*
* Copyright (C) 2002 by Dmitry Zakharov <dmit@crp.bank.gov.ua>
*
- * Licensed under the GPL v2 or later, see the file LICENSE in this tarball.
+ * Licensed under GPLv2 or later, see file LICENSE in this source tree.
*/
+//config:config AWK
+//config: bool "awk (22 kb)"
+//config: default y
+//config: help
+//config: Awk is used as a pattern scanning and processing language.
+//config:
+//config:config FEATURE_AWK_LIBM
+//config: bool "Enable math functions (requires libm)"
+//config: default y
+//config: depends on AWK
+//config: help
+//config: Enable math functions of the Awk programming language.
+//config: NOTE: This requires libm to be present for linking.
+//config:
+//config:config FEATURE_AWK_GNU_EXTENSIONS
+//config: bool "Enable a few GNU extensions"
+//config: default y
+//config: depends on AWK
+//config: help
+//config: Enable a few features from gawk:
+//config: * command line option -e AWK_PROGRAM
+//config: * simultaneous use of -f and -e on the command line.
+//config: This enables the use of awk library files.
+//config: Example: awk -f mylib.awk -e '{print myfunction($1);}' ...
+
+//applet:IF_AWK(APPLET_NOEXEC(awk, awk, BB_DIR_USR_BIN, BB_SUID_DROP, awk))
+
+//kbuild:lib-$(CONFIG_AWK) += awk.o
+
+//usage:#define awk_trivial_usage
+//usage: "[OPTIONS] [AWK_PROGRAM] [FILE]..."
+//usage:#define awk_full_usage "\n\n"
+//usage: " -v VAR=VAL Set variable"
+//usage: "\n -F SEP Use SEP as field separator"
+//usage: "\n -f FILE Read program from FILE"
+//usage: IF_FEATURE_AWK_GNU_EXTENSIONS(
+//usage: "\n -e AWK_PROGRAM"
+//usage: )
#include "libbb.h"
#include "xregex.h"
/* If you comment out one of these below, it will be #defined later
* to perform debug printfs to stderr: */
#define debug_printf_walker(...) do {} while (0)
+#define debug_printf_eval(...) do {} while (0)
+#define debug_printf_parse(...) do {} while (0)
#ifndef debug_printf_walker
# define debug_printf_walker(...) (fprintf(stderr, __VA_ARGS__))
#endif
+#ifndef debug_printf_eval
+# define debug_printf_eval(...) (fprintf(stderr, __VA_ARGS__))
+#endif
+#ifndef debug_printf_parse
+# define debug_printf_parse(...) (fprintf(stderr, __VA_ARGS__))
+#endif
+/* "+": stop on first non-option:
+ * $ awk 'BEGIN { for(i=1; i<ARGC; ++i) { print i ": " ARGV[i] }}' -argz
+ * 1: -argz
+ */
+#define OPTSTR_AWK "+" \
+ "F:v:*f:*" \
+ IF_FEATURE_AWK_GNU_EXTENSIONS("e:*") \
+ "W:"
+enum {
+ OPTBIT_F, /* define field separator */
+ OPTBIT_v, /* define variable */
+ OPTBIT_f, /* pull in awk program from file */
+ IF_FEATURE_AWK_GNU_EXTENSIONS(OPTBIT_e,) /* -e AWK_PROGRAM */
+ OPTBIT_W, /* -W ignored */
+ OPT_F = 1 << OPTBIT_F,
+ OPT_v = 1 << OPTBIT_v,
+ OPT_f = 1 << OPTBIT_f,
+ OPT_e = IF_FEATURE_AWK_GNU_EXTENSIONS((1 << OPTBIT_e)) + 0,
+ OPT_W = 1 << OPTBIT_W
+};
#define MAXVARFMT 240
#define MINNVBLOCK 64
/* simple token classes */
/* Order and hex values are very important!!! See next_token() */
-#define TC_SEQSTART 1 /* ( */
+#define TC_SEQSTART (1 << 0) /* ( */
#define TC_SEQTERM (1 << 1) /* ) */
#define TC_REGEXP (1 << 2) /* /.../ */
#define TC_OUTRDR (1 << 3) /* | > >> */
#define TC_WHILE (1 << 17)
#define TC_ELSE (1 << 18)
#define TC_BUILTIN (1 << 19)
-#define TC_GETLINE (1 << 20)
-#define TC_FUNCDECL (1 << 21) /* `function' `func' */
-#define TC_BEGIN (1 << 22)
-#define TC_END (1 << 23)
-#define TC_EOF (1 << 24)
-#define TC_VARIABLE (1 << 25)
-#define TC_ARRAY (1 << 26)
-#define TC_FUNCTION (1 << 27)
-#define TC_STRING (1 << 28)
-#define TC_NUMBER (1 << 29)
+/* This costs ~50 bytes of code.
+ * A separate class to support deprecated "length" form. If we don't need that
+ * (i.e. if we demand that only "length()" with () is valid), then TC_LENGTH
+ * can be merged with TC_BUILTIN:
+ */
+#define TC_LENGTH (1 << 20)
+#define TC_GETLINE (1 << 21)
+#define TC_FUNCDECL (1 << 22) /* 'function' 'func' */
+#define TC_BEGIN (1 << 23)
+#define TC_END (1 << 24)
+#define TC_EOF (1 << 25)
+#define TC_VARIABLE (1 << 26)
+#define TC_ARRAY (1 << 27)
+#define TC_FUNCTION (1 << 28)
+#define TC_STRING (1 << 29)
+#define TC_NUMBER (1 << 30)
#define TC_UOPPRE (TC_UOPPRE1 | TC_UOPPRE2)
/* combined token classes */
#define TC_BINOP (TC_BINOPX | TC_COMMA | TC_PIPE | TC_IN)
-#define TC_UNARYOP (TC_UOPPRE | TC_UOPPOST)
+//#define TC_UNARYOP (TC_UOPPRE | TC_UOPPOST)
