#ifdef HAVE_CONFIG_H
# include <config.h>
#endif
#include <stdbool.h>
#include <stdlib.h>
#include "format.h"
#include "c-ctype.h"
#include "xalloc.h"
#include "xerror.h"
#include "format-invalid.h"
#include "error.h"
#include "error-progname.h"
#include "gettext.h"
#define _(str) gettext (str)
enum format_arg_type
{
FAT_NONE = 0,
FAT_INTEGER = 1,
FAT_DOUBLE = 2,
FAT_CHAR = 3,
FAT_STRING = 4,
FAT_OBJC_OBJECT = 5,
FAT_POINTER = 6,
FAT_COUNT_POINTER = 7,
FAT_UNSIGNED = 1 << 3,
FAT_SIZE_SHORT = 1 << 4,
FAT_SIZE_CHAR = 2 << 4,
FAT_SIZE_LONG = 1 << 6,
FAT_SIZE_LONGLONG = 2 << 6,
FAT_SIZE_8_T = 1 << 8,
FAT_SIZE_16_T = 1 << 9,
FAT_SIZE_32_T = 1 << 10,
FAT_SIZE_64_T = 1 << 11,
FAT_SIZE_LEAST8_T = 1 << 12,
FAT_SIZE_LEAST16_T = 1 << 13,
FAT_SIZE_LEAST32_T = 1 << 14,
FAT_SIZE_LEAST64_T = 1 << 15,
FAT_SIZE_FAST8_T = 1 << 16,
FAT_SIZE_FAST16_T = 1 << 17,
FAT_SIZE_FAST32_T = 1 << 18,
FAT_SIZE_FAST64_T = 1 << 19,
FAT_SIZE_INTMAX_T = 1 << 20,
FAT_SIZE_INTPTR_T = 1 << 21,
FAT_SIZE_SIZE_T = 1 << 22,
FAT_SIZE_PTRDIFF_T = 1 << 23,
FAT_WIDE = FAT_SIZE_LONG,
FAT_SIZE_MASK = (FAT_SIZE_SHORT | FAT_SIZE_CHAR
| FAT_SIZE_LONG | FAT_SIZE_LONGLONG
| FAT_SIZE_8_T | FAT_SIZE_16_T
| FAT_SIZE_32_T | FAT_SIZE_64_T
| FAT_SIZE_LEAST8_T | FAT_SIZE_LEAST16_T
| FAT_SIZE_LEAST32_T | FAT_SIZE_LEAST64_T
| FAT_SIZE_FAST8_T | FAT_SIZE_FAST16_T
| FAT_SIZE_FAST32_T | FAT_SIZE_FAST64_T
| FAT_SIZE_INTMAX_T | FAT_SIZE_INTPTR_T
| FAT_SIZE_SIZE_T | FAT_SIZE_PTRDIFF_T)
};
struct numbered_arg
{
unsigned int number;
enum format_arg_type type;
};
struct unnumbered_arg
{
enum format_arg_type type;
};
struct spec
{
unsigned int directives;
unsigned int unnumbered_arg_count;
unsigned int allocated;
struct unnumbered_arg *unnumbered;
unsigned int c99_directives_count;
const char **c99_directives;
};
#undef isdigit
#define isdigit(c) ((unsigned int) ((c) - '0') < 10)
static int
numbered_arg_compare (const void *p1, const void *p2)
{
unsigned int n1 = ((const struct numbered_arg *) p1)->number;
unsigned int n2 = ((const struct numbered_arg *) p2)->number;
return (n1 > n2 ? 1 : n1 < n2 ? -1 : 0);
}
#define INVALID_C99_MACRO(directive_number) \
xasprintf (_("In the directive number %u, the token after '<' is not the name of a format specifier macro. The valid macro names are listed in ISO C 99 section 7.8.1."), directive_number)
static void *
format_parse (const char *format, bool objc_extensions, char **invalid_reason)
{
struct spec spec;
unsigned int numbered_arg_count;
struct numbered_arg *numbered;
struct spec *result;
spec.directives = 0;
numbered_arg_count = 0;
spec.unnumbered_arg_count = 0;
spec.allocated = 0;
numbered = NULL;
spec.unnumbered = NULL;
spec.c99_directives_count = 0;
spec.c99_directives = NULL;
for (; *format != '\0';)
if (*format++ == '%')
{
unsigned int number = 0;
enum format_arg_type type;
enum format_arg_type size;
spec.directives++;
if (isdigit (*format))
{
const char *f = format;
unsigned int m = 0;
do
{
m = 10 * m + (*f - '0');
f++;
}
while (isdigit (*f));
if (*f == '$')
{
if (m == 0)
{
*invalid_reason = INVALID_ARGNO_0 (spec.directives);
goto bad_format;
}
number = m;
format = ++f;
}
}
while (*format == ' ' || *format == '+' || *format == '-'
|| *format == '#' || *format == '0' || *format == '\'')
format++;
if (*format == '*')
{
unsigned int width_number = 0;
format++;
if (isdigit (*format))
{
const char *f = format;
unsigned int m = 0;
do
{
m = 10 * m + (*f - '0');
f++;
}
while (isdigit (*f));
if (*f == '$')
{
