aboutsummaryrefslogtreecommitdiffstats
path: root/builtin.c
diff options
context:
space:
mode:
Diffstat (limited to 'builtin.c')
-rw-r--r--builtin.c289
1 files changed, 200 insertions, 89 deletions
diff --git a/builtin.c b/builtin.c
index 3d59e359..b70176dc 100644
--- a/builtin.c
+++ b/builtin.c
@@ -3,7 +3,7 @@
*/
/*
- * Copyright (C) 1986, 1988, 1989, 1991-2014 the Free Software Foundation, Inc.
+ * Copyright (C) 1986, 1988, 1989, 1991-2015 the Free Software Foundation, Inc.
*
* This file is part of GAWK, the GNU implementation of the
* AWK Programming Language.
@@ -129,10 +129,14 @@ wrerror:
if (fp == stdout && errno == EPIPE)
gawk_exit(EXIT_FATAL);
+
/* otherwise die verbosely */
- fatal(_("%s to \"%s\" failed (%s)"), from,
- rp ? rp->value : _("standard output"),
- errno ? strerror(errno) : _("reason unknown"));
+ if ((rp != NULL) ? is_non_fatal_redirect(rp->value) : is_non_fatal_std(fp))
+ update_ERRNO_int(errno);
+ else
+ fatal(_("%s to \"%s\" failed (%s)"), from,
+ rp ? rp->value : _("standard output"),
+ errno ? strerror(errno) : _("reason unknown"));
}
/* do_exp --- exponential function */
@@ -247,7 +251,6 @@ do_fflush(int nargs)
return make_number((AWKNUM) status);
}
-#if MBS_SUPPORT
/* strncasecmpmbs --- like strncasecmp (multibyte string version) */
int
@@ -327,14 +330,6 @@ index_multibyte_buffer(char* src, char* dest, int len)
dest[idx] = mbclen;
}
}
-#else
-/* a dummy function */
-static void
-index_multibyte_buffer(char* src ATTRIBUTE_UNUSED, char* dest ATTRIBUTE_UNUSED, int len ATTRIBUTE_UNUSED)
-{
- cant_happen();
-}
-#endif
/* do_index --- find index of a string */
@@ -345,7 +340,6 @@ do_index(int nargs)
const char *p1, *p2;
size_t l1, l2;
long ret;
-#if MBS_SUPPORT
bool do_single_byte = false;
mbstate_t mbs1, mbs2;
@@ -353,7 +347,6 @@ do_index(int nargs)
memset(& mbs1, 0, sizeof(mbstate_t));
memset(& mbs2, 0, sizeof(mbstate_t));
}
-#endif
POP_TWO_SCALARS(s1, s2);
@@ -383,7 +376,6 @@ do_index(int nargs)
goto out;
}
-#if MBS_SUPPORT
if (gawk_mb_cur_max > 1) {
s1 = force_wstring(s1);
s2 = force_wstring(s2);
@@ -394,14 +386,12 @@ do_index(int nargs)
do_single_byte = ((s1->wstlen == 0 && s1->stlen > 0)
|| (s2->wstlen == 0 && s2->stlen > 0));
}
-#endif
/* IGNORECASE will already be false if posix */
if (IGNORECASE) {
while (l1 > 0) {
if (l2 > l1)
break;
-#if MBS_SUPPORT
if (! do_single_byte && gawk_mb_cur_max > 1) {
const wchar_t *pos;
@@ -412,21 +402,18 @@ do_index(int nargs)
ret = pos - s1->wstptr + 1; /* 1-based */
goto out;
} else {
-#endif
- /*
- * Could use tolower(*p1) == tolower(*p2) here.
- * See discussion in eval.c as to why not.
- */
- if (casetable[(unsigned char)*p1] == casetable[(unsigned char)*p2]
- && (l2 == 1 || strncasecmp(p1, p2, l2) == 0)) {
- ret = 1 + s1->stlen - l1;
- break;
- }
- l1--;
- p1++;
-#if MBS_SUPPORT
+ /*
+ * Could use tolower(*p1) == tolower(*p2) here.
+ * See discussion in eval.c as to why not.