#define TC_OPERAND (TC_VARIABLE | TC_ARRAY | TC_FUNCTION \
- | TC_BUILTIN | TC_GETLINE | TC_SEQSTART | TC_STRING | TC_NUMBER)
+ | TC_BUILTIN | TC_LENGTH | TC_GETLINE \
+ | TC_SEQSTART | TC_STRING | TC_NUMBER)
#define TC_STATEMNT (TC_STATX | TC_WHILE)
#define TC_OPTERM (TC_SEMICOL | TC_NEWLINE)
/* word tokens, cannot mean something else if not expected */
-#define TC_WORD (TC_IN | TC_STATEMNT | TC_ELSE | TC_BUILTIN \
- | TC_GETLINE | TC_FUNCDECL | TC_BEGIN | TC_END)
+#define TC_WORD (TC_IN | TC_STATEMNT | TC_ELSE \
+ | TC_BUILTIN | TC_LENGTH | TC_GETLINE \
+ | TC_FUNCDECL | TC_BEGIN | TC_END)
/* discard newlines after these */
#define TC_NOTERM (TC_COMMA | TC_GRPSTART | TC_GRPTERM \
* For builtins it has different meaning: n n s3 s2 s1 v3 v2 v1,
* n - min. number of args, vN - resolve Nth arg to var, sN - resolve to string
*/
+#undef P
+#undef PRIMASK
+#undef PRIMASK2
#define P(x) (x << 24)
#define PRIMASK 0x7F000000
#define PRIMASK2 0x7E000000
/* tokens and their corresponding info values */
-#define NTC "\377" /* switch to next token class (tc<<1) */
-#define NTCC '\377'
-
-#define OC_B OC_BUILTIN
+#define NTC "\377" /* switch to next token class (tc<<1) */
+#define NTCC '\377'
static const char tokenlist[] ALIGN1 =
- "\1(" NTC
- "\1)" NTC
- "\1/" NTC /* REGEXP */
- "\2>>" "\1>" "\1|" NTC /* OUTRDR */
- "\2++" "\2--" NTC /* UOPPOST */
- "\2++" "\2--" "\1$" NTC /* UOPPRE1 */
- "\2==" "\1=" "\2+=" "\2-=" /* BINOPX */
- "\2*=" "\2/=" "\2%=" "\2^="
- "\1+" "\1-" "\3**=" "\2**"
- "\1/" "\1%" "\1^" "\1*"
- "\2!=" "\2>=" "\2<=" "\1>"
- "\1<" "\2!~" "\1~" "\2&&"
- "\2||" "\1?" "\1:" NTC
- "\2in" NTC
- "\1," NTC
- "\1|" NTC
- "\1+" "\1-" "\1!" NTC /* UOPPRE2 */
- "\1]" NTC
- "\1{" NTC
- "\1}" NTC
- "\1;" NTC
- "\1\n" NTC
- "\2if" "\2do" "\3for" "\5break" /* STATX */
- "\10continue" "\6delete" "\5print"
- "\6printf" "\4next" "\10nextfile"
- "\6return" "\4exit" NTC
- "\5while" NTC
- "\4else" NTC
-
- "\3and" "\5compl" "\6lshift" "\2or"
- "\6rshift" "\3xor"
- "\5close" "\6system" "\6fflush" "\5atan2" /* BUILTIN */
- "\3cos" "\3exp" "\3int" "\3log"
- "\4rand" "\3sin" "\4sqrt" "\5srand"
- "\6gensub" "\4gsub" "\5index" "\6length"
- "\5match" "\5split" "\7sprintf" "\3sub"
- "\6substr" "\7systime" "\10strftime" "\6mktime"
- "\7tolower" "\7toupper" NTC
- "\7getline" NTC
- "\4func" "\10function" NTC
- "\5BEGIN" NTC
- "\3END" "\0"
+ "\1(" NTC /* TC_SEQSTART */
+ "\1)" NTC /* TC_SEQTERM */
+ "\1/" NTC /* TC_REGEXP */
+ "\2>>" "\1>" "\1|" NTC /* TC_OUTRDR */
+ "\2++" "\2--" NTC /* TC_UOPPOST */
+ "\2++" "\2--" "\1$" NTC /* TC_UOPPRE1 */
+ "\2==" "\1=" "\2+=" "\2-=" /* TC_BINOPX */
+ "\2*=" "\2/=" "\2%=" "\2^="
+ "\1+" "\1-" "\3**=" "\2**"
+ "\1/" "\1%" "\1^" "\1*"
+ "\2!=" "\2>=" "\2<=" "\1>"
+ "\1<" "\2!~" "\1~" "\2&&"
+ "\2||" "\1?" "\1:" NTC
+ "\2in" NTC /* TC_IN */
+ "\1," NTC /* TC_COMMA */
+ "\1|" NTC /* TC_PIPE */
+ "\1+" "\1-" "\1!" NTC /* TC_UOPPRE2 */
+ "\1]" NTC /* TC_ARRTERM */
+ "\1{" NTC /* TC_GRPSTART */
+ "\1}" NTC /* TC_GRPTERM */
+ "\1;" NTC /* TC_SEMICOL */
+ "\1\n" NTC /* TC_NEWLINE */
+ "\2if" "\2do" "\3for" "\5break" /* TC_STATX */
+ "\10continue" "\6delete" "\5print"
+ "\6printf" "\4next" "\10nextfile"
+ "\6return" "\4exit" NTC
+ "\5while" NTC /* TC_WHILE */
+ "\4else" NTC /* TC_ELSE */
+ "\3and" "\5compl" "\6lshift" "\2or" /* TC_BUILTIN */
+ "\6rshift" "\3xor"
+ "\5close" "\6system" "\6fflush" "\5atan2"
+ "\3cos" "\3exp" "\3int" "\3log"
+ "\4rand" "\3sin" "\4sqrt" "\5srand"
+ "\6gensub" "\4gsub" "\5index" /* "\6length" was here */
+ "\5match" "\5split" "\7sprintf" "\3sub"
+ "\6substr" "\7systime" "\10strftime" "\6mktime"
+ "\7tolower" "\7toupper" NTC
+ "\6length" NTC /* TC_LENGTH */
+ "\7getline" NTC /* TC_GETLINE */
+ "\4func" "\10function" NTC /* TC_FUNCDECL */
+ "\5BEGIN" NTC /* TC_BEGIN */
+ "\3END" /* TC_END */
+ /* compiler adds trailing "\0" */
;
+#define OC_B OC_BUILTIN
+
static const uint32_t tokeninfo[] = {
0,
0,
OC_REGEXP,
- xS|'a', xS|'w', xS|'|',
- OC_UNARY|xV|P(9)|'p', OC_UNARY|xV|P(9)|'m',
- OC_UNARY|xV|P(9)|'P', OC_UNARY|xV|P(9)|'M',
- OC_FIELD|xV|P(5),
- OC_COMPARE|VV|P(39)|5, OC_MOVE|VV|P(74),
- OC_REPLACE|NV|P(74)|'+', OC_REPLACE|NV|P(74)|'-',
- OC_REPLACE|NV|P(74)|'*', OC_REPLACE|NV|P(74)|'/',
- OC_REPLACE|NV|P(74)|'%', OC_REPLACE|NV|P(74)|'&',
- OC_BINARY|NV|P(29)|'+', OC_BINARY|NV|P(29)|'-',