if (m == 0)
{
*invalid_reason =
INVALID_WIDTH_ARGNO_0 (spec.directives);
goto bad_format;
}
width_number = m;
format = ++f;
}
}
if (width_number)
{
if (spec.unnumbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == numbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
numbered = (struct numbered_arg *) xrealloc (numbered, spec.allocated * sizeof (struct numbered_arg));
}
numbered[numbered_arg_count].number = width_number;
numbered[numbered_arg_count].type = FAT_INTEGER;
numbered_arg_count++;
}
else
{
if (numbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == spec.unnumbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
spec.unnumbered = (struct unnumbered_arg *) xrealloc (spec.unnumbered, spec.allocated * sizeof (struct unnumbered_arg));
}
spec.unnumbered[spec.unnumbered_arg_count].type = FAT_INTEGER;
spec.unnumbered_arg_count++;
}
}
else if (isdigit (*format))
{
do format++; while (isdigit (*format));
}
if (*format == '.')
{
format++;
if (*format == '*')
{
unsigned int precision_number = 0;
format++;
if (isdigit (*format))
{
const char *f = format;
unsigned int m = 0;
do
{
m = 10 * m + (*f - '0');
f++;
}
while (isdigit (*f));
if (*f == '$')
{
if (m == 0)
{
*invalid_reason =
INVALID_PRECISION_ARGNO_0 (spec.directives);
goto bad_format;
}
precision_number = m;
format = ++f;
}
}
if (precision_number)
{
if (spec.unnumbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == numbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
numbered = (struct numbered_arg *) xrealloc (numbered, spec.allocated * sizeof (struct numbered_arg));
}
numbered[numbered_arg_count].number = precision_number;
numbered[numbered_arg_count].type = FAT_INTEGER;
numbered_arg_count++;
}
else
{
if (numbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == spec.unnumbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
spec.unnumbered = (struct unnumbered_arg *) xrealloc (spec.unnumbered, spec.allocated * sizeof (struct unnumbered_arg));
}
spec.unnumbered[spec.unnumbered_arg_count].type = FAT_INTEGER;
spec.unnumbered_arg_count++;
}
}
else if (isdigit (*format))
{
do format++; while (isdigit (*format));
}
}
if (*format == '<')
{
spec.c99_directives =
(const char **)
xrealloc (spec.c99_directives,
2 * (spec.c99_directives_count + 1)
* sizeof (const char *));
spec.c99_directives[2 * spec.c99_directives_count] = format;
format++;
if (*format != 'P')
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
format++;
if (*format != 'R')
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
format++;
if (*format != 'I')
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
format++;
switch (*format)
{
case 'i': case 'd':
type = FAT_INTEGER;
break;
case 'u': case 'o': case 'x': case 'X':
type = FAT_INTEGER | FAT_UNSIGNED;
break;
default:
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
format++;
if (format[0] == 'M' && format[1] == 'A' && format[2] == 'X')
{
type |= FAT_SIZE_INTMAX_T;
format += 3;
}
else if (format[0] == 'P' && format[1] == 'T' && format[2] == 'R')
{
type |= FAT_SIZE_INTPTR_T;
format += 3;
}
else
{
if (format[0] == 'L' && format[1] == 'E' && format[2] == 'A'
&& format[3] == 'S' && format[4] == 'T')
{
format += 5;
if (format[0] == '8')
{
type |= FAT_SIZE_LEAST8_T;
format++;
}
else if (format[0] == '1' && format[1] == '6')
{
type |= FAT_SIZE_LEAST16_T;
format += 2;
}
else if (format[0] == '3' && format[1] == '2')
{
type |= FAT_SIZE_LEAST32_T;
format += 2;
}
else if (format[0] == '6' && format[1] == '4')
