Files
mercury/runtime/mercury_string.h
Julien Fischer bec35c03ff Fix compilation on Windows.
runtime/mercury_string.h:
     Fix a cut-and-paste error.
2023-10-18 14:58:44 +11:00

521 lines
25 KiB
C

// vim: ts=4 sw=4 expandtab ft=c
// Copyright (C) 1995-2005, 2007, 2011-2012 The University of Melbourne.
// Copyright (C) 2015-2016, 2018-2019, 2021, 2023 The Mercury team.
// This file is distributed under the terms specified in COPYING.LIB.
// mercury_string.h - string handling
#ifndef MERCURY_STRING_H
#define MERCURY_STRING_H
#include "mercury_heap.h" // for MR_offset_incr_hp_atomic
#include <string.h> // for strcmp() etc.
#include <stdarg.h>
#include <stdio.h>
#if defined(MR_WIN32)
#include <wchar.h>
#endif
// On Windows, snprintf/vsnprintf may be synonyms for _snprintf/_vsnprintf and
// thus not conform to the C99 specification. Since it is next to impossible to
// tell at compile time which implementation we are getting, just define a
// wrapper over _vsnprintf if it exists.
// Beginning with the UCRT in Visual Studio 2015 and Windows 10, snprintf and
// vsnprintf are C99 standard compliant so we may be able to drop this code
// eventually.
#if defined(MR_HAVE__VSNPRINTF)
extern int
MR_vsnprintf(char *str, size_t size, const char *format, va_list ap);
#elif defined(MR_HAVE_VSNPRINTF)
#define MR_vsnprintf vsnprintf
#else
#error "Missing both vsnprintf and _vsnprintf"
#endif
#if defined(MR_HAVE__SNPRINTF)
extern int MR_snprintf(char *str, size_t size, const char *format, ...);
#elif defined(MR_HAVE_SNPRINTF)
#define MR_snprintf snprintf
#else
#error "Missing both snprintf and _snprintf"
#endif
// Mercury characters (Unicode code points) are given type `MR_Char',
// which is a typedef for `MR_int_least32_t'.
// Mercury strings are stored as pointers to '\0'-terminated arrays of `char'.
// Strings are UTF-8 encoded.
// Mercury strings must not contain null characters. Unexpected null characters
// are a source of security vulnerabilities.
//
// The actual typedefs are in mercury_types.h to avoid problems with
// circular #includes.
//
// typedef MR_int_least32_t MR_Char;
// typedef MR_uint_least32_t MR_UnsignedChar;
//
// typedef char *MR_String;
// typedef const char *MR_ConstString;
// MR_string_const("...", len):
//
// Given a C string literal and its length, returns a Mercury string.
#define MR_string_const(string, len) ((MR_String) string)
#define MR_make_string_const(string) \
MR_string_const((string), sizeof(string) - 1)
// MR_bool MR_string_equal(MR_ConstString s1, MR_ConstString s2):
// Return true iff the two Mercury strings s1 and s2 are equal.
#define MR_string_equal(s1,s2) (strcmp((char*)(s1),(char*)(s2))==0)
// void MR_make_aligned_string(MR_String ptr, const char *string):
//
// Given a C string `string', set `ptr' to be a Mercury string with the
// same contents. (`ptr' must be an lvalue.) If the resulting Mercury string
// is to be used by Mercury code, then the string pointed to by `string'
// should have been either statically allocated or allocated on the
// Mercury heap.
//
// BEWARE: this may modify `MR_hp', so it must only be called from places
// where `MR_hp' is valid. If calling it from inside a C function, rather than
// inside Mercury code, you may need to call MR_{save/restore}_transient_hp().
//
// Algorithm: if the string is aligned, just set ptr equal to it.
// Otherwise, allocate space on the heap and copy the C string to
// the Mercury string.
#define MR_make_aligned_string(ptr, string) \
do { \
if (MR_tag((MR_Word) (string)) != 0) { \
MR_make_aligned_string_copy((ptr), (string)); \
} else { \
/* The cast is there to cast away const, if needed. */ \
(ptr) = (MR_String) (string); \
} \
} while (0)
// void MR_make_aligned_string_copy(MR_String ptr, const char * string):
//
// Same as MR_make_aligned_string(ptr, string), except that the string
// is guaranteed to be copied. This is useful for copying C strings
// onto the Mercury heap.
//
// BEWARE: this may modify `MR_hp', so it must only be called from places
// where `MR_hp' is valid. If calling it from inside a C function, rather than
// inside Mercury code, you may need to call MR_{save/restore}_transient_hp().
