diff --git a/TODO b/TODO
index adf05021628b45df71a065930f06a77e4ef714ea..c5a0c9550e93fd5e2c1eb00aa0df5fb70eea8e87 100644
--- a/TODO
+++ b/TODO
@@ -1,8 +1,6 @@
Next (in order of descending priority/precedence):
* Full standards compliance review for the engine and all fully implemented
sieve extensions. Issues discovered so far:
- - Encoded character processing must not raise an error on substitutions
- that are not syntactically valid.
- :matches : match values must only be changed when the test matches.
- If an address is not syntactically valid, then it will not be matched
by tests specifying ":localpart" or ":domain".
diff --git a/src/lib-sieve/ext-encoded-character.c b/src/lib-sieve/ext-encoded-character.c
index 08a43e4c65f9acc11c1dd4fe83c58d04341f4adf..1097b3eeebbcd4bf7d417a2fa53284ce981299f2 100644
--- a/src/lib-sieve/ext-encoded-character.c
+++ b/src/lib-sieve/ext-encoded-character.c
@@ -118,12 +118,12 @@ static bool _parse_hexint
return ( digit > 0 );
}
-static int _decode_hex
+static bool _decode_hex
(const char **in, const char *inend, string_t *result)
{
int values = 0;
- for (;;) {
+ while ( *in < inend ) {
unsigned int hexpair;
if ( !_skip_whitespace(in, inend) ) return FALSE;
@@ -134,19 +134,19 @@ static int _decode_hex
values++;
}
- return ( values > 0 ? 1 : 0 );
+ return ( values > 0 );
}
static int _decode_unicode
-(struct sieve_validator *validator, struct sieve_command_context *cmd,
- const char **in, const char *inend, string_t *result)
+(const char **in, const char *inend, string_t *result, unsigned int *error_hex)
{
int values = 0;
+ bool valid = TRUE;
- for (;;) {
+ while ( *in < inend ) {
unsigned int unicode_hex;
- if ( !_skip_whitespace(in, inend) ) return 0;
+ if ( !_skip_whitespace(in, inend) ) return FALSE;
if ( !_parse_hexint(in, inend, 0, &unicode_hex) ) break;
@@ -154,15 +154,13 @@ static int _decode_unicode
(unicode_hex >= 0xE000 && unicode_hex <= 0x10FFFF) )
uni_ucs4_to_utf8_c((unichar_t) unicode_hex, result);
else {
- sieve_command_validate_error(validator, cmd,
- "invalid unicode character 0x%08x in encoded character substitution",
- unicode_hex);
- return -1;
+ if ( valid ) *error_hex = unicode_hex;
+ valid = FALSE;
}
values++;
}
- return ( values > 0 ? 1 : 0 );
+ return ( values > 0 );
}
bool arg_encoded_string_validate
@@ -170,7 +168,6 @@ bool arg_encoded_string_validate
struct sieve_command_context *cmd)
{
bool result = TRUE;
- int ret;
enum { ST_NONE, ST_OPEN, ST_TYPE, ST_CLOSE }
state = ST_NONE;
string_t *str = sieve_ast_argument_str(*arg);
@@ -178,6 +175,7 @@ bool arg_encoded_string_validate
const char *p, *mark, *strstart, *substart = NULL;
const char *strval = (const char *) str_data(str);
const char *strend = strval + str_len(str);
+ unsigned int error_hex = 0;
T_BEGIN {
tmpstr = t_str_new(32);
@@ -219,19 +217,13 @@ bool arg_encoded_string_validate
if ( strncasecmp(mark, "hex", p - mark) == 0 ) {
/* Hexadecimal */
p++;
- ret = _decode_hex(&p, strend, tmpstr);
- if ( ret <= 0 ) {
+ if ( !_decode_hex(&p, strend, tmpstr) )
state = ST_NONE;
- if ( ret < 0 ) result = FALSE;
- }
} else if ( strncasecmp(mark, "unicode", p - mark) == 0 ) {