{
type |= FAT_SIZE_LEAST64_T;
format += 2;
}
else
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
}
else if (format[0] == 'F' && format[1] == 'A'
&& format[2] == 'S' && format[3] == 'T')
{
format += 4;
if (format[0] == '8')
{
type |= FAT_SIZE_FAST8_T;
format++;
}
else if (format[0] == '1' && format[1] == '6')
{
type |= FAT_SIZE_FAST16_T;
format += 2;
}
else if (format[0] == '3' && format[1] == '2')
{
type |= FAT_SIZE_FAST32_T;
format += 2;
}
else if (format[0] == '6' && format[1] == '4')
{
type |= FAT_SIZE_FAST64_T;
format += 2;
}
else
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
}
else
{
if (format[0] == '8')
{
type |= FAT_SIZE_8_T;
format++;
}
else if (format[0] == '1' && format[1] == '6')
{
type |= FAT_SIZE_16_T;
format += 2;
}
else if (format[0] == '3' && format[1] == '2')
{
type |= FAT_SIZE_32_T;
format += 2;
}
else if (format[0] == '6' && format[1] == '4')
{
type |= FAT_SIZE_64_T;
format += 2;
}
else
{
*invalid_reason = INVALID_C99_MACRO (spec.directives);
goto bad_format;
}
}
}
if (*format != '>')
{
*invalid_reason =
xasprintf (_("In the directive number %u, the token after '<' is not followed by '>'."), spec.directives);
goto bad_format;
}
spec.c99_directives[2 * spec.c99_directives_count + 1] = format;
spec.c99_directives_count++;
}
else
{
size = 0;
for (;; format++)
{
if (*format == 'h')
{
if (size & (FAT_SIZE_SHORT | FAT_SIZE_CHAR))
size = FAT_SIZE_CHAR;
else
size = FAT_SIZE_SHORT;
}
else if (*format == 'l')
{
if (size & (FAT_SIZE_LONG | FAT_SIZE_LONGLONG))
size = FAT_SIZE_LONGLONG;
else
size = FAT_SIZE_LONG;
}
else if (*format == 'L')
size = FAT_SIZE_LONGLONG;
else if (*format == 'q')
size = FAT_SIZE_LONGLONG;
else if (*format == 'j')
size = FAT_SIZE_INTMAX_T;
else if (*format == 'z' || *format == 'Z')
size = FAT_SIZE_SIZE_T;
else if (*format == 't')
size = FAT_SIZE_PTRDIFF_T;
else
break;
}
switch (*format)
{
case '%':
case 'm':
type = FAT_NONE;
break;
case 'c':
type = FAT_CHAR;
type |= (size & (FAT_SIZE_LONG | FAT_SIZE_LONGLONG)
? FAT_WIDE : 0);
break;
case 'C':
type = FAT_CHAR | FAT_WIDE;
break;
case 's':
type = FAT_STRING;
type |= (size & (FAT_SIZE_LONG | FAT_SIZE_LONGLONG)
? FAT_WIDE : 0);
break;
case 'S':
type = FAT_STRING | FAT_WIDE;
break;
case 'i': case 'd':
type = FAT_INTEGER;
type |= (size & FAT_SIZE_MASK);
break;
case 'u': case 'o': case 'x': case 'X':
type = FAT_INTEGER | FAT_UNSIGNED;
type |= (size & FAT_SIZE_MASK);
break;
case 'e': case 'E': case 'f': case 'F': case 'g': case 'G':
case 'a': case 'A':
type = FAT_DOUBLE;
type |= (size & FAT_SIZE_LONGLONG);
break;
case '@':
if (objc_extensions)
{
type = FAT_OBJC_OBJECT;
break;
}
goto other;
case 'p':
type = FAT_POINTER;
break;
case 'n':
type = FAT_COUNT_POINTER;
type |= (size & FAT_SIZE_MASK);
break;
other:
default:
*invalid_reason =
(*format == '\0'
? INVALID_UNTERMINATED_DIRECTIVE ()
: INVALID_CONVERSION_SPECIFIER (spec.directives, *format));
goto bad_format;
}
}
if (type != FAT_NONE)
{
if (number)
{
if (spec.unnumbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == numbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
numbered = (struct numbered_arg *) xrealloc (numbered, spec.allocated * sizeof (struct numbered_arg));
}
numbered[numbered_arg_count].number = number;
numbered[numbered_arg_count].type = type;
numbered_arg_count++;
}
else
{
if (numbered_arg_count > 0)
{
*invalid_reason = INVALID_MIXES_NUMBERED_UNNUMBERED ();
goto bad_format;
}
if (spec.allocated == spec.unnumbered_arg_count)
