symbol", ops - eval_ops); |
debug("eval%d symbol", prec(ops)); |
sym = findsym(cp); |
sym = findsym(&cp); |
if (sym < 0 && !symlist) |
if (sym < 0) { |
return (LT_IF); |
lt = LT_IF; |
if (value[sym] == NULL) |
cp = skipargs(cp); |
|
} else if (value[sym] == NULL) { |
*valp = 0; |
*valp = 0; |
else { |
lt = LT_FALSE; |
|
} else { |
*valp = strtol(value[sym], &ep, 0); |
*valp = strtol(value[sym], &ep, 0); |
if (*ep != '\0' || ep == value[sym]) |
if (*ep != '\0' || ep == value[sym]) |
return (LT_IF); |
return (LT_ERROR); |
|
lt = *valp ? LT_TRUE : LT_FALSE; |
|
cp = skipargs(cp); |
} |
} |
cp = skipsym(cp); |
constexpr = false; |
keepthis = false; |
} else { |
} else |
debug("eval%d bad expr", prec(ops)); |
return (LT_IF); |
return (LT_ERROR); |
|
} |
|
|
*cpp = cp; |
*cpp = cp; |
debug("eval%d = %d", ops - eval_ops, *valp); |
debug("eval%d = %d", prec(ops), *valp); |
return (*valp ? LT_TRUE : LT_FALSE); |
return (lt); |
} |
} |
|
|
/* |
/* |
* Table-driven evaluation of binary operators. |
* Table-driven evaluation of binary operators. |
*/ |
*/ |
static Linetype |
static Linetype |
eval_table(const struct ops *ops, int *valp, const char **cpp) |
eval_table(const struct ops *ops, long *valp, const char **cpp) |
{ |
{ |
const struct op *op; |
const struct op *op; |
const char *cp; |
const char *cp; |
int val; |
long val; |
|
Linetype lt, rt; |
|
|
debug("eval%d", ops - eval_ops); |
debug("eval%d", prec(ops)); |
cp = *cpp; |
cp = *cpp; |
if (ops->inner(ops+1, valp, &cp) == LT_IF) |
lt = ops->inner(ops+1, valp, &cp); |
return (LT_IF); |
if (lt == LT_ERROR) |
|
return (LT_ERROR); |
for (;;) { |
for (;;) { |
cp = skipcomment(cp); |
cp = skipcomment(cp); |
for (op = ops->op; op->str != NULL; op++) |
for (op = ops->op; op->str != NULL; op++) |
|
|
if (op->str == NULL) |
if (op->str == NULL) |
break; |
break; |
cp += strlen(op->str); |
cp += strlen(op->str); |
debug("eval%d %s", ops - eval_ops, op->str); |
debug("eval%d %s", prec(ops), op->str); |
if (ops->inner(ops+1, &val, &cp) == LT_IF) |
rt = ops->inner(ops+1, &val, &cp); |
return (LT_IF); |
if (rt == LT_ERROR) |
*valp = op->fn(*valp, val); |
return (LT_ERROR); |
|
lt = op->fn(valp, lt, *valp, rt, val); |
} |
} |
|
|
*cpp = cp; |
*cpp = cp; |
debug("eval%d = %d", ops - eval_ops, *valp); |
debug("eval%d = %d", prec(ops), *valp); |
return (*valp ? LT_TRUE : LT_FALSE); |
debug("eval%d lt = %s", prec(ops), linetype_name[lt]); |
|
return (lt); |
} |
} |
|
|
/* |
/* |
|
|
static Linetype |
static Linetype |
ifeval(const char **cpp) |
ifeval(const char **cpp) |
{ |
{ |
int ret; |
Linetype ret; |
int val; |
long val = 0; |
|
|
debug("eval %s", *cpp); |
debug("eval %s", *cpp); |
keepthis = killconsts ? false : true; |
constexpr = killconsts ? false : true; |
ret = eval_table(eval_ops, &val, cpp); |
ret = eval_table(eval_ops, &val, cpp); |
return (keepthis ? LT_IF : ret); |
debug("eval = %d", val); |
|
return (constexpr ? LT_IF : ret == LT_ERROR ? LT_IF : ret); |
} |
} |
|
|
/* |
/* |
* Skip over comments and stop at the next character position that is |
* Read a line and examine its initial part to determine if it is a |
