* modified by Herbert Xu to be used as built-in in ash.
* modified by Erik Andersen <andersen@codepoet.org> to be used
* in busybox.
- * modified by Bernhard Fischer to be useable (i.e. a bit less bloaty).
+ * modified by Bernhard Reutner-Fischer to be useable (i.e. a bit less bloaty).
*
* Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
*
* Original copyright notice states:
* "This program is in the Public Domain."
*/
-
#include "libbb.h"
#include <setjmp.h>
* This is true regardless of PREFER_APPLETS and STANDALONE_SHELL
* state. */
-
/* test(1) accepts the following grammar:
oexpr ::= aexpr | aexpr "-o" oexpr ;
aexpr ::= nexpr | nexpr "-a" aexpr ;
operand ::= <any legal UNIX file name>
*/
+/* TODO: handle [[ expr ]] bashism bash-compatibly.
+ * [[ ]] is meant to be a "better [ ]", with less weird syntax
+ * and without the risk of variables and quoted strings misinterpreted
+ * as operators.
+ * This will require support from shells - we need to know quote status
+ * of each parameter (see below).
+ *
+ * Word splitting and pathname expansion should NOT be performed:
+ * # a="a b"; [[ $a = "a b" ]] && echo YES
+ * YES
+ * # [[ /bin/m* ]] && echo YES
+ * YES
+ *
+ * =~ should do regexp match
+ * = and == should do pattern match against right side:
+ * # [[ *a* == bab ]] && echo YES
+ * # [[ bab == *a* ]] && echo YES
+ * YES
+ * != does the negated == (i.e., also with pattern matching).
+ * Pattern matching is quotation-sensitive:
+ * # [[ bab == "b"a* ]] && echo YES
+ * YES
+ * # [[ bab == b"a*" ]] && echo YES
+ *
+ * Conditional operators such as -f must be unquoted literals to be recognized:
+ * # [[ -e /bin ]] && echo YES
+ * YES
+ * # [[ '-e' /bin ]] && echo YES
+ * bash: conditional binary operator expected...
+ * # A='-e'; [[ $A /bin ]] && echo YES
+ * bash: conditional binary operator expected...
+ *
+ * || and && should work as -o and -a work in [ ]
+ * -a and -o aren't recognized (&& and || are to be used instead)
+ * ( and ) do not need to be quoted unlike in [ ]:
+ * # [[ ( abc ) && '' ]] && echo YES
+ * # [[ ( abc ) || '' ]] && echo YES
+ * YES
+ * # [[ ( abc ) -o '' ]] && echo YES
+ * bash: syntax error in conditional expression...
+ *
+ * Apart from the above, [[ expr ]] should work as [ expr ]
+ */
+
+#define TEST_DEBUG 0
+
enum token {
EOI,
- FILRD,
+
+ FILRD, /* file access */
FILWR,
FILEX,
+
FILEXIST,
- FILREG,
+
+ FILREG, /* file type */
FILDIR,
FILCDEV,
FILBDEV,
FILFIFO,
FILSOCK,
+
FILSYM,
FILGZ,
FILTT,
- FILSUID,
+
+ FILSUID, /* file bit */
FILSGID,
FILSTCK,
- FILNT,
+
+ FILNT, /* file ops */
FILOT,
FILEQ,
+
FILUID,
FILGID,
- STREZ,
+
+ STREZ, /* str ops */
STRNZ,
STREQ,
STRNE,
STRLT,
STRGT,
- INTEQ,
+
+ INTEQ, /* int ops */
INTNE,
INTGE,
INTGT,
INTLE,
INTLT,
+
UNOT,
BAND,
BOR,
#define is_file_access(a) (((unsigned char)((a) - FILRD)) <= 2)
#define is_file_type(a) (((unsigned char)((a) - FILREG)) <= 5)
#define is_file_bit(a) (((unsigned char)((a) - FILSUID)) <= 2)
+
+#if TEST_DEBUG
+int depth;
+#define nest_msg(...) do { \
+ depth++; \
+ fprintf(stderr, "%*s", depth*2, ""); \
+ fprintf(stderr, __VA_ARGS__); \
+} while (0)
