server: Accept utf-8 data as input in registry files.
diff --git a/server/registry.c b/server/registry.c
index d691e04..fa39279 100644
--- a/server/registry.c
+++ b/server/registry.c
@@ -168,12 +168,6 @@
* - REG_EXPAND_SZ and REG_MULTI_SZ are saved as strings instead of hex
*/
-static inline char to_hex( char ch )
-{
- if (isdigit(ch)) return ch - '0';
- return tolower(ch) - 'a' + 10;
-}
-
/* dump the full path of a key */
static void dump_path( const struct key *key, const struct key *base, FILE *f )
{
@@ -1076,65 +1070,6 @@
fprintf( stderr, "<fd>:%d: %s '%s'\n", info->line, err, info->buffer );
}
-/* parse an escaped string back into Unicode */
-/* return the number of chars read from the input, or -1 on output overflow */
-static int parse_strW( WCHAR *dest, data_size_t *len, const char *src, char endchar )
-{
- data_size_t count = sizeof(WCHAR); /* for terminating null */
- const char *p = src;
- while (*p && *p != endchar)
- {
- if (*p != '\\') *dest = (WCHAR)*p++;
- else
- {
- p++;
- switch(*p)
- {
- case 'a': *dest = '\a'; p++; break;
- case 'b': *dest = '\b'; p++; break;
- case 'e': *dest = '\e'; p++; break;
- case 'f': *dest = '\f'; p++; break;
- case 'n': *dest = '\n'; p++; break;
- case 'r': *dest = '\r'; p++; break;
- case 't': *dest = '\t'; p++; break;
- case 'v': *dest = '\v'; p++; break;
- case 'x': /* hex escape */
- p++;
- if (!isxdigit(*p)) *dest = 'x';
- else
- {
- *dest = to_hex(*p++);
- if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
- if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
- if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
- }
- break;
- case '0':
- case '1':
- case '2':
- case '3':
- case '4':
- case '5':
- case '6':
- case '7': /* octal escape */
- *dest = *p++ - '0';
- if (*p >= '0' && *p <= '7') *dest = (*dest * 8) + (*p++ - '0');
- if (*p >= '0' && *p <= '7') *dest = (*dest * 8) + (*p++ - '0');
- break;
- default:
- *dest = (WCHAR)*p++;
- break;
- }
- }
- if ((count += sizeof(WCHAR)) > *len) return -1; /* dest buffer overflow */
- dest++;
- }
- *dest = 0;
- if (!*p) return -1; /* delimiter not found */
- *len = count;
- return p + 1 - src;
-}
-
/* convert a data type tag to a value type */
static int get_data_type( const char *buffer, int *type, int *parse_type )
{
@@ -1176,10 +1111,11 @@
WCHAR *p;
struct unicode_str name;
int res, modif;
- data_size_t len = strlen(buffer) * sizeof(WCHAR);
+ data_size_t len;
- if (!get_file_tmp_space( info, len )) return NULL;
+ if (!get_file_tmp_space( info, strlen(buffer) * sizeof(WCHAR) )) return NULL;
+ len = info->tmplen;
if ((res = parse_strW( info->tmp, &len, buffer, ']' )) == -1)
{
file_read_error( "Malformed key", info );
@@ -1233,10 +1169,10 @@
struct key_value *value;
struct unicode_str name;
int index;
- data_size_t maxlen = strlen(buffer) * sizeof(WCHAR);
- if (!get_file_tmp_space( info, maxlen )) return NULL;
+ if (!get_file_tmp_space( info, strlen(buffer) * sizeof(WCHAR) )) return NULL;
name.str = info->tmp;
+ name.len = info->tmplen;
if (buffer[0] == '@')
{
name.len = 0;
@@ -1244,10 +1180,10 @@
}
else
{
- int r = parse_strW( info->tmp, &maxlen, buffer + 1, '\"' );
+ int r = parse_strW( info->tmp, &name.len, buffer + 1, '\"' );
if (r == -1) goto error;
*len = r + 1; /* for initial quote */
- name.len = maxlen - sizeof(WCHAR);
+ name.len -= sizeof(WCHAR); /* terminating null */
}
while (isspace(buffer[*len])) (*len)++;
if (buffer[*len] != '=') goto error;
@@ -1277,8 +1213,8 @@
switch(parse_type)
{
case REG_SZ:
- len = strlen(buffer) * sizeof(WCHAR);
- if (!get_file_tmp_space( info, len )) return 0;
+ if (!get_file_tmp_space( info, strlen(buffer) * sizeof(WCHAR) )) return 0;
+ len = info->tmplen;
if ((res = parse_strW( info->tmp, &len, buffer, '\"' )) == -1) goto error;
ptr = info->tmp;
break;
@@ -1332,10 +1268,11 @@
{
WCHAR *p;
int res;
- data_size_t len = strlen(name) * sizeof(WCHAR);
+ data_size_t len;
- if (!get_file_tmp_space( info, len )) return 0;
+ if (!get_file_tmp_space( info, strlen(name) * sizeof(WCHAR) )) return 0;
+ len = info->tmplen;
if ((res = parse_strW( info->tmp, &len, name, ']' )) == -1)
{
file_read_error( "Malformed key", info );
diff --git a/server/unicode.c b/server/unicode.c
index cdbf597..3be1f55 100644
--- a/server/unicode.c