#define MR_make_aligned_string_copy(ptr, string) \
MR_make_aligned_string_copy_msg((ptr), (string), NULL)
#define MR_make_aligned_string_copy_msg(ptr, string, alloc_id) \
do { \
MR_Word make_aligned_string_tmp; \
char *make_aligned_string_ptr; \
\
MR_offset_incr_hp_atomic_msg(make_aligned_string_tmp, 0, \
(strlen(string) + sizeof(MR_Word)) / sizeof(MR_Word), \
(alloc_id), "string.string/0"); \
make_aligned_string_ptr = (char *) make_aligned_string_tmp; \
strcpy(make_aligned_string_ptr, (string)); \
(ptr) = make_aligned_string_ptr; \
} while (0)
// void MR_make_aligned_string_copy_saved_hp(MR_String ptr,
// const char * string):
//
// Same as MR_make_aligned_string_copy(ptr, string), except that it uses
// MR_offset_incr_saved_hp_atomic instead of MR_offset_incr_hp_atomic.
#define MR_make_aligned_string_copy_saved_hp(ptr, string, alloc_id) \
do { \
MR_Word make_aligned_string_tmp; \
char *make_aligned_string_ptr; \
\
MR_offset_incr_saved_hp_atomic(make_aligned_string_tmp, 0, \
(strlen(string) + sizeof(MR_Word)) / sizeof(MR_Word), \
(alloc_id), "string.string/0"); \
make_aligned_string_ptr = (char *) make_aligned_string_tmp; \
strcpy(make_aligned_string_ptr, (string)); \
(ptr) = make_aligned_string_ptr; \
} while (0)
// void MR_make_aligned_string_copy_saved_hp_quote(MR_String ptr,
// const char * string):
//
// Same as MR_make_aligned_string_copy_saved_hp(ptr, string), except that
// it puts double quote marks at the start and end of the string.
#define MR_make_aligned_string_copy_saved_hp_quote(ptr, string, alloc_id) \
do { \
MR_Word make_aligned_string_tmp; \
char *make_aligned_string_ptr; \
\
MR_offset_incr_saved_hp_atomic(make_aligned_string_tmp, 0, \
(strlen(string) + 2 + sizeof(MR_Word)) / sizeof(MR_Word), \
(alloc_id), "string.string/0"); \
make_aligned_string_ptr = (char *) make_aligned_string_tmp; \
sprintf(make_aligned_string_ptr, "%c%s%c", '"', string, '"'); \
(ptr) = make_aligned_string_ptr; \
} while (0)
// void MR_allocate_aligned_string_msg(MR_String ptr, size_t len,
// MR_Code *proclabel):
//
// Allocate enough word aligned memory to hold len characters. Also record
// the location, proclabel, of the allocation if profiling is enabled.
//
// BEWARE: this may modify `MR_hp', so it must only be called from places
// where `MR_hp' is valid. If calling it from inside a C function, rather than
// inside Mercury code, you may need to call MR_{save/restore}_transient_hp().
#define MR_allocate_aligned_string_msg(ptr, len, alloc_id) \
do { \
MR_Word make_aligned_string_tmp; \
char *make_aligned_string_ptr; \
\
MR_offset_incr_hp_atomic_msg(make_aligned_string_tmp, 0, \
((len) + sizeof(MR_Word)) / sizeof(MR_Word), \
(alloc_id), "string.string/0"); \
make_aligned_string_ptr = (char *) make_aligned_string_tmp; \
(ptr) = make_aligned_string_ptr; \
} while (0)
#define MR_allocate_aligned_string_saved_hp(ptr, len, alloc_id) \
do { \
MR_Word make_aligned_string_tmp; \
char *make_aligned_string_ptr; \
\
MR_offset_incr_saved_hp_atomic(make_aligned_string_tmp, 0, \
((len) + sizeof(MR_Word)) / sizeof(MR_Word), \
(alloc_id), "string.string/0"); \
make_aligned_string_ptr = (char *) make_aligned_string_tmp; \
(ptr) = make_aligned_string_ptr; \
} while (0)
// MR_do_hash_string{,2,3,4,5,6}(int & hash, MR_Word string):
//
// Given a Mercury string `string', set `hash' to the hash value
// for that string. (`hash' must be an lvalue.)
//
// This is an implementation detail used to implement
// MR_hash_string{,2,3,4,5,6}(). It should not be used directly.
// Use MR_hash_string{,2,3,4,5,6}() instead.
//
// Note that these functions are also defined in library/string.m.
// The definition here and in string.m must be kept equivalent.