+#define unnest_msg(...) do { \
+ fprintf(stderr, "%*s", depth*2, ""); \
+ fprintf(stderr, __VA_ARGS__); \
+ depth--; \
+} while (0)
+#define dbg_msg(...) do { \
+ fprintf(stderr, "%*s", depth*2, ""); \
+ fprintf(stderr, __VA_ARGS__); \
+} while (0)
+#define unnest_msg_and_return(expr, ...) do { \
+ number_t __res = (expr); \
+ fprintf(stderr, "%*s", depth*2, ""); \
+ fprintf(stderr, __VA_ARGS__, res); \
+ depth--; \
+ return __res; \
+} while (0)
+static const char *const TOKSTR[] = {
+ "EOI",
+ "FILRD",
+ "FILWR",
+ "FILEX",
+ "FILEXIST",
+ "FILREG",
+ "FILDIR",
+ "FILCDEV",
+ "FILBDEV",
+ "FILFIFO",
+ "FILSOCK",
+ "FILSYM",
+ "FILGZ",
+ "FILTT",
+ "FILSUID",
+ "FILSGID",
+ "FILSTCK",
+ "FILNT",
+ "FILOT",
+ "FILEQ",
+ "FILUID",
+ "FILGID",
+ "STREZ",
+ "STRNZ",
+ "STREQ",
+ "STRNE",
+ "STRLT",
+ "STRGT",
+ "INTEQ",
+ "INTNE",
+ "INTGE",
+ "INTGT",
+ "INTLE",
+ "INTLT",
+ "UNOT",
+ "BAND",
+ "BOR",
+ "LPAREN",
+ "RPAREN",
+ "OPERAND"
+};
+#else
+#define nest_msg(...) ((void)0)
+#define unnest_msg(...) ((void)0)
+#define dbg_msg(...) ((void)0)
+#define unnest_msg_and_return(expr, ...) return expr
+#endif
+
enum token_types {
UNOP,
BINOP,
PAREN
};
-static const struct t_op {
+struct operator_t {
char op_text[4];
unsigned char op_num, op_type;
-} ops[] = {
+};
+
+static const struct operator_t ops[] = {
{ "-r", FILRD , UNOP },
{ "-w", FILWR , UNOP },
{ "-x", FILEX , UNOP },
#if ENABLE_FEATURE_TEST_64
-typedef int64_t arith_t;
+typedef int64_t number_t;
#else
-typedef int arith_t;
+typedef int number_t;
#endif
/* We try to minimize both static and stack usage. */
-struct statics {
- char **t_wp;
- const struct t_op *t_wp_op;
+struct test_statics {
+ char **args;
+ /* set only by check_operator(), either to bogus struct
+ * or points to matching operator_t struct. Never NULL. */
+ const struct operator_t *last_operator;
gid_t *group_array;
int ngroups;
jmp_buf leaving;
};
-/* Make it reside in writable memory, yet make compiler understand
- * that it is not going to change. */
-static struct statics *const ptr_to_statics __attribute__ ((section (".data")));
+/* See test_ptr_hack.c */
+extern struct test_statics *const test_ptr_to_statics;
-#define S (*ptr_to_statics)
-#define t_wp (S.t_wp )
-#define t_wp_op (S.t_wp_op )
+#define S (*test_ptr_to_statics)
+#define args (S.args )
+#define last_operator (S.last_operator)
#define group_array (S.group_array )
#define ngroups (S.ngroups )
#define leaving (S.leaving )
#define INIT_S() do { \
- (*(struct statics**)&ptr_to_statics) = xzalloc(sizeof(S)); \
+ (*(struct test_statics**)&test_ptr_to_statics) = xzalloc(sizeof(S)); \
+ barrier(); \
} while (0)
#define DEINIT_S() do { \
- free(ptr_to_statics); \
+ free(test_ptr_to_statics); \
} while (0)
-static arith_t primary(enum token n);
+static number_t primary(enum token n);
-static void syntax(const char *op, const char *msg) ATTRIBUTE_NORETURN;
+static void syntax(const char *op, const char *msg) NORETURN;
static void syntax(const char *op, const char *msg)
{
if (op && *op) {
/* atoi with error detection */
//XXX: FIXME: duplicate of existing libbb function?