- OC_REPLACE|NV|P(74)|'&', OC_BINARY|NV|P(15)|'&',
- OC_BINARY|NV|P(25)|'/', OC_BINARY|NV|P(25)|'%',
- OC_BINARY|NV|P(15)|'&', OC_BINARY|NV|P(25)|'*',
- OC_COMPARE|VV|P(39)|4, OC_COMPARE|VV|P(39)|3,
- OC_COMPARE|VV|P(39)|0, OC_COMPARE|VV|P(39)|1,
- OC_COMPARE|VV|P(39)|2, OC_MATCH|Sx|P(45)|'!',
- OC_MATCH|Sx|P(45)|'~', OC_LAND|Vx|P(55),
- OC_LOR|Vx|P(59), OC_TERNARY|Vx|P(64)|'?',
- OC_COLON|xx|P(67)|':',
- OC_IN|SV|P(49),
+ xS|'a', xS|'w', xS|'|',
+ OC_UNARY|xV|P(9)|'p', OC_UNARY|xV|P(9)|'m',
+ OC_UNARY|xV|P(9)|'P', OC_UNARY|xV|P(9)|'M', OC_FIELD|xV|P(5),
+ OC_COMPARE|VV|P(39)|5, OC_MOVE|VV|P(74), OC_REPLACE|NV|P(74)|'+', OC_REPLACE|NV|P(74)|'-',
+ OC_REPLACE|NV|P(74)|'*', OC_REPLACE|NV|P(74)|'/', OC_REPLACE|NV|P(74)|'%', OC_REPLACE|NV|P(74)|'&',
+ OC_BINARY|NV|P(29)|'+', OC_BINARY|NV|P(29)|'-', OC_REPLACE|NV|P(74)|'&', OC_BINARY|NV|P(15)|'&',
+ OC_BINARY|NV|P(25)|'/', OC_BINARY|NV|P(25)|'%', OC_BINARY|NV|P(15)|'&', OC_BINARY|NV|P(25)|'*',
+ OC_COMPARE|VV|P(39)|4, OC_COMPARE|VV|P(39)|3, OC_COMPARE|VV|P(39)|0, OC_COMPARE|VV|P(39)|1,
+ OC_COMPARE|VV|P(39)|2, OC_MATCH|Sx|P(45)|'!', OC_MATCH|Sx|P(45)|'~', OC_LAND|Vx|P(55),
+ OC_LOR|Vx|P(59), OC_TERNARY|Vx|P(64)|'?', OC_COLON|xx|P(67)|':',
+ OC_IN|SV|P(49), /* TC_IN */
OC_COMMA|SS|P(80),
OC_PGETLINE|SV|P(37),
- OC_UNARY|xV|P(19)|'+', OC_UNARY|xV|P(19)|'-',
- OC_UNARY|xV|P(19)|'!',
- 0,
- 0,
+ OC_UNARY|xV|P(19)|'+', OC_UNARY|xV|P(19)|'-', OC_UNARY|xV|P(19)|'!',
+ 0, /* ] */
0,
0,
0,
- ST_IF, ST_DO, ST_FOR, OC_BREAK,
- OC_CONTINUE, OC_DELETE|Vx, OC_PRINT,
- OC_PRINTF, OC_NEXT, OC_NEXTFILE,
- OC_RETURN|Vx, OC_EXIT|Nx,
+ 0, /* \n */
+ ST_IF, ST_DO, ST_FOR, OC_BREAK,
+ OC_CONTINUE, OC_DELETE|Vx, OC_PRINT,
+ OC_PRINTF, OC_NEXT, OC_NEXTFILE,
+ OC_RETURN|Vx, OC_EXIT|Nx,
ST_WHILE,
- 0,
-
+ 0, /* else */
OC_B|B_an|P(0x83), OC_B|B_co|P(0x41), OC_B|B_ls|P(0x83), OC_B|B_or|P(0x83),
OC_B|B_rs|P(0x83), OC_B|B_xo|P(0x83),
OC_FBLTIN|Sx|F_cl, OC_FBLTIN|Sx|F_sy, OC_FBLTIN|Sx|F_ff, OC_B|B_a2|P(0x83),
OC_FBLTIN|Nx|F_co, OC_FBLTIN|Nx|F_ex, OC_FBLTIN|Nx|F_in, OC_FBLTIN|Nx|F_lg,
OC_FBLTIN|F_rn, OC_FBLTIN|Nx|F_si, OC_FBLTIN|Nx|F_sq, OC_FBLTIN|Nx|F_sr,
- OC_B|B_ge|P(0xd6), OC_B|B_gs|P(0xb6), OC_B|B_ix|P(0x9b), OC_FBLTIN|Sx|F_le,
+ OC_B|B_ge|P(0xd6), OC_B|B_gs|P(0xb6), OC_B|B_ix|P(0x9b), /* OC_FBLTIN|Sx|F_le, was here */
OC_B|B_ma|P(0x89), OC_B|B_sp|P(0x8b), OC_SPRINTF, OC_B|B_su|P(0xb6),
OC_B|B_ss|P(0x8f), OC_FBLTIN|F_ti, OC_B|B_ti|P(0x0b), OC_B|B_mt|P(0x0b),
OC_B|B_lo|P(0x49), OC_B|B_up|P(0x49),
+ OC_FBLTIN|Sx|F_le, /* TC_LENGTH */
OC_GETLINE|SV|P(0),
- 0, 0,
+ 0, 0,
0,
- 0
+ 0 /* TC_END */
};
/* internal variable names and their initial values */
ORS, RS, RT, FILENAME,
SUBSEP, F0, ARGIND, ARGC,
ARGV, ERRNO, FNR, NR,
- NF, IGNORECASE, ENVIRON, NUM_INTERNAL_VARS
+ NF, IGNORECASE, ENVIRON, NUM_INTERNAL_VARS
};
static const char vNames[] ALIGN1 =
smallint nextrec;
smallint nextfile;
smallint is_f0_split;
+ smallint t_rollback;
};
struct globals2 {
uint32_t t_info; /* often used */
uint32_t t_tclass;
char *t_string;
int t_lineno;
- int t_rollback;
var *intvar[NUM_INTERNAL_VARS]; /* often used */
#define nextrec (G1.nextrec )
#define nextfile (G1.nextfile )
#define is_f0_split (G1.is_f0_split )
+#define t_rollback (G1.t_rollback )
#define t_info (G.t_info )
#define t_tclass (G.t_tclass )
#define t_string (G.t_string )
#define t_lineno (G.t_lineno )
-#define t_rollback (G.t_rollback )
#define intvar (G.intvar )
#define fsplitter (G.fsplitter )
#define rsplitter (G.rsplitter )
static const char EMSG_NOT_ARRAY[] ALIGN1 = "Not an array";
static const char EMSG_POSSIBLE_ERROR[] ALIGN1 = "Possible syntax error";
static const char EMSG_UNDEF_FUNC[] ALIGN1 = "Call to undefined function";
-#if !ENABLE_FEATURE_AWK_LIBM
static const char EMSG_NO_MATH[] ALIGN1 = "Math support is not compiled in";
-#endif
+static const char EMSG_NEGATIVE_FIELD[] ALIGN1 = "Access to negative field";
static void zero_out_var(var *vp)
{
pps = *s;
if (c == '\\')
c = bb_process_escape_sequence((const char**)s);
- if (c == '\\' && *s == pps)
- c = *(*s)++;
+ /* Example awk statement:
+ * s = "abc\"def"
+ * we must treat \" as "
+ */
+ if (c == '\\' && *s == pps) { /* unrecognized \z? */
+ c = *(*s); /* yes, fetch z */
+ if (c)
+ (*s)++; /* advance unless z = NUL */
+ }
return c;
}
+/* TODO: merge with strcpy_and_process_escape_sequences()?