#define MR_do_hash_string(hash, s) \
{ \
int len; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
len = 0; \
hash = 0; \
while (((const unsigned char *)(s))[len]) { \
hash ^= (hash << 5); \
hash ^= ((const unsigned char *)(s))[len]; \
len++; \
} \
hash ^= len; \
}
#define MR_do_hash_string2(hash, s) \
{ \
int len; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
len = 0; \
hash = 0; \
while (((const unsigned char *)(s))[len]) { \
hash = hash * 37; \
hash += ((const unsigned char *)(s))[len]; \
len++; \
} \
hash ^= len; \
}
#define MR_do_hash_string3(hash, s) \
{ \
int len; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
len = 0; \
hash = 0; \
while (((const unsigned char *)(s))[len]) { \
hash = hash * 49; \
hash += ((const unsigned char *)(s))[len]; \
len++; \
} \
hash ^= len; \
}
#define MR_keep_30_bits(x) \
((x) & ((1 << 30) - 1))
#define MR_do_hash_string4(hash, s) \
{ \
int i; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
hash = 0; \
for (i = 0; ((const unsigned char *)(s))[i] != 0; i++) { \
hash = MR_keep_30_bits(hash ^ (hash << 5)); \
hash = hash ^ ((const unsigned char *)(s))[i]; \
} \
hash ^= i; \
}
#define MR_do_hash_string5(hash, s) \
{ \
int i; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
hash = 0; \
for (i = 0; ((const unsigned char *)(s))[i] != 0; i++) { \
hash = MR_keep_30_bits(hash * 37); \
hash = MR_keep_30_bits(hash + ((const unsigned char *)(s))[i]); \
} \
hash ^= i; \
}
#define MR_do_hash_string6(hash, s) \
{ \
int i; \
MR_CHECK_EXPR_TYPE(hash, int); \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
hash = 0; \
for (i = 0; ((const unsigned char *)(s))[i] != 0; i++) { \
hash = MR_keep_30_bits(hash * 49); \
hash = MR_keep_30_bits(hash + ((const unsigned char *)(s))[i]); \
} \
hash ^= i; \
}
// MR_hash_string{,2,3,4,5,6}(s):
//
// Given a Mercury string `s', return a hash value for that string.
MR_Integer MR_hash_string(MR_ConstString);
MR_Integer MR_hash_string2(MR_ConstString);
MR_Integer MR_hash_string3(MR_ConstString);
MR_Integer MR_hash_string4(MR_ConstString);
MR_Integer MR_hash_string5(MR_ConstString);
MR_Integer MR_hash_string6(MR_ConstString);
#if defined(MR_GNUC)
#define MR_hash_string(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string(hash_string_result, s); \
hash_string_result; \
})
#define MR_hash_string2(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string2(hash_string_result, s); \
hash_string_result; \
})
#define MR_hash_string3(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string3(hash_string_result, s); \
hash_string_result; \
})
#define MR_hash_string4(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string4(hash_string_result, s); \
hash_string_result; \
})
#define MR_hash_string5(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string5(hash_string_result, s); \
hash_string_result; \
})
#define MR_hash_string6(s) \
({ \
MR_Integer hash_string_result; \
MR_CHECK_EXPR_TYPE(s, MR_ConstString); \
MR_do_hash_string6(hash_string_result, s); \
hash_string_result; \
})
#endif
// If we are not using gcc, the actual definitions of these functions
// are runtime/mercury_string.c; they use the macros below.
#define MR_HASH_STRING_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string(hash_string_result, s); \
return hash_string_result;
#define MR_HASH_STRING2_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string2(hash_string_result, s); \
return hash_string_result;
#define MR_HASH_STRING3_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string3(hash_string_result, s); \
return hash_string_result;
#define MR_HASH_STRING4_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string4(hash_string_result, s); \
return hash_string_result;
#define MR_HASH_STRING5_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string5(hash_string_result, s); \
return hash_string_result;
#define MR_HASH_STRING6_FUNC_BODY \
MR_Integer hash_string_result; \
MR_do_hash_string6(hash_string_result, s); \
return hash_string_result;
// A version of strcmp to which we can pass Mercury words
// without having to cast the arguments first.
#define MR_strcmp(s, t) strcmp((const char *)(s), (const char *)(t))
// Assuming that the first n code units of s and t are equal,
// are the rest of s and t equal?
#define MR_offset_streq(n, s, t) \
(strcmp((const char *)((s)+(n)), (const char *)((t)+(n))) == 0)
// Return the kth code unit in a string.
#define MR_nth_code_unit(s, k) \
((unsigned) (((const unsigned char *) (s))[(k)]))
// Return an MR_String which has been created using the format string, fmt,
// passed to sprintf. If memory profiling is turned on, record the allocation
// as coming from proclabel. The MR_String returned has been allocated
// on the mercury heap using MR_allocate_aligned_string_msg.
//
// BEWARE: this may modify the saved copy of `MR_hp', so it must only be called
// from places where the saved copy of `MR_hp' is valid. You will generally
// need to call MR_{save/restore}_transient_hp() before/after calling
// this function.