* not whitespace. Between calls we keep the comment state in the |
* preprocessor directive. Returns NULL on EOF, or a pointer to a |
* global variable incomment, and we also adjust the global variable |
* preprocessor directive name, or a pointer to the zero byte at the |
* linestate when we see a newline. |
* end of the line. |
|
*/ |
|
static const char * |
|
skiphash(void) |
|
{ |
|
const char *cp; |
|
|
|
linenum++; |
|
if (fgets(tline, MAXLINE, input) == NULL) { |
|
if (ferror(input)) |
|
err(2, "can't read %s", filename); |
|
else |
|
return (NULL); |
|
} |
|
cp = skipcomment(tline); |
|
if (linestate == LS_START && *cp == '#') { |
|
linestate = LS_HASH; |
|
return (skipcomment(cp + 1)); |
|
} else if (*cp == '\0') { |
|
return (cp); |
|
} else { |
|
return (skipline(cp)); |
|
} |
|
} |
|
|
|
/* |
|
* Mark a line dirty and consume the rest of it, keeping track of the |
|
* lexical state. |
|
*/ |
|
static const char * |
|
skipline(const char *cp) |
|
{ |
|
if (*cp != '\0') |
|
linestate = LS_DIRTY; |
|
while (*cp != '\0') |
|
cp = skipcomment(cp + 1); |
|
return (cp); |
|
} |
|
|
|
/* |
|
* Skip over comments, strings, and character literals and stop at the |
|
* next character position that is not whitespace. Between calls we keep |
|
* the comment state in the global variable incomment, and we also adjust |
|
* the global variable linestate when we see a newline. |
* XXX: doesn't cope with the buffer splitting inside a state transition. |
* XXX: doesn't cope with the buffer splitting inside a state transition. |
*/ |
*/ |
static const char * |
static const char * |
|
|
return (cp); |
return (cp); |
} |
} |
while (*cp != '\0') |
while (*cp != '\0') |
if (strncmp(cp, "\\\n", 2) == 0) |
/* don't reset to LS_START after a line continuation */ |
|
if (strncmp(cp, "\\\r\n", 3) == 0) |
|
cp += 3; |
|
else if (strncmp(cp, "\\\n", 2) == 0) |
cp += 2; |
cp += 2; |
else switch (incomment) { |
else switch (incomment) { |
case NO_COMMENT: |
case NO_COMMENT: |
if (strncmp(cp, "/\\\n", 3) == 0) { |
if (strncmp(cp, "/\\\r\n", 4) == 0) { |
incomment = STARTING_COMMENT; |
incomment = STARTING_COMMENT; |
|
cp += 4; |
|
} else if (strncmp(cp, "/\\\n", 3) == 0) { |
|
incomment = STARTING_COMMENT; |
cp += 3; |
cp += 3; |
} else if (strncmp(cp, "/*", 2) == 0) { |
} else if (strncmp(cp, "/*", 2) == 0) { |
incomment = C_COMMENT; |
incomment = C_COMMENT; |
|
|
} else if (strncmp(cp, "//", 2) == 0) { |
} else if (strncmp(cp, "//", 2) == 0) { |
incomment = CXX_COMMENT; |
incomment = CXX_COMMENT; |
cp += 2; |
cp += 2; |
|
} else if (strncmp(cp, "\'", 1) == 0) { |
|
incomment = CHAR_LITERAL; |
|
linestate = LS_DIRTY; |
|
cp += 1; |
|
} else if (strncmp(cp, "\"", 1) == 0) { |
|
incomment = STRING_LITERAL; |
|
linestate = LS_DIRTY; |
|
cp += 1; |
} else if (strncmp(cp, "\n", 1) == 0) { |
} else if (strncmp(cp, "\n", 1) == 0) { |
linestate = LS_START; |
linestate = LS_START; |
cp += 1; |
cp += 1; |
} else if (strchr(" \t", *cp) != NULL) { |
} else if (strchr(" \r\t", *cp) != NULL) { |
cp += 1; |
cp += 1; |
} else |
} else |
return (cp); |
return (cp); |
|
|
} |
} |
cp += 1; |
cp += 1; |
continue; |
continue; |
|
case CHAR_LITERAL: |