+ */
+static void unescape_string_in_place(char *s1)
+{
+ char *s = s1;
+ while ((*s1 = nextchar(&s)) != '\0')
+ s1++;
+}
+
static ALWAYS_INLINE int isalnum_(int c)
{
return (isalnum(c) || c == '_');
static double my_strtod(char **pp)
{
-#if ENABLE_DESKTOP
- if ((*pp)[0] == '0'
- && ((((*pp)[1] | 0x20) == 'x') || isdigit((*pp)[1]))
- ) {
- return strtoull(*pp, pp, 0);
+ char *cp = *pp;
+ if (ENABLE_DESKTOP && cp[0] == '0') {
+ /* Might be hex or octal integer: 0x123abc or 07777 */
+ char c = (cp[1] | 0x20);
+ if (c == 'x' || isdigit(cp[1])) {
+ unsigned long long ull = strtoull(cp, pp, 0);
+ if (c == 'x')
+ return ull;
+ c = **pp;
+ if (!isdigit(c) && c != '.')
+ return ull;
+ /* else: it may be a floating number. Examples:
+ * 009.123 (*pp points to '9')
+ * 000.123 (*pp points to '.')
+ * fall through to strtod.
+ */
+ }
}
-#endif
- return strtod(*pp, pp);
+ return strtod(cp, pp);
}
/* -------- working with variables (set/get/copy/etc) -------- */
v->number = 0;
s = v->string;
if (s && *s) {
+ debug_printf_eval("getvar_i: '%s'->", s);
v->number = my_strtod(&s);
+ debug_printf_eval("%f (s:'%s')\n", v->number, s);
if (v->type & VF_USER) {
s = skip_spaces(s);
if (*s != '\0')
v->type &= ~VF_USER;
}
} else {
+ debug_printf_eval("getvar_i: '%s'->zero\n", s);
v->type &= ~VF_USER;
}
v->type |= VF_CACHED;
}
+ debug_printf_eval("getvar_i: %f\n", v->number);
return v->number;
}
if (dest != src) {
clrvar(dest);
dest->type |= (src->type & ~(VF_DONTTOUCH | VF_FSTR));
+ debug_printf_eval("copyvar: number:%f string:'%s'\n", src->number, src->string);
dest->number = src->number;
if (src->string)
dest->string = xstrdup(src->string);
const char *tl;
uint32_t tc;
const uint32_t *ti;
- int l;
if (t_rollback) {
t_rollback = FALSE;
-
} else if (concat_inserted) {
concat_inserted = FALSE;
t_tclass = save_tclass;
t_info = save_info;
-
} else {
p = g_pos;
readnext:
if (*p == '\0') {
tc = TC_EOF;
-
+ debug_printf_parse("%s: token found: TC_EOF\n", __func__);
} else if (*p == '\"') {
/* it's a string */
t_string = s = ++p;
while (*p != '\"') {
- char *pp = p;
+ char *pp;
if (*p == '\0' || *p == '\n')
syntax_error(EMSG_UNEXP_EOS);
+ pp = p;
*s++ = nextchar(&pp);
p = pp;
}
p++;
*s = '\0';
tc = TC_STRING;
-
+ debug_printf_parse("%s: token found:'%s' TC_STRING\n", __func__, t_string);
} else if ((expected & TC_REGEXP) && *p == '/') {
/* it's regexp */
t_string = s = ++p;
p++;
*s = '\0';
tc = TC_REGEXP;
+ debug_printf_parse("%s: token found:'%s' TC_REGEXP\n", __func__, t_string);
} else if (*p == '.' || isdigit(*p)) {
/* it's a number */
char *pp = p;
t_double = my_strtod(&pp);
p = pp;
- if (*pp == '.')
+ if (*p == '.')
syntax_error(EMSG_UNEXP_TOKEN);
tc = TC_NUMBER;
-
+ debug_printf_parse("%s: token found:%f TC_NUMBER\n", __func__, t_double);
} else {
/* search for something known */
tl = tokenlist;
tc = 0x00000001;
ti = tokeninfo;
while (*tl) {
- l = *tl++;
- if (l == NTCC) {
+ int l = (unsigned char) *tl++;
+ if (l == (unsigned char) NTCC) {
tc <<= 1;
continue;
}
- /* if token class is expected, token
- * matches and it's not a longer word,
- * then this is what we are looking for
+ /* if token class is expected,
+ * token matches,
+ * and it's not a longer word,
*/
if ((tc & (expected | TC_WORD | TC_NEWLINE))
- && *tl == *p && strncmp(p, tl, l) == 0
+ && strncmp(p, tl, l) == 0
&& !((tc & TC_WORD) && isalnum_(p[l]))
) {
+ /* then this is what we are looking for */
t_info = *ti;
+ debug_printf_parse("%s: token found:'%.*s' t_info:%x\n", __func__, l, p, t_info);
p += l;
- break;
+ goto token_found;
}
ti++;
tl += l;
}
-
- if (!*tl) {
- /* it's a name (var/array/function),
- * otherwise it's something wrong
- */
- if (!isalnum_(*p))
- syntax_error(EMSG_UNEXP_TOKEN);
-
- t_string = --p;
- while (isalnum_(*++p)) {
- p[-1] = *p;
- }
- p[-1] = '\0';
- tc = TC_VARIABLE;
- /* also consume whitespace between functionname and bracket */
- if (!(expected & TC_VARIABLE) || (expected & TC_ARRAY))
- p = skip_spaces(p);
- if (*p == '(') {
- tc = TC_FUNCTION;
- } else {
- if (*p == '[') {
- p++;
- tc = TC_ARRAY;
- }
- }
+ /* not a known token */
+
+ /* is it a name? (var/array/function) */
+ if (!isalnum_(*p))
+ syntax_error(EMSG_UNEXP_TOKEN); /* no */
+ /* yes */
+ t_string = --p;
+ while (isalnum_(*++p)) {
+ p[-1] = *p;
+ }
+ p[-1] = '\0';
+ tc = TC_VARIABLE;
+ /* also consume whitespace between functionname and bracket */
+ if (!(expected & TC_VARIABLE) || (expected & TC_ARRAY))
+ p = skip_spaces(p);
+ if (*p == '(') {
+ tc = TC_FUNCTION;
+ debug_printf_parse("%s: token found:'%s' TC_FUNCTION\n", __func__, t_string);
+ } else {
+ if (*p == '[') {
+ p++;
+ tc = TC_ARRAY;
+ debug_printf_parse("%s: token found:'%s' TC_ARRAY\n", __func__, t_string);
+ } else
+ debug_printf_parse("%s: token found:'%s' TC_VARIABLE\n", __func__, t_string);
}
}
+ token_found:
g_pos = p;
/* skipping newlines in some cases */
ltclass = t_tclass;
/* Are we ready for this? */
- if (!(ltclass & expected))
+ if (!(ltclass & expected)) {
syntax_error((ltclass & (TC_NEWLINE | TC_EOF)) ?