+ */
+ if (casetable[(unsigned char)*p1] == casetable[(unsigned char)*p2]
+ && (l2 == 1 || strncasecmp(p1, p2, l2) == 0)) {
+ ret = 1 + s1->stlen - l1;
+ break;
+ }
+ l1--;
+ p1++;
}
-#endif
}
} else {
while (l1 > 0) {
@@ -437,7 +424,6 @@ do_index(int nargs)
ret = 1 + s1->stlen - l1;
break;
}
-#if MBS_SUPPORT
if (! do_single_byte && gawk_mb_cur_max > 1) {
const wchar_t *pos;
@@ -451,10 +437,6 @@ do_index(int nargs)
l1--;
p1++;
}
-#else
- l1--;
- p1++;
-#endif
}
}
out:
@@ -532,6 +514,9 @@ do_length(int nargs)
* Support for deferred loading of array elements requires that
* we use the array length interface even though it isn't
* necessary for the built-in array types.
+ *
+ * 1/2015: The deferred arrays are gone, but this is probably
+ * still a good idea.
*/
size = assoc_length(tmp);
@@ -544,7 +529,6 @@ do_length(int nargs)
lintwarn(_("length: received non-string argument"));
tmp = force_string(tmp);
-#if MBS_SUPPORT
if (gawk_mb_cur_max > 1) {
tmp = force_wstring(tmp);
len = tmp->wstlen;
@@ -555,7 +539,6 @@ do_length(int nargs)
if (len == 0 && tmp->stlen > 0)
len = tmp->stlen;
} else
-#endif
len = tmp->stlen;
DEREF(tmp);
@@ -928,7 +911,10 @@ check_pos:
case '*':
if (cur == NULL)
break;
- if (! do_traditional && isdigit((unsigned char) *s1)) {
+ if (! do_traditional && used_dollar && ! isdigit((unsigned char) *s1)) {
+ fatal(_("fatal: must use `count$' on all formats or none"));
+ break; /* silence warnings */
+ } else if (! do_traditional && isdigit((unsigned char) *s1)) {
int val = 0;
for (; n0 > 0 && *s1 && isdigit((unsigned char) *s1); s1++, n0--) {
@@ -1058,7 +1044,6 @@ check_pos:
(void) force_number(arg);
if ((arg->flags & NUMBER) != 0) {
uval = get_number_uj(arg);
-#if MBS_SUPPORT
if (gawk_mb_cur_max > 1) {
char buf[100];
wchar_t wc;
@@ -1099,7 +1084,7 @@ out0:
;
/* else,
fall through */
-#endif
+
cpbuf[0] = uval;
prec = 1;
cp = cpbuf;
@@ -1113,7 +1098,6 @@ out0:
*/
cp = arg->stptr;
prec = 1;
-#if MBS_SUPPORT
/*
* First character can be multiple bytes if
* it's a multibyte character. Grr.
@@ -1131,7 +1115,6 @@ out0:
fw += count - 1;
}
}
-#endif
goto pr_tail;
case 's':
need_format = false;
@@ -1660,7 +1643,7 @@ do_printf(int nargs, int redirtype)
FILE *fp = NULL;
NODE *tmp;
struct redirect *rp = NULL;
- int errflg; /* not used, sigh */
+ int errflg = 0;
NODE *redir_exp = NULL;
if (nargs == 0) {
@@ -1671,7 +1654,7 @@ do_printf(int nargs, int redirtype)
redir_exp = TOP();
if (redir_exp->type != Node_val)
fatal(_("attempt to use array `%s' in a scalar context"), array_vname(redir_exp));
- rp = redirect(redir_exp, redirtype, & errflg);
+ rp = redirect(redir_exp, redirtype, & errflg, true);
DEREF(redir_exp);
decr_sp();
}
@@ -1684,9 +1667,13 @@ do_printf(int nargs, int redirtype)
redir_exp = PEEK(nargs);
if (redir_exp->type != Node_val)
fatal(_("attempt to use array `%s' in a scalar context"), array_vname(redir_exp));
- rp = redirect(redir_exp, redirtype, & errflg);
+ rp = redirect(redir_exp, redirtype, & errflg, true);
if (rp != NULL)
fp = rp->output.fp;
+ else if (errflg) {
+ update_ERRNO_int(errflg);
+ return;
+ }
} else if (do_debug) /* only the debugger can change the default output */