|
case STRING_LITERAL: |
|
if ((incomment == CHAR_LITERAL && cp[0] == '\'') || |
|
(incomment == STRING_LITERAL && cp[0] == '\"')) { |
|
incomment = NO_COMMENT; |
|
cp += 1; |
|
} else if (cp[0] == '\\') { |
|
if (cp[1] == '\0') |
|
cp += 1; |
|
else |
|
cp += 2; |
|
} else if (strncmp(cp, "\n", 1) == 0) { |
|
if (incomment == CHAR_LITERAL) |
|
error("unterminated char literal"); |
|
else |
|
error("unterminated string literal"); |
|
} else |
|
cp += 1; |
|
continue; |
case C_COMMENT: |
case C_COMMENT: |
if (strncmp(cp, "*\\\n", 3) == 0) { |
if (strncmp(cp, "*\\\r\n", 4) == 0) { |
incomment = FINISHING_COMMENT; |
incomment = FINISHING_COMMENT; |
|
cp += 4; |
|
} else if (strncmp(cp, "*\\\n", 3) == 0) { |
|
incomment = FINISHING_COMMENT; |
cp += 3; |
cp += 3; |
} else if (strncmp(cp, "*/", 2) == 0) { |
} else if (strncmp(cp, "*/", 2) == 0) { |
incomment = NO_COMMENT; |
incomment = NO_COMMENT; |
|
|
incomment = C_COMMENT; |
incomment = C_COMMENT; |
continue; |
continue; |
default: |
default: |
/* bug */ |
abort(); /* bug */ |
abort(); |
|
} |
} |
return (cp); |
return (cp); |
} |
} |
|
|
/* |
/* |
|
* Skip macro arguments. |
|
*/ |
|
static const char * |
|
skipargs(const char *cp) |
|
{ |
|
const char *ocp = cp; |
|
int level = 0; |
|
cp = skipcomment(cp); |
|
if (*cp != '(') |
|
return (cp); |
|
do { |
|
if (*cp == '(') |
|
level++; |
|
if (*cp == ')') |
|
level--; |
|
cp = skipcomment(cp+1); |
|
} while (level != 0 && *cp != '\0'); |
|
if (level == 0) |
|
return (cp); |
|
else |
|
/* Rewind and re-detect the syntax error later. */ |
|
return (ocp); |
|
} |
|
|
|
/* |
* Skip over an identifier. |
* Skip over an identifier. |
*/ |
*/ |
static const char * |
static const char * |
|
|
} |
} |
|
|
/* |
/* |
|
* Skip whitespace and take a copy of any following identifier. |
|
*/ |
|
static const char * |
|
getsym(const char **cpp) |
|
{ |
|
const char *cp = *cpp, *sym; |
|
|
|
cp = skipcomment(cp); |
|
cp = skipsym(sym = cp); |
|
if (cp == sym) |
|
return NULL; |
|
*cpp = cp; |
|
return (xstrdup(sym, cp)); |
|
} |
|
|
|
/* |
|
* Check that s (a symbol) matches the start of t, and that the |
|
* following character in t is not a symbol character. Returns a |
|
* pointer to the following character in t if there is a match, |
|
* otherwise NULL. |
|
*/ |
|
static const char * |
|
matchsym(const char *s, const char *t) |
|
{ |
|
while (*s != '\0' && *t != '\0') |
|
if (*s != *t) |
|
return (NULL); |
|
else |
|
++s, ++t; |
|
if (*s == '\0' && endsym(*t)) |
|
return(t); |
|
else |
|
return(NULL); |
|
} |
|
|
|
/* |
* Look for the symbol in the symbol table. If it is found, we return |
* Look for the symbol in the symbol table. If it is found, we return |
* the symbol table index, else we return -1. |
* the symbol table index, else we return -1. |
*/ |
*/ |
static int |
static int |
findsym(const char *str) |
findsym(const char **strp) |
{ |
{ |
const char *cp; |
const char *str; |
int symind; |
int symind; |
|
|
cp = skipsym(str); |
str = *strp; |
if (cp == str) |
*strp = skipsym(str); |
return (-1); |
if (symlist) { |
if (symlist) |
if (*strp == str) |
printf("%.*s\n", (int)(cp-str), str); |