EMSG_UNEXP_EOS : EMSG_UNEXP_TOKEN);
+ }
return ltclass;
#undef concat_inserted
uint32_t tc, xtc;
var *v;
+ debug_printf_parse("%s(%x)\n", __func__, iexp);
+
sn.info = PRIMASK;
sn.r.n = glptr = NULL;
xtc = TC_OPERAND | TC_UOPPRE | TC_REGEXP | iexp;
while (!((tc = next_token(xtc)) & iexp)) {
+
if (glptr && (t_info == (OC_COMPARE | VV | P(39) | 2))) {
/* input redirection (<) attached to glptr node */
+ debug_printf_parse("%s: input redir\n", __func__);
cn = glptr->l.n = new_node(OC_CONCAT | SS | P(37));
cn->a.n = glptr;
xtc = TC_OPERAND | TC_UOPPRE;
glptr = NULL;
} else if (tc & (TC_BINOP | TC_UOPPOST)) {
+ debug_printf_parse("%s: TC_BINOP | TC_UOPPOST\n", __func__);
/* for binary and postfix-unary operators, jump back over
* previous operators with higher priority */
vn = cn;
vn->a.n = cn;
} else {
+ debug_printf_parse("%s: other\n", __func__);
/* for operands and prefix-unary operators, attach them
* to last node */
vn = cn;
cn->a.n = vn;
xtc = TC_OPERAND | TC_UOPPRE | TC_REGEXP;
if (tc & (TC_OPERAND | TC_REGEXP)) {
+ debug_printf_parse("%s: TC_OPERAND | TC_REGEXP\n", __func__);
xtc = TC_UOPPRE | TC_UOPPOST | TC_BINOP | TC_OPERAND | iexp;
/* one should be very careful with switch on tclass -
* only simple tclasses should be used! */
switch (tc) {
case TC_VARIABLE:
case TC_ARRAY:
+ debug_printf_parse("%s: TC_VARIABLE | TC_ARRAY\n", __func__);
cn->info = OC_VAR;
v = hash_search(ahash, t_string);
if (v != NULL) {
case TC_NUMBER:
case TC_STRING:
+ debug_printf_parse("%s: TC_NUMBER | TC_STRING\n", __func__);
cn->info = OC_VAR;
v = cn->l.v = xzalloc(sizeof(var));
if (tc & TC_NUMBER)
break;
case TC_REGEXP:
+ debug_printf_parse("%s: TC_REGEXP\n", __func__);
mk_re_node(t_string, cn, xzalloc(sizeof(regex_t)*2));
break;
case TC_FUNCTION:
+ debug_printf_parse("%s: TC_FUNCTION\n", __func__);
cn->info = OC_FUNC;
cn->r.f = newfunc(t_string);
cn->l.n = condition();
break;
case TC_SEQSTART:
+ debug_printf_parse("%s: TC_SEQSTART\n", __func__);
cn = vn->r.n = parse_expr(TC_SEQTERM);
+ if (!cn)
+ syntax_error("Empty sequence");
cn->a.n = vn;
break;
case TC_GETLINE:
+ debug_printf_parse("%s: TC_GETLINE\n", __func__);
glptr = cn;
xtc = TC_OPERAND | TC_UOPPRE | TC_BINOP | iexp;
break;
case TC_BUILTIN:
+ debug_printf_parse("%s: TC_BUILTIN\n", __func__);
cn->l.n = condition();
break;
+
+ case TC_LENGTH:
+ debug_printf_parse("%s: TC_LENGTH\n", __func__);
+ next_token(TC_SEQSTART | TC_OPTERM | TC_GRPTERM);
+ rollback_token();
+ if (t_tclass & TC_SEQSTART) {
+ /* It was a "(" token. Handle just like TC_BUILTIN */
+ cn->l.n = condition();
+ }
+ break;
}
}
}
}
+
+ debug_printf_parse("%s() returns %p\n", __func__, sn.r.n);
return sn.r.n;
}
} while (c & TC_NEWLINE);
if (c & TC_GRPSTART) {
+ debug_printf_parse("%s: TC_GRPSTART\n", __func__);
while (next_token(TC_GRPSEQ | TC_GRPTERM) != TC_GRPTERM) {
+ debug_printf_parse("%s: !TC_GRPTERM\n", __func__);
if (t_tclass & TC_NEWLINE)
continue;
rollback_token();
chain_group();
}
+ debug_printf_parse("%s: TC_GRPTERM\n", __func__);
} else if (c & (TC_OPSEQ | TC_OPTERM)) {
+ debug_printf_parse("%s: TC_OPSEQ | TC_OPTERM\n", __func__);
rollback_token();
chain_expr(OC_EXEC | Vx);
- } else { /* TC_STATEMNT */
+ } else {
+ /* TC_STATEMNT */
+ debug_printf_parse("%s: TC_STATEMNT(?)\n", __func__);
switch (t_info & OPCLSMASK) {
case ST_IF:
+ debug_printf_parse("%s: ST_IF\n", __func__);
n = chain_node(OC_BR | Vx);
n->l.n = condition();
chain_group();
break;
case ST_WHILE:
+ debug_printf_parse("%s: ST_WHILE\n", __func__);
n2 = condition();
n = chain_loop(NULL);
n->l.n = n2;
break;
case ST_DO:
+ debug_printf_parse("%s: ST_DO\n", __func__);
n2 = chain_node(OC_EXEC);
n = chain_loop(NULL);
n2->a.n = n->a.n;
break;
case ST_FOR:
+ debug_printf_parse("%s: ST_FOR\n", __func__);
next_token(TC_SEQSTART);
n2 = parse_expr(TC_SEMICOL | TC_SEQTERM);
if (t_tclass & TC_SEQTERM) { /* for-in */
- if ((n2->info & OPCLSMASK) != OC_IN)
+ if (!n2 || (n2->info & OPCLSMASK) != OC_IN)
syntax_error(EMSG_UNEXP_TOKEN);
n = chain_node(OC_WALKINIT | VV);
n->l.n = n2->l.n;
case OC_PRINT:
case OC_PRINTF:
+ debug_printf_parse("%s: OC_PRINT[F]\n", __func__);
n = chain_node(t_info);
n->l.n = parse_expr(TC_OPTERM | TC_OUTRDR | TC_GRPTERM);
if (t_tclass & TC_OUTRDR) {
break;
case OC_BREAK:
+ debug_printf_parse("%s: OC_BREAK\n", __func__);
n = chain_node(OC_EXEC);
n->a.n = break_ptr;
+ chain_expr(t_info);
break;
case OC_CONTINUE:
+ debug_printf_parse("%s: OC_CONTINUE\n", __func__);
n = chain_node(OC_EXEC);
n->a.n = continue_ptr;
+ chain_expr(t_info);
break;
/* delete, next, nextfile, return, exit */
default:
+ debug_printf_parse("%s: default\n", __func__);
chain_expr(t_info);
}
}
while ((tclass = next_token(TC_EOF | TC_OPSEQ | TC_GRPSTART |
TC_OPTERM | TC_BEGIN | TC_END | TC_FUNCDECL)) != TC_EOF) {
- if (tclass & TC_OPTERM)
+ if (tclass & TC_OPTERM) {
+ debug_printf_parse("%s: TC_OPTERM\n", __func__);
continue;
+ }
seq = &mainseq;
if (tclass & TC_BEGIN) {
+ debug_printf_parse("%s: TC_BEGIN\n", __func__);
seq = &beginseq;
chain_group();
-
} else if (tclass & TC_END) {
+ debug_printf_parse("%s: TC_END\n", __func__);
seq = &endseq;
chain_group();
-
} else if (tclass & TC_FUNCDECL) {
+ debug_printf_parse("%s: TC_FUNCDECL\n", __func__);
next_token(TC_FUNCTION);
g_pos++;
f = newfunc(t_string);
seq = &f->body;
chain_group();
clear_array(ahash);
-
} else if (tclass & TC_OPSEQ) {
+ debug_printf_parse("%s: TC_OPSEQ\n", __func__);
rollback_token();
cn = chain_node(OC_TEST);
cn->l.n = parse_expr(TC_OPTERM | TC_EOF | TC_GRPSTART);
if (t_tclass & TC_GRPSTART) {
+ debug_printf_parse("%s: TC_GRPSTART\n", __func__);
rollback_token();
chain_group();
} else {
+ debug_printf_parse("%s: !TC_GRPSTART\n", __func__);
chain_node(OC_PRINT);
}
cn->r.n = mainseq.last;
-
} else /* if (tclass & TC_GRPSTART) */ {
+ debug_printf_parse("%s: TC_GRPSTART(?)\n", __func__);
rollback_token();
chain_group();
}
}
+ debug_printf_parse("%s: TC_EOF\n", __func__);
}
regfree(re);
regfree(ire); // TODO: nuke ire, use re+1?
