PostgreSQL Source Code  git master
saslprep.c File Reference
#include "postgres.h"
#include "common/saslprep.h"
#include "common/string.h"
#include "common/unicode_norm.h"
#include "mb/pg_wchar.h"
Include dependency graph for saslprep.c:

Go to the source code of this file.

Macros

#define STRDUP(s)   pstrdup(s)
 
#define ALLOC(size)   palloc(size)
 
#define FREE(size)   pfree(size)
 
#define IS_CODE_IN_TABLE(code, map)   is_code_in_table(code, map, lengthof(map))
 

Functions

static int codepoint_range_cmp (const void *a, const void *b)
 
static bool is_code_in_table (pg_wchar code, const pg_wchar *map, int mapsize)
 
static int pg_utf8_string_len (const char *source)
 
pg_saslprep_rc pg_saslprep (const char *input, char **output)
 

Variables

static const pg_wchar non_ascii_space_ranges []
 
static const pg_wchar commonly_mapped_to_nothing_ranges []
 
static const pg_wchar prohibited_output_ranges []
 
static const pg_wchar unassigned_codepoint_ranges []
 
static const pg_wchar RandALCat_codepoint_ranges []
 
static const pg_wchar LCat_codepoint_ranges []
 

Macro Definition Documentation

◆ ALLOC

#define ALLOC (   size)    palloc(size)

Definition at line 39 of file saslprep.c.

◆ FREE

#define FREE (   size)    pfree(size)

Definition at line 40 of file saslprep.c.

◆ IS_CODE_IN_TABLE

#define IS_CODE_IN_TABLE (   code,
  map 
)    is_code_in_table(code, map, lengthof(map))

Definition at line 965 of file saslprep.c.

◆ STRDUP

#define STRDUP (   s)    pstrdup(s)

Definition at line 38 of file saslprep.c.

Function Documentation

◆ codepoint_range_cmp()

static int codepoint_range_cmp ( const void *  a,
const void *  b 
)
static

Definition at line 968 of file saslprep.c.

969 {
970  const pg_wchar *key = (const pg_wchar *) a;
971  const pg_wchar *range = (const pg_wchar *) b;
972 
973  if (*key < range[0])
974  return -1; /* less than lower bound */
975  if (*key > range[1])
976  return 1; /* greater than upper bound */
977 
978  return 0; /* within range */
979 }
int b
Definition: isn.c:70
int a
Definition: isn.c:69
unsigned int pg_wchar
Definition: mbprint.c:31
static struct cvec * range(struct vars *v, chr a, chr b, int cases)
Definition: regc_locale.c:412

References a, b, sort-test::key, and range().

Referenced by is_code_in_table().

◆ is_code_in_table()

static bool is_code_in_table ( pg_wchar  code,
const pg_wchar map,
int  mapsize 
)
static

Definition at line 982 of file saslprep.c.

983 {
984  Assert(mapsize % 2 == 0);
985 
986  if (code < map[0] || code > map[mapsize - 1])
987  return false;
988 
989  if (bsearch(&code, map, mapsize / 2, sizeof(pg_wchar) * 2,
991  return true;
992  else
993  return false;
994 }
#define Assert(condition)
Definition: c.h:861
static int codepoint_range_cmp(const void *a, const void *b)
Definition: saslprep.c:968

References Assert, and codepoint_range_cmp().

◆ pg_saslprep()

pg_saslprep_rc pg_saslprep ( const char *  input,
char **  output 
)

Definition at line 1046 of file saslprep.c.