fp = output_fp;
else
@@ -1805,13 +1792,11 @@ do_substr(int nargs)
if (nargs == 2) { /* third arg. missing */
/* use remainder of string */
length = t1->stlen - indx; /* default to bytes */
-#if MBS_SUPPORT
if (gawk_mb_cur_max > 1) {
t1 = force_wstring(t1);
if (t1->wstlen > 0) /* use length of wide char string if we have one */
length = t1->wstlen - indx;
}
-#endif
d_length = length; /* set here in case used in diagnostics, below */
}
@@ -1824,12 +1809,10 @@ do_substr(int nargs)
}
/* get total len of input string, for following checks */
-#if MBS_SUPPORT
if (gawk_mb_cur_max > 1) {
t1 = force_wstring(t1);
src_len = t1->wstlen;
} else
-#endif
src_len = t1->stlen;
if (indx >= src_len) {
@@ -1847,7 +1830,6 @@ do_substr(int nargs)
length = src_len - indx;
}
-#if MBS_SUPPORT
/* force_wstring() already called */
if (gawk_mb_cur_max == 1 || t1->wstlen == t1->stlen)
/* single byte case */
@@ -1877,9 +1859,6 @@ do_substr(int nargs)
*cp = '\0';
r = make_str_node(substr, cp - substr, ALREADY_MALLOCED);
}
-#else
- r = make_string(t1->stptr + indx, length);
-#endif
DEREF(t1);
return r;
@@ -2107,7 +2086,7 @@ void
do_print(int nargs, int redirtype)
{
struct redirect *rp = NULL;
- int errflg; /* not used, sigh */
+ int errflg = 0;
FILE *fp = NULL;
int i;
NODE *redir_exp = NULL;
@@ -2119,9 +2098,13 @@ do_print(int nargs, int redirtype)
redir_exp = PEEK(nargs);
if (redir_exp->type != Node_val)
fatal(_("attempt to use array `%s' in a scalar context"), array_vname(redir_exp));
- rp = redirect(redir_exp, redirtype, & errflg);
+ rp = redirect(redir_exp, redirtype, & errflg, true);
if (rp != NULL)
fp = rp->output.fp;
+ else if (errflg) {
+ update_ERRNO_int(errflg);
+ return;
+ }
} else if (do_debug) /* only the debugger can change the default output */
fp = output_fp;
else
@@ -2177,13 +2160,13 @@ do_print_rec(int nargs, int redirtype)
FILE *fp = NULL;
NODE *f0;
struct redirect *rp = NULL;
- int errflg; /* not used, sigh */
+ int errflg = 0;
NODE *redir_exp = NULL;
assert(nargs == 0);
if (redirtype != 0) {
redir_exp = TOP();
- rp = redirect(redir_exp, redirtype, & errflg);
+ rp = redirect(redir_exp, redirtype, & errflg, true);
if (rp != NULL)
fp = rp->output.fp;
DEREF(redir_exp);
@@ -2191,6 +2174,11 @@ do_print_rec(int nargs, int redirtype)
} else
fp = output_fp;
+ if (errflg) {
+ update_ERRNO_int(errflg);
+ return;
+ }
+
if (fp == NULL)
return;
@@ -2211,7 +2199,6 @@ do_print_rec(int nargs, int redirtype)
rp->output.gawk_fflush(rp->output.fp, rp->output.opaque);
}
-#if MBS_SUPPORT
/* is_wupper --- function version of iswupper for passing function pointers */
@@ -2276,7 +2263,6 @@ wide_tolower(wchar_t *wstr, size_t wlen)
{
wide_change_case(wstr, wlen, is_wupper, to_wlower);
}
-#endif
/* do_tolower --- lower case a string */
@@ -2299,14 +2285,11 @@ do_tolower(int nargs)
cp < cp2; cp++)
if (isupper(*cp))
*cp = tolower(*cp);
- }
-#if MBS_SUPPORT
- else {
+ } else {
force_wstring(t2);
wide_tolower(t2->wstptr, t2->wstlen);
wstr2str(t2);
}
-#endif
DEREF(t1);
return t2;
@@ -2333,14 +2316,11 @@ do_toupper(int nargs)
cp < cp2; cp++)
if (islower(*cp))
*cp = toupper(*cp);
- }
-#if MBS_SUPPORT
- else {
+ } else {
force_wstring(t2);