-static arith_t getn(const char *s)
+static number_t getn(const char *s)
{
char *p;
#if ENABLE_FEATURE_TEST_64
*/
-static enum token t_lex(char *s)
+static enum token check_operator(char *s)
{
- const struct t_op *op;
+ static const struct operator_t no_op = {
+ .op_num = -1,
+ .op_type = -1
+ };
+ const struct operator_t *op;
- t_wp_op = NULL;
+ last_operator = &no_op;
if (s == NULL) {
return EOI;
}
op = ops;
do {
if (strcmp(s, op->op_text) == 0) {
- t_wp_op = op;
+ last_operator = op;
return op->op_num;
}
op++;
static int binop(void)
{
const char *opnd1, *opnd2;
- struct t_op const *op;
- arith_t val1, val2;
+ const struct operator_t *op;
+ number_t val1, val2;
- opnd1 = *t_wp;
- (void) t_lex(*++t_wp);
- op = t_wp_op;
+ opnd1 = *args;
+ check_operator(*++args);
+ op = last_operator;
- opnd2 = *++t_wp;
+ opnd2 = *++args;
if (opnd2 == NULL)
syntax(op->op_text, "argument expected");
return val1 > val2;
if (op->op_num == INTLE)
return val1 <= val2;
- if (op->op_num == INTLT)
- return val1 < val2;
+ /*if (op->op_num == INTLT)*/
+ return val1 < val2;
}
if (is_str_op(op->op_num)) {
val1 = strcmp(opnd1, opnd2);
return val1 != 0;
if (op->op_num == STRLT)
return val1 < 0;
- if (op->op_num == STRGT)
- return val1 > 0;
+ /*if (op->op_num == STRGT)*/
+ return val1 > 0;
}
/* We are sure that these three are by now the only binops we didn't check
* yet, so we do not check if the class is correct:
return b1.st_mtime > b2.st_mtime;
if (op->op_num == FILOT)
return b1.st_mtime < b2.st_mtime;
- if (op->op_num == FILEQ)
- return b1.st_dev == b2.st_dev && b1.st_ino == b2.st_ino;
+ /*if (op->op_num == FILEQ)*/
+ return b1.st_dev == b2.st_dev && b1.st_ino == b2.st_ino;
}
- return 1; /* NOTREACHED */
+ /*return 1; - NOTREACHED */
}
static void initialize_group_array(void)
{
- ngroups = getgroups(0, NULL);
- if (ngroups > 0) {
+ int n;
+
+ /* getgroups may be expensive, try to use it only once */
+ ngroups = 32;
+ do {
/* FIXME: ash tries so hard to not die on OOM,
* and we spoil it with just one xrealloc here */
/* We realloc, because test_main can be entered repeatedly by shell.
* Testcase (ash): 'while true; do test -x some_file; done'
* and watch top. (some_file must have owner != you) */
- group_array = xrealloc(group_array, ngroups * sizeof(gid_t));
- getgroups(ngroups, group_array);
- }
+ n = ngroups;
+ group_array = xrealloc(group_array, n * sizeof(gid_t));
+ ngroups = getgroups(n, group_array);
+ } while (ngroups > n);
}
static int filstat(char *nm, enum token mode)
{
struct stat s;
- int i = i; /* gcc 3.x thinks it can be used uninitialized */
+ unsigned i = i; /* gcc 3.x thinks it can be used uninitialized */
if (mode == FILSYM) {
#ifdef S_IFLNK
}
-static arith_t nexpr(enum token n)
+static number_t nexpr(enum token n)
{
- if (n == UNOT)
- return !nexpr(t_lex(*++t_wp));
- return primary(n);
+ number_t res;
+
+ nest_msg(">nexpr(%s)\n", TOKSTR[n]);
+ if (n == UNOT) {
+ n = check_operator(*++args);
+ if (n == EOI) {
+ /* special case: [ ! ], [ a -a ! ] are valid */
+ /* IOW, "! ARG" may miss ARG */
+ unnest_msg("<nexpr:1 (!EOI)\n");
+ return 1;
+ }
+ res = !nexpr(n);
+ unnest_msg("<nexpr:%lld\n", res);
+ return res;
+ }
+ res = primary(n);
+ unnest_msg("<nexpr:%lld\n", res);
+ return res;
}
-static arith_t aexpr(enum token n)
+static number_t aexpr(enum token n)
{
- arith_t res;
+ number_t res;
+ nest_msg(">aexpr(%s)\n", TOKSTR[n]);
res = nexpr(n);
- if (t_lex(*++t_wp) == BAND)
- return aexpr(t_lex(*++t_wp)) && res;
- t_wp--;
+ dbg_msg("aexpr: nexpr:%lld, next args:%s\n", res, args[1]);
+ if (check_operator(*++args) == BAND) {
+ dbg_msg("aexpr: arg is AND, next args:%s\n", args[1]);
+ res = aexpr(check_operator(*++args)) && res;
+ unnest_msg("<aexpr:%lld\n", res);
+ return res;
+ }
+ args--;
+ unnest_msg("<aexpr:%lld, args:%s\n", res, args[0]);
return res;
}
-static arith_t oexpr(enum token n)
+static number_t oexpr(enum token n)
{
- arith_t res;