MR_String MR_make_string(MR_AllocSiteInfoPtr alloc_id, const char *fmt, ...);
// Given a Mercury string `string', make a copy that inserts any required
// character escapes and places double quote marks at the start and end of
// the copy. On success, returns MR_TRUE and sets `ptr' to point to the copy
// of the string. Returns MR_FALSE if `string' is not a valid UTF-8 encoded
// string.
extern MR_bool MR_escape_string_quote(MR_String *ptr, const char * string);
// True if c is an ASCII code point, i.e. U+0000..U+007f.
#define MR_is_ascii(c) ((unsigned) (c) <= 0x7f)
// True if c is a Unicode surrogate code point, i.e. U+D800..U+DFFF.
#define MR_is_surrogate(c) (((unsigned) (c) & 0x1FF800) == 0xD800)
// True if c is a Unicode control code point, i.e. U+0000..U+001f,
// U+007f..U+009f.
#define MR_is_control(c) \
((0x00 <= (unsigned) (c) && (unsigned) (c) <= 0x1f) || \
(0x7f <= (unsigned) (c) && (unsigned) (c) <= 0x9f))
// UTF-8 manipulation
#define MR_utf8_is_single_byte(c) (((unsigned) (c) & 0x80) == 0)
#define MR_utf8_is_lead_byte(c) (((unsigned) (c) - 0xC0) < 0x3E)
#define MR_utf8_is_trail_byte(c) (((unsigned) (c) & 0xC0) == 0x80)
// XXX ILSEQ The following functions should be rethought to make dealing
// with ill-formed code unit sequences easier.
// Advance `*pos' to the beginning of the next code point in `s'.
// If `*pos' is already at the end of the string, return MR_FALSE
// without modifying `*pos'.
// This function simply searches for a single or lead byte without decoding,
// and so may skip over bytes in ill-formed sequences.
extern MR_bool MR_utf8_next(const MR_String s_, MR_Integer *pos);
// Rewind `*pos' to the beginning of the previous code point in `s'.
// If `*pos' is already at the beginning of the string, return MR_FALSE
// without modifying `*pos'.
// This function simply searches for a single or lead byte without decoding,
// and so may skip over bytes in ill-formed sequences.
extern MR_bool MR_utf8_prev(const MR_String s_, MR_Integer *pos);
// Decode and return the code point beginning at `pos' in `s'.
// Return 0 if at the end of the string (i.e. the NUL terminator).
// Return -2 if the code unit sequence beginning at that offset is ill-formed.
//
// The _mb version requires s[pos] to be the lead byte of a multibyte code
// point.
extern MR_int_least32_t MR_utf8_get(const MR_String s, MR_Integer pos);
extern MR_int_least32_t MR_utf8_get_mb(const MR_String s, MR_Integer pos,
int *width);
// Decode the code point beginning at `pos' in `s', and advance `*pos'.
// The _mb version requires s[pos] to be the lead byte of a multibyte code
// point.
extern MR_int_least32_t MR_utf8_get_next(const MR_String s, MR_Integer *pos);
extern MR_int_least32_t MR_utf8_get_next_mb(const MR_String s,
MR_Integer *pos);
// Rewind `*pos' to the beginning of the previous code point in `s'
// and return that code code.
// Return -1 if `*pos' is already at the beginning of the string.
extern MR_int_least32_t MR_utf8_prev_get(const MR_String s, MR_Integer *pos);
// Return the number of bytes required to encode the code point `c' in UTF-8.
// Returns 0 for surrogate code points or illegal values.
extern size_t MR_utf8_width(MR_Char c);
// Encode the code point `c' into the buffer `s'.
// Return the number of bytes used.
// Returns 0 if `c' is a surrogate or not a valid code point.
extern size_t MR_utf8_encode(char s[], MR_Char c);
// Return MR_TRUE iff `s' contains a valid UTF-8 encoded string.
extern MR_bool MR_utf8_verify(const MR_String s);
// If `s' contains an invalid UTF-8 encoded string, return the number of
// UTF-8 code units between its start and the first invalid UTF-8 character.
// This number may be zero or positive.
// If `s' contains a valid UTF-8 encoded string, return a negative value.
extern MR_Integer MR_utf8_find_ill_formed_char(const MR_String s);
// Accessing Unicode file names on Windows requires that we use the functions
// taking wide character strings. The following functions handle the
// conversions.
#if defined(MR_WIN32)
extern wchar_t *MR_utf8_to_wide(const char *s);
extern char *MR_wide_to_utf8(const wchar_t *ws,
MR_AllocSiteInfoPtr alloc_id);
#endif
#endif // not MERCURY_STRING_H