{
spec.allocated = 2 * spec.allocated + 1;
spec.unnumbered = (struct unnumbered_arg *) xrealloc (spec.unnumbered, spec.allocated * sizeof (struct unnumbered_arg));
}
spec.unnumbered[spec.unnumbered_arg_count].type = type;
spec.unnumbered_arg_count++;
}
}
format++;
}
if (numbered_arg_count > 1)
{
unsigned int i, j;
bool err;
qsort (numbered, numbered_arg_count,
sizeof (struct numbered_arg), numbered_arg_compare);
err = false;
for (i = j = 0; i < numbered_arg_count; i++)
if (j > 0 && numbered[i].number == numbered[j-1].number)
{
enum format_arg_type type1 = numbered[i].type;
enum format_arg_type type2 = numbered[j-1].type;
enum format_arg_type type_both;
if (type1 == type2)
type_both = type1;
else
{
type_both = FAT_NONE;
if (!err)
*invalid_reason =
INVALID_INCOMPATIBLE_ARG_TYPES (numbered[i].number);
err = true;
}
numbered[j-1].type = type_both;
}
else
{
if (j < i)
{
numbered[j].number = numbered[i].number;
numbered[j].type = numbered[i].type;
}
j++;
}
numbered_arg_count = j;
if (err)
goto bad_format;
}
if (numbered_arg_count > 0)
{
unsigned int i;
for (i = 0; i < numbered_arg_count; i++)
if (numbered[i].number != i + 1)
{
*invalid_reason =
xasprintf (_("The string refers to argument number %u but ignores argument number %u."), numbered[i].number, i + 1);
goto bad_format;
}
spec.unnumbered_arg_count = numbered_arg_count;
spec.allocated = spec.unnumbered_arg_count;
spec.unnumbered = (struct unnumbered_arg *) xmalloc (spec.allocated * sizeof (struct unnumbered_arg));
for (i = 0; i < spec.unnumbered_arg_count; i++)
spec.unnumbered[i].type = numbered[i].type;
free (numbered);
numbered_arg_count = 0;
}
result = (struct spec *) xmalloc (sizeof (struct spec));
*result = spec;
return result;
bad_format:
if (numbered != NULL)
free (numbered);
if (spec.unnumbered != NULL)
free (spec.unnumbered);
if (spec.c99_directives != NULL)
free (spec.c99_directives);
return NULL;
}
static void *
format_c_parse (const char *format, char **invalid_reason)
{
return format_parse (format, false, invalid_reason);
}
static void *
format_objc_parse (const char *format, char **invalid_reason)
{
return format_parse (format, true, invalid_reason);
}
static void
format_free (void *descr)
{
struct spec *spec = (struct spec *) descr;
if (spec->unnumbered != NULL)
free (spec->unnumbered);
if (spec->c99_directives != NULL)
free (spec->c99_directives);
free (spec);
}
static int
format_get_number_of_directives (void *descr)
{
struct spec *spec = (struct spec *) descr;
return spec->directives;
}
static bool
format_check (const lex_pos_ty *pos, void *msgid_descr, void *msgstr_descr,
bool equality, bool noisy, const char *pretty_msgstr)
{
struct spec *spec1 = (struct spec *) msgid_descr;
struct spec *spec2 = (struct spec *) msgstr_descr;
bool err = false;
unsigned int i;
if (equality
? spec1->unnumbered_arg_count != spec2->unnumbered_arg_count
: spec1->unnumbered_arg_count < spec2->unnumbered_arg_count)
{
if (noisy)
{
error_with_progname = false;
error_at_line (0, 0, pos->file_name, pos->line_number,
_("number of format specifications in 'msgid' and '%s' does not match"),
pretty_msgstr);
error_with_progname = true;
}
err = true;
}
else
for (i = 0; i < spec2->unnumbered_arg_count; i++)
if (spec1->unnumbered[i].type != spec2->unnumbered[i].type)
{
if (noisy)
{
error_with_progname = false;
error_at_line (0, 0, pos->file_name, pos->line_number,
_("format specifications in 'msgid' and '%s' for argument %u are not the same"),