wide_toupper(t2->wstptr, t2->wstlen);
wstr2str(t2);
}
-#endif
DEREF(t1);
return t2;
@@ -2551,13 +2531,12 @@ do_match(int nargs)
size_t *wc_indices = NULL;
rlength = REEND(rp, t1->stptr) - RESTART(rp, t1->stptr); /* byte length */
-#if MBS_SUPPORT
if (rlength > 0 && gawk_mb_cur_max > 1) {
t1 = str2wstr(t1, & wc_indices);
rlength = wc_indices[rstart + rlength - 1] - wc_indices[rstart] + 1;
rstart = wc_indices[rstart];
}
-#endif
+
rstart++; /* now it's 1-based indexing */
/* Build the array only if the caller wants the optional subpatterns */
@@ -2579,12 +2558,10 @@ do_match(int nargs)
start = t1->stptr + s;
subpat_start = s;
subpat_len = len = SUBPATEND(rp, t1->stptr, ii) - s;
-#if MBS_SUPPORT
if (len > 0 && gawk_mb_cur_max > 1) {
subpat_start = wc_indices[s];
subpat_len = wc_indices[s + len - 1] - subpat_start + 1;
}
-#endif
it = make_string(start, len);
it->flags |= MAYBE_NUM; /* user input */
@@ -2803,6 +2780,8 @@ do_sub(int nargs, unsigned int flags)
if ((t1->flags & NUMCUR) != 0)
goto set_how_many;
+ warning(_("gensub: third argument `%.*s' treated as 1"),
+ (int) t1->stlen, t1->stptr);
how_many = 1;
}
} else {
@@ -2815,8 +2794,8 @@ set_how_many:
how_many = d;
else
how_many = LONG_MAX;
- if (d == 0)
- warning(_("gensub: third argument of 0 treated as 1"));
+ if (d <= 0)
+ warning(_("gensub: third argument %g treated as 1"), d);
}
DEREF(t1);
@@ -3093,6 +3072,146 @@ done:
return make_number((AWKNUM) matches);
}
+/* call_sub --- call do_sub indirectly */
+
+NODE *
+call_sub(const char *name, int nargs)
+{
+ unsigned int flags = 0;
+ NODE *regex, *replace, *glob_flag;
+ NODE **lhs, *rhs;
+ NODE *zero = make_number(0.0);
+ NODE *result;
+
+ if (name[0] == 'g') {
+ if (name[1] == 'e')
+ flags = GENSUB;
+ else
+ flags = GSUB;
+ }
+
+ if (flags == 0 || flags == GSUB) {
+ /* sub or gsub */
+ if (nargs != 2)
+ fatal(_("%s: can be called indirectly only with two arguments"), name);
+
+ replace = POP_STRING();
+ regex = POP(); /* the regex */
+ /*
+ * push regex
+ * push replace
+ * push $0
+ */
+ regex = make_regnode(Node_regex, regex);
+ PUSH(regex);
+ PUSH(replace);
+ lhs = r_get_field(zero, (Func_ptr *) 0, true);
+ nargs++;
+ PUSH_ADDRESS(lhs);
+ } else {
+ /* gensub */
+ if (nargs == 4)
+ rhs = POP();
+ else
+ rhs = NULL;
+ glob_flag = POP_STRING();
+ replace = POP_STRING();
+ regex = POP(); /* the regex */
+ /*
+ * push regex
+ * push replace
+ * push glob_flag
+ * if (nargs = 3) {
+ * push $0
+ * nargs++
+ * }
+ */
+ regex = make_regnode(Node_regex, regex);
+ PUSH(regex);
+ PUSH(replace);
+ PUSH(glob_flag);
+ if (rhs == NULL) {
+ lhs = r_get_field(zero, (Func_ptr *) 0, true);
+ rhs = *lhs;
+ UPREF(rhs);
+ PUSH(rhs);
+ nargs++;
+ }
+ PUSH(rhs);
+ }
+
+
+ unref(zero);
+ result = do_sub(nargs, flags);
+ if (flags != GENSUB)
+ reset_record();
+ return result;
+}
+
+/* call_match --- call do_match indirectly */
+
+NODE *
+call_match(int nargs)
+{
+ NODE *regex, *text, *array;
+ NODE *result;
+
+ regex = text = array = NULL;
+ if (nargs == 3)
+ array = POP();
+ regex = POP();
+
+ /* Don't need to pop the string just to push it back ... */
+
+ regex = make_regnode(Node_regex, regex);