}
- if (strlen(s) > 1) {
+ if (s[0] && s[1]) { /* strlen(s) > 1 */
mk_re_node(s, n, re);
} else {
- n->info = (uint32_t) *s;
+ n->info = (uint32_t) s[0];
}
return n;
if (size >= maxfields) {
i = maxfields;
maxfields = size + 16;
- Fields = xrealloc(Fields, maxfields * sizeof(var));
+ Fields = xrealloc(Fields, maxfields * sizeof(Fields[0]));
for (; i < maxfields; i++) {
Fields[i].type = VF_SPECIAL;
Fields[i].string = NULL;
}
}
-
- if (size < nfields) {
- for (i = size; i < nfields; i++) {
- clrvar(Fields + i);
- }
+ /* if size < nfields, clear extra field variables */
+ for (i = size; i < nfields; i++) {
+ clrvar(Fields + i);
}
nfields = size;
}
static int awk_split(const char *s, node *spl, char **slist)
{
- int l, n = 0;
+ int l, n;
char c[4];
char *s1;
regmatch_t pmatch[2]; // TODO: why [2]? [1] is enough...
if (*getvar_s(intvar[RS]) == '\0')
c[2] = '\n';
+ n = 0;
if ((spl->info & OPCLSMASK) == OC_REGEXP) { /* regex split */
if (!*s)
return n; /* "": zero fields */
}
if (*s1)
n++;
- while ((s1 = strpbrk(s1, c))) {
+ while ((s1 = strpbrk(s1, c)) != NULL) {
*s1++ = '\0';
n++;
}
is_f0_split = FALSE;
} else if (v == intvar[FS]) {
- mk_splitter(getvar_s(v), &fsplitter);
+ /*
+ * The POSIX-2008 standard says that changing FS should have no effect on the
+ * current input line, but only on the next one. The language is:
+ *
+ * > Before the first reference to a field in the record is evaluated, the record
+ * > shall be split into fields, according to the rules in Regular Expressions,
+ * > using the value of FS that was current at the time the record was read.
+ *
+ * So, split up current line before assignment to FS:
+ */
+ split_f0();
+ mk_splitter(getvar_s(v), &fsplitter);
} else if (v == intvar[RS]) {
mk_splitter(getvar_s(v), &rsplitter);
-
} else if (v == intvar[IGNORECASE]) {
icase = istrue(v);
-
} else { /* $n */
n = getvar_i(intvar[NF]);
setvar_i(intvar[NF], n > v-Fields ? n : v-Fields+1);
int fd, so, eo, r, rp;
char c, *m, *s;
+ debug_printf_eval("entered %s()\n", __func__);
+
/* we're using our own buffer since we need access to accumulating
* characters
*/
rsm->pos = p - eo;
rsm->size = size;
+ debug_printf_eval("returning from %s(): %d\n", __func__, r);
+
return r;
}
char c;
const char *s = format;
- if (int_as_int && n == (int)n) {
- r = snprintf(b, size, "%d", (int)n);
+ if (int_as_int && n == (long long)n) {
+ r = snprintf(b, size, "%lld", (long long)n);
} else {
do { c = *s; } while (c && *++s);
if (strchr("diouxX", c)) {
//gsub(/\<b*/,"") on "abc" will reach this point, advance to "bc"
//... and will erroneously match "b" even though it is NOT at the word start.
//we need REG_NOTBOW but it does not exist...
+//TODO: if EXTRA_COMPAT=y, use GNU matching and re_search,
+//it should be able to do it correctly.
/* Subtle: this is safe only because
* qrealloc allocated at least one extra byte */
resbuf[residx] = *sp;
switch (info) {
case B_a2:
-#if ENABLE_FEATURE_AWK_LIBM
- setvar_i(res, atan2(getvar_i(av[0]), getvar_i(av[1])));
-#else
- syntax_error(EMSG_NO_MATH);
-#endif
+ if (ENABLE_FEATURE_AWK_LIBM)
+ setvar_i(res, atan2(getvar_i(av[0]), getvar_i(av[1])));
+ else
+ syntax_error(EMSG_NO_MATH);
break;
case B_sp: {
#define fnargs (G.evaluate__fnargs)
/* seed is initialized to 1 */
#define seed (G.evaluate__seed)
-#define sreg (G.evaluate__sreg)
+#define sreg (G.evaluate__sreg)
var *v1;
if (!op)
return setvar_s(res, NULL);
+ debug_printf_eval("entered %s()\n", __func__);
+
v1 = nvalloc(2);
while (op) {
opn = (opinfo & OPNMASK);
g_lineno = op->lineno;
op1 = op->l.n;
+ debug_printf_eval("opinfo:%08x opn:%08x\n", opinfo, opn);
+
+ /* "delete" is special:
+ * "delete array[var--]" must evaluate index expr only once,
+ * must not evaluate it in "execute inevitable things" part.