pretty_msgstr, i + 1);
error_with_progname = true;
}
err = true;
}
return err;
}
struct formatstring_parser formatstring_c =
{
format_c_parse,
format_free,
format_get_number_of_directives,
format_check
};
struct formatstring_parser formatstring_objc =
{
format_objc_parse,
format_free,
format_get_number_of_directives,
format_check
};
void
get_c99_format_directives (const char *string,
struct interval **intervalsp, size_t *lengthp)
{
char *invalid_reason = NULL;
struct spec *descr =
(struct spec *) format_parse (string, true, &invalid_reason);
if (descr != NULL && descr->c99_directives_count > 0)
{
unsigned int n = descr->c99_directives_count;
struct interval *intervals =
(struct interval *) xmalloc (n * sizeof (struct interval));
unsigned int i;
for (i = 0; i < n; i++)
{
intervals[i].startpos = descr->c99_directives[2 * i] - string;
intervals[i].endpos = descr->c99_directives[2 * i + 1] - string;
}
*intervalsp = intervals;
*lengthp = n;
}
else
{
*intervalsp = NULL;
*lengthp = 0;
}
if (descr != NULL)
format_free (descr);
else
free (invalid_reason);
}
#ifdef TEST
#include <stdio.h>
#include "getline.h"
static void
format_print (void *descr)
{
struct spec *spec = (struct spec *) descr;
unsigned int i;
if (spec == NULL)
{
printf ("INVALID");
return;
}
printf ("(");
for (i = 0; i < spec->unnumbered_arg_count; i++)
{
if (i > 0)
printf (" ");
if (spec->unnumbered[i].type & FAT_UNSIGNED)
printf ("[unsigned]");
switch (spec->unnumbered[i].type & FAT_SIZE_MASK)
{
case 0:
break;
case FAT_SIZE_SHORT:
printf ("[short]");
break;
case FAT_SIZE_CHAR:
printf ("[char]");
break;
case FAT_SIZE_LONG:
printf ("[long]");
break;
case FAT_SIZE_LONGLONG:
printf ("[long long]");
break;
case FAT_SIZE_8_T:
printf ("[int8_t]");
break;
case FAT_SIZE_16_T:
printf ("[int16_t]");
break;
case FAT_SIZE_32_T:
printf ("[int32_t]");
break;
case FAT_SIZE_64_T:
printf ("[int64_t]");
break;
case FAT_SIZE_LEAST8_T:
printf ("[int_least8_t]");
break;
case FAT_SIZE_LEAST16_T:
printf ("[int_least16_t]");
break;
case FAT_SIZE_LEAST32_T:
printf ("[int_least32_t]");
break;
case FAT_SIZE_LEAST64_T:
printf ("[int_least64_t]");
break;
case FAT_SIZE_FAST8_T:
printf ("[int_fast8_t]");
break;
case FAT_SIZE_FAST16_T:
printf ("[int_fast16_t]");
break;
case FAT_SIZE_FAST32_T:
printf ("[int_fast32_t]");
break;
case FAT_SIZE_FAST64_T:
printf ("[int_fast64_t]");
break;
case FAT_SIZE_INTMAX_T:
printf ("[intmax_t]");
break;
case FAT_SIZE_INTPTR_T:
printf ("[intptr_t]");
break;
case FAT_SIZE_SIZE_T:
printf ("[size_t]");
break;
case FAT_SIZE_PTRDIFF_T:
printf ("[ptrdiff_t]");
break;
default:
abort ();
}
switch (spec->unnumbered[i].type & ~(FAT_UNSIGNED | FAT_SIZE_MASK))
{
case FAT_INTEGER:
printf ("i");
break;
case FAT_DOUBLE:
printf ("f");
break;
case FAT_CHAR:
printf ("c");
break;
case FAT_STRING:
printf ("s");
break;
case FAT_OBJC_OBJECT:
printf ("@");
break;
case FAT_POINTER:
printf ("p");
break;
case FAT_COUNT_POINTER:
printf ("n");
break;
default:
abort ();
}
}
printf (")");
}
int
main ()
{
for (;;)
{
char *line = NULL;
size_t line_size = 0;
int line_len;
char *invalid_reason;
void *descr;
line_len = getline (&line, &line_size, stdin);
if (line_len < 0)
break;
if (line_len > 0 && line[line_len - 1] == '\n')
line[--line_len] = '\0';
invalid_reason = NULL;
descr = format_c_parse (line, &invalid_reason);
format_print (descr);
printf ("\n");
if (descr == NULL)
printf ("%s\n", invalid_reason);
free (invalid_reason);
free (line);
}
return 0;
}
#endif