/* Unicode */
p++;
- ret = _decode_unicode(validator, cmd, &p, strend, tmpstr);
- if ( ret <= 0 ) {
+ if ( !_decode_unicode(&p, strend, tmpstr, &error_hex) )
state = ST_NONE;
- if ( ret < 0 ) result = FALSE;
- }
} else {
/* Invalid encoding */
p++;
@@ -241,6 +233,14 @@ bool arg_encoded_string_validate
case ST_CLOSE:
if ( *p == '}' ) {
/* We now know that the substitution is valid */
+
+ if ( error_hex != 0 ) {
+ sieve_command_validate_error(validator, cmd,
+ "invalid unicode character 0x%08x in encoded character substitution",
+ error_hex);
+ result = FALSE;
+ break;
+ }
if ( newstr == NULL ) {
newstr = str_new(sieve_ast_pool((*arg)->ast), str_len(str)*2);
diff --git a/src/testsuite/tests/compile/errors/encoded-character.sieve b/src/testsuite/tests/compile/errors/encoded-character.sieve
index 698eb0801b1bca84d0aa7f440d0a3a4a596e2e7d..04d9de40d4f788af5366878f65ae168e8a76963a 100644
--- a/src/testsuite/tests/compile/errors/encoded-character.sieve
+++ b/src/testsuite/tests/compile/errors/encoded-character.sieve
@@ -11,13 +11,13 @@ require "fileinto";
fileinto "INBOX.${unicode:200000}";
# Not an error
-#fileinto "INBOX.${unicode:200000";
+fileinto "INBOX.${unicode:200000";
# Invalid unicode character (2)
fileinto "INBOX.${Unicode:DF01}";
# Not an error
-#fileinto "INBOX.${Unicode:DF01";
+fileinto "INBOX.${Unicode:DF01";
diff --git a/src/testsuite/tests/extensions/encoded-character.svtest b/src/testsuite/tests/extensions/encoded-character.svtest
index e241495ba29171833ffce9aaeefc61d401c83b3b..b05e2e4c6858091331ad717d6755c20919c37cab 100644
--- a/src/testsuite/tests/extensions/encoded-character.svtest
+++ b/src/testsuite/tests/extensions/encoded-character.svtest
@@ -83,6 +83,48 @@ test "HEX equality braindead" {
}
}
+test "Syntax errors" {
+ if anyof( not string "$" "${hex:24}", not string "$ " "${hex:24} ", not string " $" " ${hex:24}" ) {
+ test_fail "loose $ handled inappropriately";
+ }
+
+ if anyof( not string "${" "${hex:24}{", not string "a${" "a${hex:24}{", not string "${a" "${hex:24}{a" ) {
+ test_fail "loose ${ handled inappropriately";
+ }
+
+ if anyof( not string "${}" "${hex:24}{}", not string "b${}" "b${hex:24}{}", not string "${}b" "${hex:24}{}b" ) {
+ test_fail "entirely missing content handled inappropriately";
+ }
+
+ if not string "${:}" "${hex:24}{:}" {
+ test_fail "missing content handled inappropriately";
+ }
+
+ if not string "${hex:}" "${hex:24}{hex:}" {
+ test_fail "missing hex content handled inappropriately";
+ }
+
+ if not string "${unicode:}" "${hex:24}{unicode:}" {
+ test_fail "missing unicode content handled inappropriately";
+ }
+
+ if not string "${hex:sss}" "${hex:24}{hex:sss}" {
+ test_fail "erroneous hex content handled inappropriately";
+ }
+
+ if not string "${unicode:ttt}" "${hex:24}{unicode:ttt}" {
+ test_fail "erroneous unicode content handled inappropriately";
+ }
+
+ if not string "${hex:aa aa" "${hex:24}{hex:aa aa" {
+ test_fail "unterminated hex content handled inappropriately";
+ }
+
+ if not string "${unicode: aaaa aaaa" "${hex:24}{unicode: aaaa aaaa" {
+ test_fail "unterminated unicode content handled inappropriately";
+ }
+}
+
/*
* RFC Examples
*/