+ */
+ if (XC(opinfo & OPCLSMASK) == XC(OC_DELETE)) {
+ uint32_t info = op1->info & OPCLSMASK;
+ var *v;
+
+ debug_printf_eval("DELETE\n");
+ if (info == OC_VAR) {
+ v = op1->l.v;
+ } else if (info == OC_FNARG) {
+ v = &fnargs[op1->l.aidx];
+ } else {
+ syntax_error(EMSG_NOT_ARRAY);
+ }
+ if (op1->r.n) { /* array ref? */
+ const char *s;
+ s = getvar_s(evaluate(op1->r.n, v1));
+ hash_remove(iamarray(v), s);
+ } else {
+ clear_array(iamarray(v));
+ }
+ goto next;
+ }
/* execute inevitable things */
if (opinfo & OF_RES1)
L.v = evaluate(op1, v1);
if (opinfo & OF_RES2)
R.v = evaluate(op->r.n, v1+1);
- if (opinfo & OF_STR1)
+ if (opinfo & OF_STR1) {
L.s = getvar_s(L.v);
- if (opinfo & OF_STR2)
+ debug_printf_eval("L.s:'%s'\n", L.s);
+ }
+ if (opinfo & OF_STR2) {
R.s = getvar_s(R.v);
- if (opinfo & OF_NUM1)
+ debug_printf_eval("R.s:'%s'\n", R.s);
+ }
+ if (opinfo & OF_NUM1) {
L_d = getvar_i(L.v);
+ debug_printf_eval("L_d:%f\n", L_d);
+ }
+ debug_printf_eval("switch(0x%x)\n", XC(opinfo & OPCLSMASK));
switch (XC(opinfo & OPCLSMASK)) {
/* -- iterative node type -- */
break;
}
- case XC( OC_DELETE ): {
- uint32_t info = op1->info & OPCLSMASK;
- var *v;
-
- if (info == OC_VAR) {
- v = op1->l.v;
- } else if (info == OC_FNARG) {
- v = &fnargs[op1->l.aidx];
- } else {
- syntax_error(EMSG_NOT_ARRAY);
- }
-
- if (op1->r.n) {
- const char *s;
- clrvar(L.v);
- s = getvar_s(evaluate(op1->r.n, v1));
- hash_remove(iamarray(v), s);
- } else {
- clear_array(iamarray(v));
- }
- break;
- }
+ /* case XC( OC_DELETE ): - moved to happen before arg evaluation */
case XC( OC_NEWSOURCE ):
g_progname = op->l.new_progname;
/* -- recursive node type -- */
case XC( OC_VAR ):
+ debug_printf_eval("VAR\n");
L.v = op->l.v;
if (L.v == intvar[NF])
split_f0();
goto v_cont;
case XC( OC_FNARG ):
+ debug_printf_eval("FNARG[%d]\n", op->l.aidx);
L.v = &fnargs[op->l.aidx];
v_cont:
res = op->r.n ? findvar(iamarray(L.v), R.s) : L.v;
break;
case XC( OC_MOVE ):
+ debug_printf_eval("MOVE\n");
/* if source is a temporary string, jusk relink it to dest */
//Disabled: if R.v is numeric but happens to have cached R.v->string,
//then L.v ends up being a string, which is wrong
var *vbeg, *v;
const char *sv_progname;
- if (!op->r.f->body.first)
+ /* The body might be empty, still has to eval the args */
+ if (!op->r.n->info && !op->r.f->body.first)
syntax_error(EMSG_UNDEF_FUNC);
vbeg = v = nvalloc(op->r.f->nargs + 1);
rsm->F = popen(L.s, "r");
rsm->is_pipe = TRUE;
} else {
- rsm->F = fopen_for_read(L.s); /* not xfopen! */
+ rsm->F = fopen_for_read(L.s); /* not xfopen! */
}
}
} else {
rsm = iF;
}
- if (!rsm->F) {
+ if (!rsm || !rsm->F) {
setvar_i(intvar[ERRNO], errno);
setvar_i(res, -1);
break;
/* simple builtins */
case XC( OC_FBLTIN ): {
- int i;
- rstream *rsm;
double R_d = R_d; /* for compiler */
switch (opn) {
case F_in:
- R_d = (int)L_d;
+ R_d = (long long)L_d;
break;
case F_rn:
R_d = (double)rand() / (double)RAND_MAX;
break;
-#if ENABLE_FEATURE_AWK_LIBM
+
case F_co:
- R_d = cos(L_d);
- break;
+ if (ENABLE_FEATURE_AWK_LIBM) {
+ R_d = cos(L_d);
+ break;
+ }
case F_ex:
- R_d = exp(L_d);
- break;
+ if (ENABLE_FEATURE_AWK_LIBM) {
+ R_d = exp(L_d);
+ break;
+ }
case F_lg:
- R_d = log(L_d);
- break;
+ if (ENABLE_FEATURE_AWK_LIBM) {
+ R_d = log(L_d);
+ break;
+ }
case F_si:
- R_d = sin(L_d);
- break;
+ if (ENABLE_FEATURE_AWK_LIBM) {
+ R_d = sin(L_d);
+ break;
+ }
case F_sq:
- R_d = sqrt(L_d);
- break;
-#else
- case F_co:
- case F_ex:
- case F_lg:
- case F_si:
- case F_sq:
+ if (ENABLE_FEATURE_AWK_LIBM) {
+ R_d = sqrt(L_d);
+ break;
+ }
+
syntax_error(EMSG_NO_MATH);
break;
-#endif
+
case F_sr:
R_d = (double)seed;
seed = op1 ? (unsigned)L_d : (unsigned)time(NULL);
break;
case F_le:
- if (!op1)
+ debug_printf_eval("length: L.s:'%s'\n", L.s);
+ if (!op1) {
L.s = getvar_s(intvar[F0]);
+ debug_printf_eval("length: L.s='%s'\n", L.s);
+ }
+ else if (L.v->type & VF_ARRAY) {
+ R_d = L.v->x.array->nel;
+ debug_printf_eval("length: array_len:%d\n", L.v->x.array->nel);
+ break;
+ }
R_d = strlen(L.s);
break;
if (!op1) {
fflush(stdout);
} else if (L.s && *L.s) {
- rsm = newfile(L.s);
+ rstream *rsm = newfile(L.s);
fflush(rsm->F);
} else {
fflush_all();
}
break;
- case F_cl:
- i = 0;
+ case F_cl: {
+ rstream *rsm;
+ int err = 0;
rsm = (rstream *)hash_search(fdhash, L.s);
+ debug_printf_eval("OC_FBLTIN F_cl rsm:%p\n", rsm);
if (rsm) {
- i = rsm->is_pipe ? pclose(rsm->F) : fclose(rsm->F);
+ debug_printf_eval("OC_FBLTIN F_cl "
+ "rsm->is_pipe:%d, ->F:%p\n",
+ rsm->is_pipe, rsm->F);
+ /* Can be NULL if open failed. Example:
+ * getline line <"doesnt_exist";
+ * close("doesnt_exist"); <--- here rsm->F is NULL
+ */
+ if (rsm->F)
+ err = rsm->is_pipe ? pclose(rsm->F) : fclose(rsm->F);
free(rsm->buffer);
hash_remove(fdhash, L.s);
}
- if (i != 0)
+ if (err)
setvar_i(intvar[ERRNO], errno);
- R_d = (double)i;
+ R_d = (double)err;
break;
}
+ } /* switch */
setvar_i(res, R_d);
break;
}
case XC( OC_FIELD ): {