+++ b/server/unicode.c
@@ -26,8 +26,130 @@
#include "unicode.h"
+/* number of following bytes in sequence based on first byte value (for bytes above 0x7f) */
+static const char utf8_length[128] =
+{
+ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 0x80-0x8f */
+ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 0x90-0x9f */
+ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 0xa0-0xaf */
+ 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, /* 0xb0-0xbf */
+ 0,0,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 0xc0-0xcf */
+ 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, /* 0xd0-0xdf */
+ 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, /* 0xe0-0xef */
+ 3,3,3,3,3,0,0,0,0,0,0,0,0,0,0,0 /* 0xf0-0xff */
+};
+
+/* first byte mask depending on UTF-8 sequence length */
+static const unsigned char utf8_mask[4] = { 0x7f, 0x1f, 0x0f, 0x07 };
+
+/* minimum Unicode value depending on UTF-8 sequence length */
+static const unsigned int utf8_minval[4] = { 0x0, 0x80, 0x800, 0x10000 };
+
+static inline char to_hex( char ch )
+{
+ if (isdigit(ch)) return ch - '0';
+ return tolower(ch) - 'a' + 10;
+}
+
+/* parse an escaped string back into Unicode */
+/* return the number of chars read from the input, or -1 on output overflow */
+int parse_strW( WCHAR *buffer, data_size_t *len, const char *src, char endchar )
+{
+ WCHAR *dest = buffer;
+ WCHAR *end = buffer + *len / sizeof(WCHAR);
+ const char *p = src;
+ unsigned char ch;
+
+ while (*p && *p != endchar && dest < end)
+ {
+ if (*p == '\\')
+ {
+ p++;
+ if (!*p) break;
+ switch(*p)
+ {
+ case 'a': *dest++ = '\a'; p++; continue;
+ case 'b': *dest++ = '\b'; p++; continue;
+ case 'e': *dest++ = '\e'; p++; continue;
+ case 'f': *dest++ = '\f'; p++; continue;
+ case 'n': *dest++ = '\n'; p++; continue;
+ case 'r': *dest++ = '\r'; p++; continue;
+ case 't': *dest++ = '\t'; p++; continue;
+ case 'v': *dest++ = '\v'; p++; continue;
+ case 'x': /* hex escape */
+ p++;
+ if (!isxdigit(*p)) *dest = 'x';
+ else
+ {
+ *dest = to_hex(*p++);
+ if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
+ if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
+ if (isxdigit(*p)) *dest = (*dest * 16) + to_hex(*p++);
+ }
+ dest++;
+ continue;
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7': /* octal escape */
+ *dest = *p++ - '0';
+ if (*p >= '0' && *p <= '7') *dest = (*dest * 8) + (*p++ - '0');
+ if (*p >= '0' && *p <= '7') *dest = (*dest * 8) + (*p++ - '0');
+ dest++;
+ continue;
+ }
+ /* unrecognized escape: fall through to normal char handling */
+ }
+
+
+ ch = *p++;
+ if (ch < 0x80) *dest++ = ch;
+ else /* parse utf8 char */
+ {
+ int charlen = utf8_length[ch-0x80];
+ unsigned int res = ch & utf8_mask[charlen];
+
+ switch(charlen)
+ {
+ case 3:
+ if ((ch = *p ^ 0x80) >= 0x40) break;
+ res = (res << 6) | ch;
+ p++;
+ case 2:
+ if ((ch = *p ^ 0x80) >= 0x40) break;
+ res = (res << 6) | ch;
+ p++;
+ case 1:
+ if ((ch = *p ^ 0x80) >= 0x40) break;
+ res = (res << 6) | ch;
+ p++;
+ if (res < utf8_minval[charlen]) break;
+ if (res > 0x10ffff) break;
+ if (res <= 0xffff) *dest++ = res;
+ else /* we need surrogates */
+ {
+ res -= 0x10000;
+ *dest++ = 0xd800 | (res >> 10);
+ if (dest < end) *dest++ = 0xdc00 | (res & 0x3ff);
+ }
+ continue;
+ }
+ /* ignore invalid char */
+ }
+ }
+ if (dest >= end) return -1; /* overflow */
+ *dest++ = 0;
+ if (!*p) return -1; /* delimiter not found */
+ *len = (dest - buffer) * sizeof(WCHAR);
+ return p + 1 - src;
+}
+
/* dump a Unicode string with proper escaping */
-int dump_strW( const WCHAR *str, size_t len, FILE *f, const char escape[2] )
+int dump_strW( const WCHAR *str, data_size_t len, FILE *f, const char escape[2] )
{
static const char escapes[32] = ".......abtnvfr.............e....";
char buffer[256];
diff --git a/server/unicode.h b/server/unicode.h
index f6dfb4e..dd922db 100644
--- a/server/unicode.h
+++ b/server/unicode.h
@@ -33,6 +33,7 @@
return memdup( str, len );
}
-extern int dump_strW( const WCHAR *str, size_t len, FILE *f, const char escape[2] );
+extern int parse_strW( WCHAR *buffer, data_size_t *len, const char *src, char endchar );
+extern int dump_strW( const WCHAR *str, data_size_t len, FILE *f, const char escape[2] );
#endif /* __WINE_SERVER_UNICODE_H */