1047 {
1048  pg_wchar *input_chars = NULL;
1049  pg_wchar *output_chars = NULL;
1050  int input_size;
1051  char *result;
1052  int result_size;
1053  int count;
1054  int i;
1055  bool contains_RandALCat;
1056  unsigned char *p;
1057  pg_wchar *wp;
1058 
1059  /* Ensure we return *output as NULL on failure */
1060  *output = NULL;
1061 
1062  /*
1063  * Quick check if the input is pure ASCII. An ASCII string requires no
1064  * further processing.
1065  */
1066  if (pg_is_ascii(input))
1067  {
1068  *output = STRDUP(input);
1069  if (!(*output))
1070  goto oom;
1071  return SASLPREP_SUCCESS;
1072  }
1073 
1074  /*
1075  * Convert the input from UTF-8 to an array of Unicode codepoints.
1076  *
1077  * This also checks that the input is a legal UTF-8 string.
1078  */
1079  input_size = pg_utf8_string_len(input);
1080  if (input_size < 0)
1081  return SASLPREP_INVALID_UTF8;
1082 
1083  input_chars = ALLOC((input_size + 1) * sizeof(pg_wchar));
1084  if (!input_chars)
1085  goto oom;
1086 
1087  p = (unsigned char *) input;
1088  for (i = 0; i < input_size; i++)
1089  {
1090  input_chars[i] = utf8_to_unicode(p);
1091  p += pg_utf_mblen(p);
1092  }
1093  input_chars[i] = (pg_wchar) '\0';
1094 
1095  /*
1096  * The steps below correspond to the steps listed in [RFC3454], Section
1097  * "2. Preparation Overview"
1098  */
1099 
1100  /*
1101  * 1) Map -- For each character in the input, check if it has a mapping
1102  * and, if so, replace it with its mapping.
1103  */
1104  count = 0;
1105  for (i = 0; i < input_size; i++)
1106  {
1107  pg_wchar code = input_chars[i];
1108 
1110  input_chars[count++] = 0x0020;
1112  {
1113  /* map to nothing */
1114  }
1115  else
1116  input_chars[count++] = code;
1117  }
1118  input_chars[count] = (pg_wchar) '\0';
1119  input_size = count;
1120 
1121  if (input_size == 0)
1122  goto prohibited; /* don't allow empty password */
1123 
1124  /*
1125  * 2) Normalize -- Normalize the result of step 1 using Unicode
1126  * normalization.
1127  */
1128  output_chars = unicode_normalize(UNICODE_NFKC, input_chars);
1129  if (!output_chars)
1130  goto oom;
1131 
1132  /*
1133  * 3) Prohibit -- Check for any characters that are not allowed in the
1134  * output. If any are found, return an error.
1135  */
1136  for (i = 0; i < input_size; i++)
1137  {
1138  pg_wchar code = input_chars[i];
1139 
1141  goto prohibited;
1143  goto prohibited;
1144  }
1145 
1146  /*
1147  * 4) Check bidi -- Possibly check for right-to-left characters, and if
1148  * any are found, make sure that the whole string satisfies the
1149  * requirements for bidirectional strings. If the string does not satisfy
1150  * the requirements for bidirectional strings, return an error.
1151  *
1152  * [RFC3454], Section "6. Bidirectional Characters" explains in more
1153  * detail what that means:
1154  *
1155  * "In any profile that specifies bidirectional character handling, all
1156  * three of the following requirements MUST be met:
1157  *
1158  * 1) The characters in section 5.8 MUST be prohibited.
1159  *
1160  * 2) If a string contains any RandALCat character, the string MUST NOT
1161  * contain any LCat character.
1162  *
1163  * 3) If a string contains any RandALCat character, a RandALCat character
1164  * MUST be the first character of the string, and a RandALCat character
1165  * MUST be the last character of the string."
1166  */
1167  contains_RandALCat = false;
1168  for (i = 0; i < input_size; i++)
1169  {
1170  pg_wchar code = input_chars[i];
1171 
1173  {
1174  contains_RandALCat = true;
1175  break;
1176  }
1177  }
1178 
1179  if (contains_RandALCat)
1180  {
1181  pg_wchar first = input_chars[0];
1182  pg_wchar last = input_chars[input_size - 1];
1183 
1184  for (i = 0; i < input_size; i++)
1185  {
1186  pg_wchar code = input_chars[i];
1187 
1189  goto prohibited;
1190  }
1191 
1194  goto prohibited;
1195  }
1196 
1197  /*
1198  * Finally, convert the result back to UTF-8.
1199  */
1200  result_size = 0;
1201  for (wp = output_chars; *wp; wp++)
1202  {
1203  unsigned char buf[4];
1204 
1205  unicode_to_utf8(*wp, buf);
1206  result_size += pg_utf_mblen(buf);
1207  }
1208 
1209  result = ALLOC(result_size + 1);
1210  if (!result)
1211  goto oom;
1212 
1213  /*
1214  * There are no error exits below here, so the error exit paths don't need
1215  * to worry about possibly freeing "result".
1216  */
1217  p = (unsigned char *) result;
1218  for (wp = output_chars; *wp; wp++)
1219  {
1220  unicode_to_utf8(*wp, p);
1221  p += pg_utf_mblen(p);
1222  }
1223  Assert((char *) p == result + result_size);
1224  *p = '\0';
1225 
1226  FREE(input_chars);
1227  FREE(output_chars);
1228 
1229  *output = result;
1230  return SASLPREP_SUCCESS;
1231 
1232 prohibited:
1233  if (input_chars)
1234  FREE(input_chars);
1235  if (output_chars)
1236  FREE(output_chars);
1237 
1238  return SASLPREP_PROHIBITED;
1239 
1240 oom:
1241  if (input_chars)
1242  FREE(input_chars);
1243  if (output_chars)
1244  FREE(output_chars);
1245 
1246  return SASLPREP_OOM;
1247 }
FILE * input
FILE * output
int i
Definition: isn.c:73
static pg_wchar utf8_to_unicode(const unsigned char *c)
Definition: mbprint.c:53
static char * buf
Definition: pg_test_fsync.c:73
#define pg_utf_mblen
Definition: pg_wchar.h:633
static unsigned char * unicode_to_utf8(pg_wchar c, unsigned char *utf8string)
Definition: pg_wchar.h:575
static const pg_wchar unassigned_codepoint_ranges[]
Definition: saslprep.c:157
static const pg_wchar non_ascii_space_ranges[]
Definition: saslprep.c:66
static const pg_wchar RandALCat_codepoint_ranges[]
Definition: saslprep.c:558
#define STRDUP(s)
Definition: saslprep.c:38
#define IS_CODE_IN_TABLE(code, map)
Definition: saslprep.c:965
#define ALLOC(size)
Definition: saslprep.c:39
#define FREE(size)
Definition: saslprep.c:40
static const pg_wchar LCat_codepoint_ranges[]
Definition: saslprep.c:597
static const pg_wchar commonly_mapped_to_nothing_ranges[]
Definition: saslprep.c:81
static const pg_wchar prohibited_output_ranges[]
Definition: saslprep.c:116
static int pg_utf8_string_len(const char *source)
Definition: saslprep.c:1002
@ SASLPREP_INVALID_UTF8
Definition: saslprep.h:24
@ SASLPREP_PROHIBITED
Definition: saslprep.h:25
@ SASLPREP_OOM
Definition: saslprep.h:23
@ SASLPREP_SUCCESS
Definition: saslprep.h:22
bool pg_is_ascii(const char *str)
Definition: string.c:133
pg_wchar * unicode_normalize(UnicodeNormalizationForm form, const pg_wchar *input)
Definition: unicode_norm.c:402
@ UNICODE_NFKC
Definition: unicode_norm.h:23