int i = (int)getvar_i(R.v);
+ if (i < 0)
+ syntax_error(EMSG_NEGATIVE_FIELD);
if (i == 0) {
res = intvar[F0];
} else {
case XC( OC_BINARY ):
case XC( OC_REPLACE ): {
double R_d = getvar_i(R.v);
+ debug_printf_eval("BINARY/REPLACE: R_d:%f opn:%c\n", R_d, opn);
switch (opn) {
case '+':
L_d += R_d;
L_d /= R_d;
break;
case '&':
-#if ENABLE_FEATURE_AWK_LIBM
- L_d = pow(L_d, R_d);
-#else
- syntax_error(EMSG_NO_MATH);
-#endif
+ if (ENABLE_FEATURE_AWK_LIBM)
+ L_d = pow(L_d, R_d);
+ else
+ syntax_error(EMSG_NO_MATH);
break;
case '%':
if (R_d == 0)
syntax_error(EMSG_DIV_BY_ZERO);
- L_d -= (int)(L_d / R_d) * R_d;
+ L_d -= (long long)(L_d / R_d) * R_d;
break;
}
+ debug_printf_eval("BINARY/REPLACE result:%f\n", L_d);
res = setvar_i(((opinfo & OPCLSMASK) == OC_BINARY) ? res : L.v, L_d);
break;
}
default:
syntax_error(EMSG_POSSIBLE_ERROR);
- }
+ } /* switch */
+ next:
if ((opinfo & OPCLSMASK) <= SHIFT_TIL_THIS)
op = op->a.n;
if ((opinfo & OPCLSMASK) >= RECUR_FROM_THIS)
} /* while (op) */
nvfree(v1);
+ debug_printf_eval("returning from %s(): %p\n", __func__, res);
return res;
#undef fnargs
#undef seed
* otherwise return 0 */
static int is_assignment(const char *expr)
{
- char *exprc, *s, *s0, *s1;
+ char *exprc, *val;
- exprc = xstrdup(expr);
- if (!isalnum_(*exprc) || (s = strchr(exprc, '=')) == NULL) {
- free(exprc);
+ if (!isalnum_(*expr) || (val = strchr(expr, '=')) == NULL) {
return FALSE;
}
- *s++ = '\0';
- s0 = s1 = s;
- while (*s)
- *s1++ = nextchar(&s);
+ exprc = xstrdup(expr);
+ val = exprc + (val - expr);
+ *val++ = '\0';
- *s1 = '\0';
- setvar_u(newvar(exprc), s0);
+ unescape_string_in_place(val);
+ setvar_u(newvar(exprc), val);
free(exprc);
return TRUE;
}
#define rsm (G.next_input_file__rsm)
#define files_happen (G.next_input_file__files_happen)
- FILE *F = NULL;
+ FILE *F;
const char *fname, *ind;
if (rsm.F)
rsm.F = NULL;
rsm.pos = rsm.adv = 0;
- do {
+ for (;;) {
if (getvar_i(intvar[ARGIND])+1 >= getvar_i(intvar[ARGC])) {
if (files_happen)
return NULL;
fname = "-";
F = stdin;
- } else {
- ind = getvar_s(incvar(intvar[ARGIND]));
- fname = getvar_s(findvar(iamarray(intvar[ARGV]), ind));
- if (fname && *fname && !is_assignment(fname))
- F = xfopen_stdin(fname);
+ break;
+ }
+ ind = getvar_s(incvar(intvar[ARGIND]));
+ fname = getvar_s(findvar(iamarray(intvar[ARGV]), ind));
+ if (fname && *fname && !is_assignment(fname)) {
+ F = xfopen_stdin(fname);
+ break;
}
- } while (!F);
+ }
files_happen = TRUE;
setvar_s(intvar[FILENAME], fname);
}
int awk_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
-int awk_main(int argc, char **argv)
+int awk_main(int argc UNUSED_PARAM, char **argv)
{
unsigned opt;
- char *opt_F, *opt_W;
+ char *opt_F;
llist_t *list_v = NULL;
llist_t *list_f = NULL;
+#if ENABLE_FEATURE_AWK_GNU_EXTENSIONS
+ llist_t *list_e = NULL;
+#endif
int i, j;
var *v;
var tv;
*s1 = '=';
}
}
- opt_complementary = "v::f::"; /* -v and -f can occur multiple times */
- opt = getopt32(argv, "F:v:f:W:", &opt_F, &list_v, &list_f, &opt_W);
+ opt = getopt32(argv, OPTSTR_AWK, &opt_F, &list_v, &list_f, IF_FEATURE_AWK_GNU_EXTENSIONS(&list_e,) NULL);
argv += optind;
- argc -= optind;
- if (opt & 0x1)
- setvar_s(intvar[FS], opt_F); // -F
- while (list_v) { /* -v */
+ //argc -= optind;
+ if (opt & OPT_W)
+ bb_error_msg("warning: option -W is ignored");
+ if (opt & OPT_F) {
+ unescape_string_in_place(opt_F);
+ setvar_s(intvar[FS], opt_F);
+ }
+ while (list_v) {
if (!is_assignment(llist_pop(&list_v)))
bb_show_usage();
}
- if (list_f) { /* -f */
- do {
- char *s = NULL;
- FILE *from_file;
-
- g_progname = llist_pop(&list_f);
- from_file = xfopen_stdin(g_progname);
- /* one byte is reserved for some trick in next_token */
- for (i = j = 1; j > 0; i += j) {
- s = xrealloc(s, i + 4096);
- j = fread(s + i, 1, 4094, from_file);
- }
- s[i] = '\0';
- fclose(from_file);
- parse_program(s + 1);
- free(s);
- } while (list_f);
- argc++;
- } else { // no -f: take program from 1st parameter
- if (!argc)
+ while (list_f) {
+ char *s = NULL;
+ FILE *from_file;
+
+ g_progname = llist_pop(&list_f);
+ from_file = xfopen_stdin(g_progname);
+ /* one byte is reserved for some trick in next_token */
+ for (i = j = 1; j > 0; i += j) {
+ s = xrealloc(s, i + 4096);
+ j = fread(s + i, 1, 4094, from_file);
+ }
+ s[i] = '\0';
+ fclose(from_file);
+ parse_program(s + 1);
+ free(s);
+ }
+ g_progname = "cmd. line";
+#if ENABLE_FEATURE_AWK_GNU_EXTENSIONS
+ while (list_e) {
+ parse_program(llist_pop(&list_e));
+ }
+#endif
+ if (!(opt & (OPT_f | OPT_e))) {
+ if (!*argv)
bb_show_usage();
- g_progname = "cmd. line";
parse_program(*argv++);
}
- if (opt & 0x8) // -W
- bb_error_msg("warning: unrecognized option '-W %s' ignored", opt_W);
/* fill in ARGV array */
- setvar_i(intvar[ARGC], argc);
setari_u(intvar[ARGV], 0, "awk");
i = 0;
while (*argv)
setari_u(intvar[ARGV], ++i, *argv++);
+ setvar_i(intvar[ARGC], i + 1);
evaluate(beginseq.first, &tv);
if (!mainseq.first && !endseq.first)