return (-1); |
|
if (symdepth && firstsym) |
|
printf("%s%3d", zerosyms ? "" : "\n", depth); |
|
firstsym = zerosyms = false; |
|
printf("%s%.*s%s", |
|
symdepth ? " " : "", |
|
(int)(*strp-str), str, |
|
symdepth ? "" : "\n"); |
|
/* we don't care about the value of the symbol */ |
|
return (0); |
|
} |
for (symind = 0; symind < nsyms; ++symind) { |
for (symind = 0; symind < nsyms; ++symind) { |
if (strlcmp(symname[symind], str, cp-str) == 0) { |
if (matchsym(symname[symind], str) != NULL) { |
debug("findsym %s %s", symname[symind], |
debugsym("findsym", symind); |
value[symind] ? value[symind] : ""); |
|
return (symind); |
return (symind); |
} |
} |
} |
} |
|
|
} |
} |
|
|
/* |
/* |
|
* Resolve indirect symbol values to their final definitions. |
|
*/ |
|
static void |
|
indirectsym(void) |
|
{ |
|
const char *cp; |
|
int changed, sym, ind; |
|
|
|
do { |
|
changed = 0; |
|
for (sym = 0; sym < nsyms; ++sym) { |
|
if (value[sym] == NULL) |
|
continue; |
|
cp = value[sym]; |
|
ind = findsym(&cp); |
|
if (ind == -1 || ind == sym || |
|
*cp != '\0' || |
|
value[ind] == NULL || |
|
value[ind] == value[sym]) |
|
continue; |
|
debugsym("indir...", sym); |
|
value[sym] = value[ind]; |
|
debugsym("...ectsym", sym); |
|
changed++; |
|
} |
|
} while (changed); |
|
} |
|
|
|
/* |
|
* Add a symbol to the symbol table, specified with the format sym=val |
|
*/ |
|
static void |
|
addsym1(bool ignorethis, bool definethis, char *symval) |
|
{ |
|
const char *sym, *val; |
|
|
|
sym = symval; |
|
val = skipsym(sym); |
|
if (definethis && *val == '=') { |
|
symval[val - sym] = '\0'; |
|
val = val + 1; |
|
} else if (*val == '\0') { |
|
val = definethis ? "1" : NULL; |
|
} else { |
|
usage(); |
|
} |
|
addsym2(ignorethis, sym, val); |
|
} |
|
|
|
/* |
* Add a symbol to the symbol table. |
* Add a symbol to the symbol table. |
*/ |
*/ |
static void |
static void |
addsym(bool ignorethis, bool definethis, char *sym) |
addsym2(bool ignorethis, const char *sym, const char *val) |
{ |
{ |
|
const char *cp = sym; |
int symind; |
int symind; |
char *val; |
|
|
|
symind = findsym(sym); |
symind = findsym(&cp); |
if (symind < 0) { |
if (symind < 0) { |
if (nsyms >= MAXSYMS) |
if (nsyms >= MAXSYMS) |
errx(2, "too many symbols"); |
errx(2, "too many symbols"); |
symind = nsyms++; |
symind = nsyms++; |
} |
} |
symname[symind] = sym; |
|
ignore[symind] = ignorethis; |
ignore[symind] = ignorethis; |
val = sym + (skipsym(sym) - sym); |
symname[symind] = sym; |
if (definethis) { |
value[symind] = val; |
if (*val == '=') { |
debugsym("addsym", symind); |
value[symind] = val+1; |
} |
*val = '\0'; |
|
} else if (*val == '\0') |
static void |
value[symind] = ""; |
debugsym(const char *why, int symind) |
else |
{ |
usage(); |
debug("%s %s%c%s", why, symname[symind], |
|
value[symind] ? '=' : ' ', |
|
value[symind] ? value[symind] : "undef"); |
|
} |
|
|
|
/* |
|
* Add symbols to the symbol table from a file containing |
|
* #define and #undef preprocessor directives. |
|
*/ |
|
static void |
|
defundefile(const char *fn) |
|
{ |
|
filename = fn; |
|
input = fopen(fn, "rb"); |
|
if (input == NULL) |
|