+ number_t res;
+ nest_msg(">oexpr(%s)\n", TOKSTR[n]);
res = aexpr(n);
- if (t_lex(*++t_wp) == BOR) {
- return oexpr(t_lex(*++t_wp)) || res;
+ dbg_msg("oexpr: aexpr:%lld, next args:%s\n", res, args[1]);
+ if (check_operator(*++args) == BOR) {
+ dbg_msg("oexpr: next arg is OR, next args:%s\n", args[1]);
+ res = oexpr(check_operator(*++args)) || res;
+ unnest_msg("<oexpr:%lld\n", res);
+ return res;
}
- t_wp--;
+ args--;
+ unnest_msg("<oexpr:%lld, args:%s\n", res, args[0]);
return res;
}
-
-static arith_t primary(enum token n)
+static number_t primary(enum token n)
{
- arith_t res;
+#if TEST_DEBUG
+ number_t res = res; /* for compiler */
+#else
+ number_t res;
+#endif
+ const struct operator_t *args0_op;
+ nest_msg(">primary(%s)\n", TOKSTR[n]);
if (n == EOI) {
syntax(NULL, "argument expected");
}
if (n == LPAREN) {
- res = oexpr(t_lex(*++t_wp));
- if (t_lex(*++t_wp) != RPAREN)
+ res = oexpr(check_operator(*++args));
+ if (check_operator(*++args) != RPAREN)
syntax(NULL, "closing paren expected");
+ unnest_msg("<primary:%lld\n", res);
return res;
}
- if (t_wp_op && t_wp_op->op_type == UNOP) {
+
+ /* coreutils 6.9 checks "is args[1] binop and args[2] exist?" first,
+ * do the same */
+ args0_op = last_operator;
+ /* last_operator = operator at args[1] */
+ if (check_operator(args[1]) != EOI) { /* if args[1] != NULL */
+ if (args[2]) {
+ // coreutils also does this:
+ // if (args[3] && args[0]="-l" && args[2] is BINOP)
+ // return binop(1 /* prepended by -l */);
+ if (last_operator->op_type == BINOP)
+ unnest_msg_and_return(binop(), "<primary: binop:%lld\n");
+ }
+ }
+ /* check "is args[0] unop?" second */
+ if (args0_op->op_type == UNOP) {
/* unary expression */
- if (*++t_wp == NULL)
- syntax(t_wp_op->op_text, "argument expected");
+ if (args[1] == NULL)
+// syntax(args0_op->op_text, "argument expected");
+ goto check_emptiness;
+ args++;
if (n == STREZ)
- return t_wp[0][0] == '\0';
+ unnest_msg_and_return(args[0][0] == '\0', "<primary:%lld\n");
if (n == STRNZ)
- return t_wp[0][0] != '\0';
+ unnest_msg_and_return(args[0][0] != '\0', "<primary:%lld\n");
if (n == FILTT)
- return isatty(getn(*t_wp));
- return filstat(*t_wp, n);
+ unnest_msg_and_return(isatty(getn(*args)), "<primary: isatty(%s)%lld\n", *args);
+ unnest_msg_and_return(filstat(*args, n), "<primary: filstat(%s):%lld\n", *args);
}
- t_lex(t_wp[1]);
- if (t_wp_op && t_wp_op->op_type == BINOP) {
- return binop();
+ /*check_operator(args[1]); - already done */
+ if (last_operator->op_type == BINOP) {
+ /* args[2] is known to be NULL, isn't it bound to fail? */
+ unnest_msg_and_return(binop(), "<primary:%lld\n");
}
-
- return t_wp[0][0] != '\0';
+ check_emptiness:
+ unnest_msg_and_return(args[0][0] != '\0', "<primary:%lld\n");
}
{
int res;
const char *arg0;
- bool negate = 0;
+// bool negate = 0;
arg0 = bb_basename(argv[0]);
if (arg0[0] == '[') {
* isn't likely in the case of a shell. paranoia
* prevails...
*/
- ngroups = 0;
+ /*ngroups = 0; - done by INIT_S() */
//argc--;
argv++;
res = 1;
goto ret;
}
+#if 0
+// Now it's fixed in the parser and should not be needed
if (LONE_CHAR(argv[0], '!') && argv[1]) {
negate = 1;
//argc--;
goto ret;
}
if (argv[2] && !argv[3]) {
- t_lex(argv[1]);
- if (t_wp_op && t_wp_op->op_type == BINOP) {
+ check_operator(argv[1]);
+ if (last_operator->op_type == BINOP) {
/* "test [!] arg1 <binary_op> arg2" */
- t_wp = &argv[0];
+ args = argv;
res = (binop() == 0);
goto ret;
}
//argc++;
argv--;
}
- t_wp = &argv[0];
- res = !oexpr(t_lex(*t_wp));
+#endif
+ args = argv;
+ res = !oexpr(check_operator(*args));
- if (*t_wp != NULL && *++t_wp != NULL) {
- bb_error_msg("%s: unknown operand", *t_wp);
+ if (*args != NULL && *++args != NULL) {
+ /* TODO: example when this happens? */
+ bb_error_msg("%s: unknown operand", *args);
res = 2;
}
ret:
DEINIT_S();
- return negate ? !res : res;
+// return negate ? !res : res;
+ return res;
}
projects / oweals / busybox.git / blobdiff