+ PUSH(regex);
+
+ if (array)
+ PUSH(array);
+
+ result = do_match(nargs);
+ return result;
+}
+
+/* call_split_func --- call do_split or do_pat_split indirectly */
+
+NODE *
+call_split_func(const char *name, int nargs)
+{
+ NODE *regex, *seps;
+ NODE *result;
+
+ regex = seps = NULL;
+ if (nargs < 2)
+ fatal(_("indirect call to %s requires at least two arguments"),
+ name);
+
+ if (nargs == 4)
+ seps = POP();
+
+ if (nargs >= 3) {
+ regex = POP_STRING();
+ regex = make_regnode(Node_regex, regex);
+ } else {
+ if (name[0] == 's') {
+ regex = make_regnode(Node_regex, FS_node->var_value);
+ regex->re_flags |= FS_DFLT;
+ } else
+ regex = make_regnode(Node_regex, FPAT_node->var_value);
+ nargs++;
+ }
+
+ /* Don't need to pop the string or the data array */
+
+ PUSH(regex);
+
+ if (seps)
+ PUSH(seps);
+
+ result = (name[0] == 's') ? do_split(nargs) : do_patsplit(nargs);
+
+ return result;
+}
/* make_integer - Convert an integer to a number node. */
@@ -3633,7 +3752,7 @@ do_bindtextdomain(int nargs)
return make_string(the_result, strlen(the_result));
}
-/* do_div --- do integer division, return quotient and remainder in dest array */
+/* do_intdiv --- do integer division, return quotient and remainder in dest array */
/*
* We define the semantics as:
@@ -3644,7 +3763,7 @@ do_bindtextdomain(int nargs)
*/
NODE *
-do_div(int nargs)
+do_intdiv(int nargs)
{
NODE *numerator, *denominator, *result;
double num, denom, quotient, remainder;
@@ -3652,7 +3771,7 @@ do_div(int nargs)
result = POP_PARAM();
if (result->type != Node_var_array)
- fatal(_("div: third argument is not an array"));
+ fatal(_("intdiv: third argument is not an array"));
assoc_clear(result);
denominator = POP_SCALAR();
@@ -3660,9 +3779,9 @@ do_div(int nargs)
if (do_lint) {
if ((numerator->flags & (NUMCUR|NUMBER)) == 0)
- lintwarn(_("div: received non-numeric first argument"));
+ lintwarn(_("intdiv: received non-numeric first argument"));
if ((denominator->flags & (NUMCUR|NUMBER)) == 0)
- lintwarn(_("div: received non-numeric second argument"));
+ lintwarn(_("intdiv: received non-numeric second argument"));
}
(void) force_number(numerator);
@@ -3671,7 +3790,7 @@ do_div(int nargs)
denom = double_to_int(get_number_d(denominator));
if (denom == 0.0)
- fatal(_("div: division by zero attempted"));
+ fatal(_("intdiv: division by zero attempted"));
quotient = double_to_int(num / denom);
/*
@@ -3705,7 +3824,6 @@ do_div(int nargs)
static size_t
mbc_byte_count(const char *ptr, size_t numchars)
{
-#if MBS_SUPPORT
mbstate_t cur_state;
size_t sum = 0;
int mb_len;
@@ -3726,9 +3844,6 @@ mbc_byte_count(const char *ptr, size_t numchars)
}
return sum;
-#else
- return numchars;
-#endif
}
/* mbc_char_count --- return number of m.b. chars in string, up to numbytes bytes */
@@ -3736,7 +3851,6 @@ mbc_byte_count(const char *ptr, size_t numchars)
static size_t
mbc_char_count(const char *ptr, size_t numbytes)
{
-#if MBS_SUPPORT
mbstate_t cur_state;
size_t sum = 0;
int mb_len;
@@ -3759,7 +3873,4 @@ mbc_char_count(const char *ptr, size_t numbytes)
}
return sum;
-#else
- return numbytes;
-#endif
}
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-13 15:20:08 +0000