err(2, "can't open %s", fn); |
|
linenum = 0; |
|
while (defundef()) |
|
; |
|
if (ferror(input)) |
|
err(2, "can't read %s", filename); |
|
else |
|
fclose(input); |
|
if (incomment) |
|
error("EOF in comment"); |
|
} |
|
|
|
/* |
|
* Read and process one #define or #undef directive |
|
*/ |
|
static bool |
|
defundef(void) |
|
{ |
|
const char *cp, *kw, *sym, *val, *end; |
|
|
|
cp = skiphash(); |
|
if (cp == NULL) |
|
return (false); |
|
if (*cp == '\0') |
|
goto done; |
|
/* strip trailing whitespace, and do a fairly rough check to |
|
avoid unsupported multi-line preprocessor directives */ |
|
end = cp + strlen(cp); |
|
while (end > tline && strchr(" \t\n\r", end[-1]) != NULL) |
|
--end; |
|
if (end > tline && end[-1] == '\\') |
|
Eioccc(); |
|
|
|
kw = cp; |
|
if ((cp = matchsym("define", kw)) != NULL) { |
|
sym = getsym(&cp); |
|
if (sym == NULL) |
|
error("missing macro name in #define"); |
|
if (*cp == '(') { |
|
val = "1"; |
|
} else { |
|
cp = skipcomment(cp); |
|
val = (cp < end) ? xstrdup(cp, end) : ""; |
|
} |
|
debug("#define"); |
|
addsym2(false, sym, val); |
|
} else if ((cp = matchsym("undef", kw)) != NULL) { |
|
sym = getsym(&cp); |
|
if (sym == NULL) |
|
error("missing macro name in #undef"); |
|
cp = skipcomment(cp); |
|
debug("#undef"); |
|
addsym2(false, sym, NULL); |
} else { |
} else { |
if (*val != '\0') |
error("unrecognized preprocessor directive"); |
usage(); |
|
value[symind] = NULL; |
|
} |
} |
|
skipline(cp); |
|
done: |
|
debug("parser line %d state %s comment %s line", linenum, |
|
comment_name[incomment], linestate_name[linestate]); |
|
return (true); |
} |
} |
|
|
/* |
/* |
* Compare s with n characters of t. |
* Concatenate two strings into new memory, checking for failure. |
* The same as strncmp() except that it checks that s[n] == '\0'. |
|
*/ |
*/ |
static int |
static char * |
strlcmp(const char *s, const char *t, size_t n) |
astrcat(const char *s1, const char *s2) |
{ |
{ |
while (n-- && *t != '\0') |
char *s; |
if (*s != *t) |
int len; |
return ((unsigned char)*s - (unsigned char)*t); |
size_t size; |
else |
|
++s, ++t; |
len = snprintf(NULL, 0, "%s%s", s1, s2); |
return ((unsigned char)*s); |
if (len < 0) |
|
err(2, "snprintf"); |
|
size = (size_t)len + 1; |
|
s = (char *)malloc(size); |
|
if (s == NULL) |
|
err(2, "malloc"); |
|
snprintf(s, size, "%s%s", s1, s2); |
|
return (s); |
} |
} |
|
|
/* |
/* |
|
* Duplicate a segment of a string, checking for failure. |
|
*/ |
|
static const char * |
|
xstrdup(const char *start, const char *end) |
|
{ |
|
size_t n; |
|
char *s; |
|
|
|
if (end < start) abort(); /* bug */ |
|
n = (size_t)(end - start) + 1; |
|
s = malloc(n); |
|
if (s == NULL) |
|
err(2, "malloc"); |
|
snprintf(s, n, "%s", start); |
|
return (s); |
|
} |
|
|
|
/* |
* Diagnostics. |
* Diagnostics. |
*/ |
*/ |
static void |
static void |
|
|
else |
else |
warnx("%s: %d: %s (#if line %d depth %d)", |
warnx("%s: %d: %s (#if line %d depth %d)", |
filename, linenum, msg, stifline[depth], depth); |
filename, linenum, msg, stifline[depth], depth); |
|
closeio(); |
errx(2, "output may be truncated"); |
errx(2, "output may be truncated"); |
} |
} |