References ALLOC, Assert, buf, commonly_mapped_to_nothing_ranges, FREE, i, input, IS_CODE_IN_TABLE, LCat_codepoint_ranges, non_ascii_space_ranges, output, pg_is_ascii(), pg_utf8_string_len(), pg_utf_mblen, prohibited_output_ranges, RandALCat_codepoint_ranges, SASLPREP_INVALID_UTF8, SASLPREP_OOM, SASLPREP_PROHIBITED, SASLPREP_SUCCESS, STRDUP, unassigned_codepoint_ranges, UNICODE_NFKC, unicode_normalize(), unicode_to_utf8(), and utf8_to_unicode().

Referenced by pg_be_scram_build_secret(), pg_fe_scram_build_secret(), scram_init(), and scram_verify_plain_password().

◆ pg_utf8_string_len()

static int pg_utf8_string_len ( const char *  source)
static

Definition at line 1002 of file saslprep.c.

1003 {
1004  const unsigned char *p = (const unsigned char *) source;
1005  int l;
1006  int num_chars = 0;
1007  size_t len = strlen(source);
1008 
1009  while (len)
1010  {
1011  l = pg_utf_mblen(p);
1012 
1013  if (len < l || !pg_utf8_islegal(p, l))
1014  return -1;
1015 
1016  p += l;
1017  len -= l;
1018  num_chars++;
1019  }
1020 
1021  return num_chars;
1022 }
const void size_t len
static rewind_source * source
Definition: pg_rewind.c:89
bool pg_utf8_islegal(const unsigned char *source, int length)
Definition: wchar.c:1953

References len, pg_utf8_islegal(), pg_utf_mblen, and source.

Referenced by pg_saslprep().

Variable Documentation

◆ commonly_mapped_to_nothing_ranges

const pg_wchar commonly_mapped_to_nothing_ranges[]
static
Initial value:
=
{
0x00AD, 0x00AD,
0x034F, 0x034F,
0x1806, 0x1806,
0x180B, 0x180D,
0x200B, 0x200D,
0x2060, 0x2060,
0xFE00, 0xFE0F,
0xFEFF, 0xFEFF
}

Definition at line 81 of file saslprep.c.

Referenced by pg_saslprep().

◆ LCat_codepoint_ranges

const pg_wchar LCat_codepoint_ranges[]
static

Definition at line 597 of file saslprep.c.

Referenced by pg_saslprep().

◆ non_ascii_space_ranges

const pg_wchar non_ascii_space_ranges[]
static
Initial value:
=
{
0x00A0, 0x00A0,
0x1680, 0x1680,
0x2000, 0x200B,
0x202F, 0x202F,
0x205F, 0x205F,
0x3000, 0x3000
}

Definition at line 66 of file saslprep.c.

Referenced by pg_saslprep().

◆ prohibited_output_ranges

const pg_wchar prohibited_output_ranges[]
static

Definition at line 116 of file saslprep.c.

Referenced by pg_saslprep().

◆ RandALCat_codepoint_ranges

const pg_wchar RandALCat_codepoint_ranges[]
static

Definition at line 558 of file saslprep.c.

Referenced by pg_saslprep().

◆ unassigned_codepoint_ranges

const pg_wchar unassigned_codepoint_ranges[]
static

Definition at line 157 of file saslprep.c.

Referenced by pg_saslprep().